social sciences $€ £ ¥

Article Googling : Fashion Consumer Behaviour Using Trends

Emmanuel Sirimal Silva 1,* , Hossein Hassani 2 , Dag Øivind Madsen 3 and Liz Gee 4

1 Centre for Fashion Business and Innovation Research, Fashion Business School, London College of Fashion, University of the Arts London, 272 High Holborn, London WC1V 7EY, UK 2 Research Institute of Energy Management and Planning, University of Tehran, Tehran 1417466191, Iran; hassani.stat@.com 3 Department of Business, Marketing and Law, School of Business, University of South-Eastern Norway, Bredalsveien 14, 3511 Hønefoss, Norway; [email protected] 4 Centre for Fashion Business and Innovation Research, Fashion Business School, London College of Fashion, University of the Arts London, 272 High Holborn, London WC1V 7EY, UK; [email protected] * Correspondence: [email protected]

 Received: 22 February 2019; Accepted: 29 March 2019; Published: 4 April 2019 

Abstract: This paper aims to discuss the current state of Google Trends as a useful tool for fashion consumer analytics, show the importance of being able to forecast fashion consumer trends and then presents a univariate forecast evaluation of fashion consumer Google Trends to motivate more academic research in this subject area. Using Burberry—a British luxury fashion house—as an example, we compare several parametric and nonparametric forecasting techniques to determine the best univariate forecasting model for “Burberry” Google Trends. In addition, we also introduce singular spectrum analysis as a useful tool for denoising fashion consumer Google Trends and apply a recently developed hybrid neural network model to generate forecasts. Our initial results indicate that there is no single univariate model (out of ARIMA, , TBATS, and neural network autoregression) that can provide the best forecast of fashion consumer Google Trends for Burberry across all horizons. In fact, we find neural network autoregression (NNAR) to be the worst contender. We then seek to improve the accuracy of NNAR forecasts for fashion consumer Google Trends via the introduction of singular spectrum analysis for noise reduction in fashion data. The hybrid neural network model (Denoised NNAR) succeeds in outperforming all competing models across all horizons, with a majority of statistically significant outcomes at providing the best forecast for Burberry’s highly seasonal fashion consumer Google Trends. In an era of , we show the usefulness of Google Trends, denoising and forecasting consumer behaviour for the fashion industry.

Keywords: Google Trends; fashion; forecast; neural networks; singular spectrum analysis; big data

1. Introduction The emergence of big data and advances in big data analytics led to the creation of Google Trends (Choi and Varian 2012), a tool and source for analysing big data on web searches across the globe. Given that 70% of luxury purchases are estimated to be influenced by online interactions (D’Arpizio and Levato 2017), Google Trends has the potential to play a pivotal role in the developments of big data in fashion. Fashion consumers are actively using Google, looking for ideas, finding the best designs, and buying with a tap (Boone 2016). In 2016 alone, Google was receiving more than 4 million search queries per minute from 2.4 billion internet users and was processing 20 petabytes of information per day (Wedel and Kannan 2016). Data analytics on such online behaviour lets Google predict the next big fashion trend (Bain 2016) with Google’s Online Retail Monitor indicating that in 2018, fashion has

Soc. Sci. 2019, 8, 111; doi:10.3390/socsci8040111 www.mdpi.com/journal/socsci Sustainability 2019, 11, x FOR PEER REVIEW 2 of 24

Soc. Sci. 2019, 8, 111 2 of 23 lets Google predict the next big fashion trend (Bain 2016) with Google’s Online Retail Monitor indicating that in 2018, fashion has seen the highest growth in searches, boosted by overseas shoppers seen(especially the highest from the growth EU) inseeking searches, access boosted to UK by brands overseas online shoppers (Jahshan (especially 2018). Such from trends the EU)are positive seeking accessfor the to fashion UK brands industry online in (Jahshan the UK 2018 which). Such is trendsexperiencing are positive considerable for the fashion uncertainty industry with in theBrexit UK whichlooming is experiencingon the horizon. considerable uncertainty with Brexit looming on the horizon. Google TrendsTrends is is a gooda good example example of howof how big databig da canta becan exploited be exploited and visualised and visualised in a user-friendly in a user- style.friendly The style. term The big term data big itself data struggles itself struggles to find ato universal find a universal definition definition and it can and mean it can different mean different things tothings different to different people people (Marr (Marr2015; 2015;Hassani Hassani and Silva and Silva2015b 2015b).). Nevertheless, Nevertheless, most most researchers researchers agree agree on buildingon building upon upon the threethe three defining defining dimensions dimensions of Big of Data Big (3Vs) Data as(3Vs) introduced as introduced by Laney by(2001 Laney): volume, (2001): varietyvolume, and variety velocity. and velocity. Figure1 below Figure summarises 1 below summaris the 3Vses to the give 3Vs the to reader give the an reader indication an indication on how big on datahow wasbig data initially was thought initially to thought be expanding. to be expanding. Today, the definition Today, the of bigdefinition data has of evolved big data (much has evolved like the size(much of biglike data the size which of hasbig onlydata gottenwhich biggerhas only as predictedgotten bigger by Varian as predicted 2014) and by nowVarian includes 2014) 5Vs,and withnow additionalincludes 5Vs, Vs beingwith additional veracity (which Vs being accounts veracity for (which the quality accounts of the for data) the andquality value of (whichthe data) accounts and value for analytics(which accounts on data) for (Hassani analytics and on Silva data) 2018 (Hassani). and Silva 2018).

Figure 1. The 3Vs of big data (Soubra (Soubra 20122012).).

It is no secret thatthat GoogleGoogle Trends areare increasinglyincreasingly influencinginfluencing businessbusiness decision-makingdecision-making in a variety of industries (see, for example, YuYu et al. 2019;2019; Siliverstovs and and Wochner Wochner 2018;2018; ZhangZhang et et al. al. 20182018)) given its ability to act as a leading indicator for forecasting key variables of interest. The fashion industry tootoo cancan benefit benefit from from the the exploitation exploitation of Googleof Google Trends Trends for forecastingfor forecasting fashion fashion variables, variables, from predictingfrom predicting future future purchase purchase decisions, decisions, to determining to determining the effectiveness the effectiveness of marketing of marketing campaigns campaigns and forecastingand forecasting online online consumer consumer brand brand engagement. engagement. Moreover, Moreover, there is there a need is fora need more for conclusive more conclusive research whichresearch evaluates which evaluates whether bigwhether data in big the data form in of the Google form of Trends Google can Trends help predict can help actual predict sales actual for fashion sales brands.for fashion Whilst brands. finding Whilst the answerfindingto the this answer problem to this is beyond problem the is scope beyond of our the research, scope of there our research, is reason tothere believe is reason that thisto believe could that be the this case, could since be the evidence case, since suggests evidence that Googlesuggests searches that Google can predictsearches other can typespredict of other economic types activity of economic such as activity real estate such sales as real and estate prices sales (Wu and Brynjolfssonprices (Wu and 2015 Brynjolfsson), exchange rates2015), ( Bulutexchange 2018 rates), UK (Bulut cinema 2018), admissions UK cinema (Hand admissions and Judge (Hand2011), stockand Judge market 2011), volatility stock (Hamidmarket andvolatility Heiden (Hamid 2015), inflationand Heiden expectations 2015), inflation (Guzman ex 2011pectations) and tourist(Guzman arrivals 2011) (Bangwayo-Skeete and tourist arrivals and Skeete(Bangwayo-Skeete 2015). In addition, and Skeete there 2015). could In beaddition, several ther alternatee could research be several avenues alternate which research are waiting avenues to bewhich explored are waiting not only to from be explored a fashion not design, only buying from a and design, buying perspective, and merchandising but also from Soc. Sci. 2019, 8, 111 3 of 23 a fashion management perspective (see, for example, Madsen 2016), particularly in evaluating the success of marketing and social media campaigns. We subscribe to Gordon’s (2017) view that Google Trends can be a metric for online consumer behaviour, as more than 75% of the world’s internet searches are conducted on Google (Net Market Share 2019). Therefore, we believe that the fashion industry should consider relevant Google Trends as ‘fashion consumer Google Trends’ which according to McDowell(2019) can enable brands to identify consumer patterns and profit from them. Here, it is worthwhile to define Google Trends for the reader. In brief, as one of the largest real time datasets currently available (Rogers 2016), recording data from 2004 to present (Choi and Varian 2012), Google Trends allows one to gauge consumer search interest in brands. However, instead of the raw level of queries for a given search term, it is important to note that Google Trends reports the query index, which begins with a query share (Choi and Varian 2012):

Total query volume for search term in a given geographic location Total number of queries in that region at a point in time

In other words, its normalised (which enables more accurate comparisons over time) means that Google Trends will always show the search interest on a topic as a proportion of all searches on all topics on Google at that time and location (Rogers 2016). Data quality is another important consideration, and Google Trends seeks to improve the quality of its data by excluding searches made by very few people, duplicate searches and special characters (Google 2019; Choi and Varian 2012). Furthermore, the query share based approach to computing Google Trends has its benefits in a world where big data and data mining have been marred by privacy issues and concerns (Hassani et al. 2014, 2016a). The data aggregation underlying Google Trends ensures the output is anonymised and thus no individual is identified personally (Rogers 2016). The motivations for this research (and its importance) stems from several existing studies. Firstly, Jun et al.(2018) notes that the purpose of big data utilisation is now shifting from monitoring towards forecasting, and thereby, indicating the importance of predictive analytics and forecasting for the future. Secondly, the increase in ‘research shopping’, whereby consumers are seen accessing information via one channel and purchasing through another channel (Verhoef et al. 2007; Bradlow et al. 2017), adds more importance to the potential of fashion consumer Google Trends to be a useful fashion analytics tool. Thirdly, as LaValle et al.(2011) points out, organisations are interested in what is likely to happen next, and forecasting is a tool which can provide this information. However, the emergence of big data brings about its own challenges for generating accurate forecasts (Hassani and Silva 2015b). Fourthly, Wedel and Kannan(2016) assert that trend forecasting is vital for companies to be able to identify changes in the environment and set up defences to retain market share. Furthermore, the fashion industry currently benefits from big data trend forecasts and analytics through popular and well reputed services by WGSN and Edited. As discussed later, Google Trends has the potential to complement these existing platforms. To this end, Google Trends can indicate consumer sentiment towards a brand, and has the ability to extrapolate this to potential purchasing behaviour that can help brands plan more effectively. Thus, our interest lies in understanding the benefits of Google Trends for fashion analytics and identifying the possibility of forecasting such online consumer trends into the future so that more productive managerial and marketing decisions can be made. Accordingly, the aim of this paper is to determine whether there exists a single univariate forecasting model which can predict fashion consumer Google Trends across both short and long run horizons. The following objectives are put forward to help achieve this aim. (1) Identify the uses of Google Trends for predicting fashion consumer behaviour and the need for forecasting same, (2) analyse parametric and nonparametric univariate models at forecasting fashion consumer Google Trends and (3) evaluate the importance of signal extraction and denoising for fashion analytics. Soc. Sci. 2019, 8, 111 4 of 23

Accordingly, this study has several contributions; the first of which it is the initial attempt at forecasting Google Trends for fashion. Secondly, this paper marks the introductory application of the Denoised Neural Network Autorgression (DNNAR) model of Silva et al.(2019), which incorporates SingularSustainability Spectrum 2019, 11 Analysis, x FOR PEER (SSA) REVIEW (Broomhead and King 1986a, 1986b) and the Trigonometric4 of Box–Cox 24 ARMACox trend ARMA seasonal trend seasonal (TBATS) (TBATS) model formodel improving for improving the accuracy the accuracy of forecasts of forecasts in the in fashionthe fashion industry. Thisindustry. contribution This contribution is important is important as Bradlow as Bradlow et al.(2017 et al.) points(2017) points out that out that new new research research insights insights arise eitherarise from either new from data, new new data, methods new methods or some or combination some combination of the two.of the Thirdly, two. Thirdly, the forecast the forecast evaluation presentedevaluation herewith presented compares herewith five compares popular andfive powerfulpopular and univariate powerful time univariate series analysistime series and analysis forecasting techniquesand forecasting covering techniques both parametric covering and both nonparametric parametric and models. nonparametric Section 4models. provides Section more 4 detailprovides around the importancemore detail ofaround each the chosen importance model, of what each theychosen do model, and how what they they are do used and inhow this they study. are used Fourthly, in to the bestthis ofstudy. our Fourthly, knowledge to the this best is theof our first knowledge academic th paperis is the to first take academic stock of paper the status to take of stock Google of the Trends as a usefulstatus of analytical Google Trends tool for as a the useful fashion analytical industry, tool for not the only fashion by summarising industry, not only the latestby summarising examples from the latest examples from the industry, but also by presenting several additional examples of our own. the industry, but also by presenting several additional examples of our own. The remainder of this paper is organised as follows. Section 2 presents a concise review of the statusThe remainder of Google Trends of this as paper an analytical is organised tool in as the follows. fashion Sectionindustry2 and presents provides a concise some insights review on of the statushow of Googleit could be Trends more asuseful an analyticalin future; we tool also in re thefer fashion to the need industry for statistical and provides models someand accurate insights on howforecasting it could be of more fashion useful consumer in future; Google we Trends also referthrough to the examples. need for statisticalSection 3 focuses models entirely andaccurate on forecastingintroducing of fashion the data, consumer whilst Section Google 4 briefly Trends introduces through the the forecasting examples. models. Section Section3 focuses 5 is dedicated entirely on introducingfor the forecast the data, evaluation whilst Section which 4is briefly followed introduces by a discussion the forecasting in Section models. 6. The paper Section concludes5 is dedicated in for theSection forecast 7 by pointing evaluation out whichthe key isfindings followed and bylimitations a discussion of our inresearch. Section 6. The paper concludes in Section7 by pointing out the key findings and limitations of our research. 2. Googling Fashion 2. Googling Fashion 2.1. How is the Fashion Industry Exploiting Google Trends? 2.1. HowBoone is the Fashion(2016) asserts Industry that Exploiting Google GoogleTrends Trends?can also show which styles really catch on with shoppers when search patterns and geographic factors driving the biggest fashion trends in 2016 Boone(2016) asserts that Google Trends can also show which styles really catch on with shoppers were analysed. For example, Figure 2 below shows Google Trends for bomber jackets which grew when297% search YoY patterns in the UK and and geographic 612% YoY factors in the US driving and illustrated the biggest the fashion shift towards trends genderless in 2016 were fashion analysed. For example,(Boone 2016). Figure In2 addition,below shows fashion Google companies Trends can for tr bomberack Google jackets searches which to grew identify 297% purchasing YoY in the UK and 612%decisions YoY and in theswiftly US andmeet illustrated that demand the (Hastreite shift towardsr 2016). genderless Given the fashionincreasing (Boone competition 2016). within In addition, fashionthe industry, companies such can analytics track Google can give searches any brand to or identify retailer purchasinga clear competitive decisions advantage and swiftly in terms meet of that demandmore ( Hastreiterefficient resource 2016). alloca Giventions the and increasing minimising competition waste. within the industry, such analytics can give any brand or retailer a clear competitive advantage in terms of more efficient resource allocations and minimising waste.

FigureFigure 2. Google2. Google search search trends trends forfor bomber jackets jackets inin UK UK and and US US(Boone (Boone 2016). 2016 ).

ThereThere is also is also evidence evidence of of some some fashionfashion brands looking looking into into exploiting exploiting Google Google Trends Trends to their to their advantageadvantage for enhancingfor enhancing consumer consumer engagement engagement via via improved improved online online consumer consumer experiences.experiences. One such examplesuch isexample Miinto, is aMiinto, leading a leading Norwegian Norwegian fashion fashion retailer retailer that that created created a fun a fun nostalgia-based nostalgia-based quiz quiz using Googleusing Trends Google to Trends enable to consumers enable consumers to be fashion to be fashion forward forward (Armstrong (Armstrong 2016). 2016). Google Google and and Zalando Zalando teamed up to create Project Muze to exploit the data in Google’s Fashion Trends report using teamed up to create Project Muze to exploit the data in Google’s Fashion Trends report using Neural Neural Networks for creating designs based on users’ interests (Think with Google 2017). Whilst Soc. Sci. 2019, 8, 111 5 of 23

Networks for creating designs based on users’ interests (Think with Google 2017). Whilst some have criticised the process (see for example, Perez(2016)), there is potential for further research and development to improve the outputs from such novel use of big data analytics. The lack of academic research into Google Trends as a viable analytical tool for the fashion industry is indeed worrying for an industry which is valued at 3 trillion USD and 2% of the world’s GDP (Fashion United 2019). The relative neglect of Google Trends in the context of the fashion industry is surprising since research using Google Trends has increased dramatically in the last decade with lucrative applications in IT, communications, medicine, health, business and economics (Jun et al. 2018). It is our hope that the topics covered in this paper would motivate more academic research into fashion consumer Google Trends and its value for decision-making.

2.2. How Can Google Trends Help Strategic Fashion Management & Marketing Decisions? In addition to presenting selected examples of how the fashion industry can benefit from analytics based on fashion consumer Google Trends for making more productive and strategic fashion management and marketing decisions, we also use this subsection to highlight the importance and utility of being able to forecast fashion consumer Google Trends accurately into the future.

2.2.1. Identifying Seasonal Patterns in Demand The see-now, buy-now trend was expected to flatten out the seasonal peaks in the fashion industry via cruise collections and trans-seasonality. For example, in 2016, the luxury fashion house Burberry launched its entirely straight-to-consumer collection as it aimed to create a seasonless fashion calendar, breaking all the rules at the London (Brown 2016; Rodulfo 2016; Cusick 2016). However, evidence suggests that fashion remains highly seasonal (for example, the most recent retail sales patterns in UK, as visible via the Office for National too supports this view). This could partly be due to the see-now, buy-now trend failing to become an industry standard as some early adopters such as Tom Ford found it backfired, whilst the Kering Group has so far resisted the model (Reuters 2018). Smith(2018) found, from Edited asserts, that there is growing need to understand the shifting seasonality of apparel. Whilst there is a plethora of evidence in academia indicating the importance of seasonality in fashion demand (Nenni et al. 2013), Thomassey(2014) correctly notes that not all fashion products have a seasonal demand (as continuity lines have year-round demand). Nonetheless, a considerable number of items are impacted by seasonality and thus, seasonality should beSustainability incorporated 2019, into11, x FOR fashion PEERforecasting REVIEW systems. 6 of 24 To this end, fashion consumer Google Trends could indicate consumers’ seasonality in demand From this example, it is evident that the fashion industry could benefit through the application for any given product and clearly show changes in seasonal demand patterns. Figure3 considers of statistical signal processing models which have the capability of extracting seasonal variations in Google web search patterns for ‘bomber jackets’, which not only show that there is a seasonal demand, data and providing insights and seasonal forecasts into the future. Moreover, the data in Figure 3 but also that the seasonal demand patterns have changed considerably over time. For example, there is appears to be nonstationary over time and would therefore benefit from the application of a diminishingnonparametric seasonal models, demand which do from not 2004 assume up untilstationarity. December 2013, which is followed by the seasonal demand shifting upwards prior to illustrating a downward sloping trend as visible.

Figure 3. Google Trends for bomber jackets (January 2004–February 2019) (Data Source: Google Trends, 1 FebruaryFigure 3. 2019).Google Trends for bomber jackets (January 2004–February 2019) (Data Source: Google Trends, 1 February 2019).

2.2.2. Competitor Analysis Gone are the days when luxury brands viewed online as a “mass market” (Deloitte 2018). With e-commerce expected to account for 36% of global fashion retail sales by 2020 (Meena 2018), all brands are focused on getting their digital footprint right. Moreover, recent reports indicate that consumers are increasingly choosing online shopping over the high street (BBC 2019). This preference is understandable as online shopping enables quick and easy price comparisons and ensures that physical journeys are not wasted, as retailers may not always carry full stock of sizes and colours in every store. As consumers demand better shopping experiences, luxury brands are exploiting the web to provide personalised shopping experiences (Deloitte 2018) with global investments in improving luxury online shopping now topping 1 billion USD (Gallhagher 2017). Fashion consumer Google Trends has the potential to help brands assess the success of their online marketing campaigns and enable identification of key competitors through brand search trend analysis (Willner 2017). In addition, analysts can also compare performance based on specific products. For example, Gucci was reported to be the most popular Google search for fashion brands in 2017 with Louis Vuitton coming in at number two (Bobila 2017), this trend appears to be continuing in 2019 (see, Figure 4). Such insights are useful for fashion brands as it shows them the effectiveness of their online marketing campaigns and websites in terms of engaging the consumer. In addition, brands can easily identify who their key competitors are, and this can enable them to study their competitors’ approaches to engaging the consumer online and improve upon their existing approaches. The simple example in Figure 4 also clearly illustrates how, except for Gucci, Louis Vuitton and Chanel, the other luxury brands considered here are struggling to get their online footprint set right. Interestingly, the trends shown here continue in terms of image, news, and YouTube searches online. However, it is important to remember that high search trends are not always a good thing. It is imperative that managers and analysts understand the underlying cause of emerging trends (Bradlow et al. 2017). For example, Figure 5 below shows worldwide fashion consumer Google Trends for the search term “Dolce & Gabbana”. The peak popularity for the fashion brand in November 2018 is far from positive as this coincides with its disastrous marketing campaign featuring a Chinese woman struggling to eat pizza and other Italian foods with chopsticks (Hall and Suen 2018). Therefore, it is advisable that peaks in Google Trends are always cross-checked with research into other information surrounding brands to ensure it is a positive gain as opposed to one which is fuelled by negative media. Soc. Sci. 2019, 8, 111 6 of 23

If a fashion company could find evidence of a significant positive correlation between its historical sales for a product and consumer interest in the said product, as indicated via fashion consumer Google Trends, then it is reasonable to assume that Google Trends could serve as a potential indicator for changes in future sales. It is noteworthy that Boone et al.(2017) found evidence of Google Trends improving sales forecasts. Therefore, it would be useful if fashion brands can accurately forecast future seasonality movements in fashion consumer Google Trends, enabling better decisions on what to stock, when to stock and how much to stock, in addition to using consumer trends to determine when to reduce and increase the price points from a marketing perspective. Hastreiter(2016) supports the view that tracking Google searches could be useful for identifying purchasing decisions and that acting upon such analytics to swiftly meet the growing demand could strengthen top lines through value realisation which enables removal of discount traps and more full price sales. From this example, it is evident that the fashion industry could benefit through the application of statistical signal processing models which have the capability of extracting seasonal variations in data and providing insights and seasonal forecasts into the future. Moreover, the data in Figure3 appears to be nonstationary over time and would therefore benefit from the application of nonparametric models, which do not assume stationarity.

2.2.2. Competitor Analysis Gone are the days when luxury brands viewed online as a “mass market” (Deloitte 2018). With e-commerce expected to account for 36% of global fashion retail sales by 2020 (Meena 2018), all brands are focused on getting their digital footprint right. Moreover, recent reports indicate that consumers are increasingly choosing online shopping over the high street (BBC 2019). This preference is understandable as online shopping enables quick and easy price comparisons and ensures that physical journeys are not wasted, as retailers may not always carry full stock of sizes and colours in every store. As consumers demand better shopping experiences, luxury brands are exploiting the web to provide personalised shopping experiences (Deloitte 2018) with global investments in improving luxury online shopping now topping 1 billion USD (Gallhagher 2017). Fashion consumer Google Trends has the potential to help brands assess the success of their online marketing campaigns and enable identification of key competitors through brand search trend analysis (Willner 2017). In addition, analysts can also compare performance based on specific products. For example, Gucci was reported to be the most popular Google search for fashion brands in 2017 with Louis Vuitton coming in at number two (Bobila 2017), this trend appears to be continuing in 2019 (see, Figure4). Such insights are useful for fashion brands as it shows them the effectiveness of their online marketing campaigns and websites in terms of engaging the consumer. In addition, brands can easily identify who their key competitors are, and this can enable them to study their competitors’ approaches to engaging the consumer online and improve upon their existing approaches. The simple example in Figure4 also clearly illustrates how, except for Gucci, Louis Vuitton and Chanel, the other luxury brands considered here are struggling to get their online footprint set right. Interestingly, the trends shown here continue in terms of image, news, Google Shopping and YouTube searches online. However, it is important to remember that high search trends are not always a good thing. It is imperative that managers and analysts understand the underlying cause of emerging trends (Bradlow et al. 2017). For example, Figure5 below shows worldwide fashion consumer Google Trends for the search term “Dolce & Gabbana”. The peak popularity for the fashion brand in November 2018 is far from positive as this coincides with its disastrous marketing campaign featuring a Chinese woman struggling to eat pizza and other Italian foods with chopsticks (Hall and Suen 2018). Therefore, it is advisable that peaks in Google Trends are always cross-checked with research into other information surrounding brands to ensure it is a positive gain as opposed to one which is fuelled by negative media. Sustainability 2019, 11, x FOR PEER REVIEW 7 of 24 Sustainability 2019, 11, x FOR PEER REVIEW 7 of 24 Such in-depth analysis of fashion consumer Google Trends data is useful for forecasting models lookingSuch at in-depthpredicting analysis the future of fashion of competition consumer based Google on historicalTrends data data. is usefulThis is forbecause forecasting outliers models could lookingbe removed at predicting via time seriesthe future analysis of competition models which based enable on historical denoising; data. this This would is because ensure outliers the forecasts could bebeing removed generated via timeare more series realistic analysis and models accounts which for enablenegative denoising; press. Those this interestedwould ensure in an the example forecasts of beinga denoised generated time are series more whereby realistic signaland accounts processing for ne wasgative used press. to remove Those interested an outlier in are an examplereferred ofto aHassani denoised et al. time (2018, series Figure whereby 5). signal processing was used to remove an outlier are referred to Soc. Sci. 2019, 8, 111 7 of 23 Hassani et al. (2018, Figure 5).

Figure 4. Google Trends for selected luxury fashion brands (January 2004–February 2019) (Data Figure 4. Google Trends for selected luxury fashion brands (January 2004–February 2019) (Data Source: FigureSource: 4.Google Google Trends, Trends 20 forFebruary selected 2019). luxury fashion brands (January 2004–February 2019) (Data Google Trends, 20 February 2019). Source: Google Trends, 20 February 2019).

FigureFigure 5. 5.Google Google Trends Trends for for “Dolce “Dolce & Gabbana”& Gabbana” over over the pastthe 12past months. 12 months. (Data (Data Source: Source: Google Google Trends, 12 March 2019). FigureTrends, 5. 12 Google March 2019).Trends for “Dolce & Gabbana” over the past 12 months. (Data Source: Google SuchTrends, in-depth 12 March analysis 2019). of fashion consumer Google Trends data is useful for forecasting models 2.2.3. Identifying Brand Extension Opportunities looking at predicting the future of competition based on historical data. This is because outliers could 2.2.3. Identifying Brand Extension Opportunities be removedBrand extension, via time series a popular analysis and models fundamental which luxu enablery brand denoising; marketing this would strategy ensure (Eren-Erdogmus the forecasts beinget al. Brand generated2018) refersextension, are to more the a launchingpopular realistic and and of fundamentala accounts new prod foruct luxu negative underry brand an press. existing marketing Those brand interested strategy name. inInvesting(Eren-Erdogmus an example in new of a denoisedetproducts al. 2018) time within refers series theto the wherebyfashion launching industry signal of processing a cannew be prod a costly wasuct usedunder affair. to an removeFashion existing anbrands brand outlier nowname. are have referred Investing the tooption Hassaniin new of etproductsanalysing al.(2018 ,within Figurefashion 5the ).consumer fashion Googleindustry Trends can be for a costlythe pr affair.oducts/sectors Fashion theybrands plan now on haveexpanding the option into, ofto analysingunderstand fashion the consumer consumer demand/sentiment Google Trends for towards the pr oducts/sectorsthe said product they or sector.plan on For expanding example, into, Figure to 2.2.3.understand6 below Identifying shows the consumerfashion Brand Extensionconsumer demand/sentiment OpportunitiesGoogle Trends towards for smart the said watches. product It or is sector.evident For that example, the consumer Figure interest in smart watches is increasing over time and is also becoming more seasonal in demand. Such 6 belowBrand shows extension, fashion a popularconsumer and Google fundamental Trends for luxury smart brand watches. marketing It is evident strategy that (Eren-Erdogmus the consumer information can give confidence to decision makers about investing in new products and show the etinterest al. 2018 in) smart refers watches to the launching is increasing of a over new time product and underis also becoming an existing more brand seasonal name. in Investing demand. in Such new best period for launching said investments in their brands. productsinformation within can thegive fashion confidence industry to decision can be makers a costly about affair. investing Fashion brandsin new products now have and the show option the of best periodSolanki for (2018) launching discusses said the investments importance in of their plus-s brands.ize as a growing fashion sector, and notes how analysing fashion consumer Google Trends for the products/sectors they plan on expanding into, GoogleSolanki Trends (2018) not discussesonly shows the an importance increasing of number plus-size of assearches, a growing but fashionalso that sector, the term and reachednotes how its to understand the consumer demand/sentiment towards the said product or sector. For example, Googlepeak popularity Trends not in onlyUK in shows June 2018.an increasing Another numberexample of is searches, presented but by also Young that (2018) the term with reached regard itsto Figure6 below shows fashion consumer Google Trends for smart watches. It is evident that the peak popularity in UK in June 2018. Another example is presented by Young (2018) with regard to consumer interest in smart watches is increasing over time and is also becoming more seasonal in demand. Such information can give confidence to decision makers about investing in new products and show the best period for launching said investments in their brands. Solanki(2018) discusses the importance of plus-size as a growing fashion sector, and notes how Google Trends not only shows an increasing number of searches, but also that the term reached its peak popularity in UK in June 2018. Another example is presented by Young(2018) with regard to the future of vegan fashion; she discusses Google Trends as a tool which can indicate the future for this Soc.Sustainability Sci. 2019, 201982019, 111,, 1111,, xx FORFOR PEERPEER REVIEWREVIEW 88 ofof8 of 2424 23

thethe futurefuture ofof veganvegan fashion;fashion; sheshe discussesdiscusses GoogleGoogle TrTrends as a tool which can indicate the future for particularthisthis particularparticular sector. sector.sector. Overall, Overall,Overall, it isevident itit isis evidentevident that the thatthat ability thethe abilityability to accurately toto accuratelyaccurately forecast forecastforecast such such trendssuch trendstrends intothe intointo future thethe basedfuturefuture on basedbased historical onon historicalhistorical fashion fashiofashio consumern consumer Google Google Trends Trends data can data be can of utmost be of utmost importance importance for fashion for brandsfashionfashion to brandsbrands develop toto developdevelop better and betterbetter more andand successful moremore successfulsuccessful brand extension brandbrand extensionextension strategies. strategies.strategies.

FigureFigure 6.6.Google Google Trends Trends for for smart smart watches watches (January (Januar 2004–Februaryy 2004–February 2019). 2019). (Data (Data Source: Source: Google Google Trends, 20Trends, February 20 February 2019). 2019). 2.2.4. Identifying Better Marketing Terms 2.2.4. Identifying Better Marketing Terms Fashion marketing and advertising are more complex today than ever before, with consumers Fashion marketing and advertising are more complex today than ever before, with consumers demanding ethical advertising and use of appropriate language (Bae et al. 2015). As Forni(2018) demanding ethical advertising and use of appropriate language (Bae et al. 2015). As Forni (2018) asserts, when marketing fashion online, the failure to understand the need for geographical language asserts, when marketing fashion online,line, thethe failurefailure toto understandunderstand thethe needneed forfor geographicalgeographical languagelanguage changes can be detrimental. Fashion consumer Google Trends can help fashion brands ensure they changes can be detrimental. Fashion consumer Googlele TrendsTrends cancan helphelp fashionfashion brandsbrands ensureensure theythey are using the most appropriate terms online for marketing and search engine optimisation in the are using the most appropriate terms online for marketingrketing andand searchsearch engineengine optimisationoptimisation inin thethe countriescountries inin whichwhich theythey operate.operate. For example, example, Fi Figuregure 77 belowbelow shows the fashion consumer consumer Google Google TrendsTrends for for two two productproduct categoriescategories in two different different markets markets (i.e., (i.e., United United States States and and United United Kingdom). Kingdom). ThisThis exampleexample illustratesillustrates thatthat usingusing the term ‘Men’s Clothing’Clothing’ asas as opposedopposed opposed toto to ‘Menswear’‘Menswear’ ‘Menswear’ onon on aa a brand’sbrand’s brand’s websiteswebsites or or eveneven offlineoffline marketingmarketing campaigns is is lik likelyely to to attract attract more more engagement engagement from from the the consumer consumer ininin thethethe UnitedUnitedUnited States.States.States. On On the the other other hand,hand, itit it alsoalso also showsshows shows thatthat that usingusing using thethe the termterm term ‘Men’s‘Men’s ‘Men’s Clothing’Clothing’ Clothing’ inin in thethe the UnitedUnited Kingdom Kingdom market market will will have have a completely a completely different different effect effect as the as term the ‘Menswear’term ‘Menswear’ is more is popular. more Thepopular. ability The to analyse ability fashionto analyse consumer fashion Google consumer Trends Google by country Trends adds by country more value adds as more it enables value brands as it toenables personalise brands their to personalise websites according their websites to the according consumer to searchthe consumer interests search in any interests given country in any given across thecountry globe. across Once the again, globe. being Once able again, to forecast being these able to fashionto forecastforecast consumer thesethese fashionfashion Google consumer Trends accurately Google Trends into the futureaccurately can helpinto the improve future search can help engine improve optimisation search engine for fashion optimisation brands. for fashion brands.

Figure 7. Cont. Soc. Sci. 2019, 8, 111 9 of 23 Sustainability 2019, 11, x FOR PEER REVIEW 9 of 24

. Figure 7. Google Trends for different product categories in USA and UK (January 2014–March 2019). Figure 7. Google Trends for different product categories in USA and UK (January 2014–March 2019). (Data Source: Google Trends, 24 March 2019). (Data Source: Google Trends, 24 March 2019). 3. Data 3. Data Burberry Google Trends Burberry Google Trends The choice of Burberry as the data set of interest for this study was influenced by several factors. First, BurberryThe choice is of one Burberry of the UK’sas the highestdata set profileof interest exporters, for this butstudy supply was influenced chain disruptions by several resulting factors. fromFirst, Brexit Burberry are expectedis one of tothe add UK’s millions highest in profil extrae tradeexporters, costs but for supply this fashion chain housedisruptions (Rovnick resulting 2019). Accordingly,from Brexit thereare expected is a need to for add more millions data analytics in extra intotrade this costs brand for this to help fashion improve house its (Rovnick profitability 2019). via moreAccordingly, lucrative there resource is a allocations.need for more Secondly, data analytic as evidenceds into this in brand the seasonal to help plot improve in Figure its profitability8, Burberry appearsvia more to belucrative struggling resource in terms allocations. of increasing Secondly, and maintaining as evidenced online in the consumer seasonal interest plot in theFigure brand 8, name.Burberry Thus, appears there is to need be struggling for further in analysis terms of and increasing forecasting and ofmaintaining future online online behavioural consumer trends interest for Burberryin the brand so that name. the management Thus, there can is put need in placefor furt a seriesher analysis of actions and for improvingforecasting itsof online future footprint online andbehavioural be better positionedtrends for toBurberry compete so with that otherthe mana luxurygement fashion can brands. put in Thirdly,place a Burberryseries of actions believes for in theimproving importance its online of digital footprint innovation and be and better sees positioned ‘online’ as to the compete first access with other point luxury to its brand fashion (Burberry brands. Thirdly, Burberry believes in the importance of digital innovation and sees ‘online’ as the first access 2018a). Therefore, Google Trends-based analytics has the potential to help the brand ensure its online point to its brand (Burberry 2018a). Therefore, Google Trends-based analytics has the potential to content remains highly relevant to its consumers. For example, Burberry(2018a) states the brand help the brand ensure its online content remains highly relevant to its consumers. For example, wishes to personalise online homepages, and fashion consumer Google Trends-based analytics can Burberry (2018a) states the brand wishes to personalise online homepages, and fashion consumer help the brand in identifying the most popular product categories and terminology for any given Google Trends-based analytics can help the brand in identifying the most popular product categories market (see Section 2.2.4). Fourthly, as a brand, Burberry has recently demonstrated its willingness to and terminology for any given market (see Section 2.2.4). Fourthly, as a brand, Burberry has recently change for the better and become more sustainable. For example, in July 2018 Burberry was called out demonstrated its willingness to change for the better and become more sustainable. For example, in for burning millions of its products to protect its brand (BBC 2018). Swift to act, by September 2018 July 2018 Burberry was called out for burning millions of its products to protect its brand (BBC 2018). Burberry announced that it will stop the practice of destroying unsaleable products and the use of Swift to act, by September 2018 Burberry announced that it will stop the practice of destroying real fur, and work towards reusing, repairing, donating or recycling their unsaleable goods (Burberry unsaleable products and the use of real fur, and work towards reusing, repairing, donating or 2018b; Hanbury 2018; Bain 2018). Thus, Burberry is a brand to watch, and is a brand that is responsive recycling their unsaleable goods (Burberry 2018b; Hanbury 2018; Bain 2018). Thus, Burberry is a to the consumers’ needs and wants. brand to watch, and is a brand that is responsive to the consumers’ needs and wants. The data considered within the forecast evaluation was extracted through Google Trends and relates to monthly web search history for the search term “Burberry” as recorded worldwide from January 2004–February 2019. The chosen time period is influenced by the availability of data and the fact that longer time horizons can capture the historical trends and changes in seasonal variations visible in time series. Analysing such detailed information should enable forecasting models to generate more accurate parameters for modelling and forecasting future consumer trends. However, it should be noted that depending on the objective of the forecasting exercise, the use of shorter time series can be more useful under certain scenarios. For example, if a fashion company does not wish to rely on a nonparametric forecasting model, then analysing shorter time series can be a better option as longer time series are more likely to be nonstationary over time (Bradlow et al. 2017). Soc. Sci. 2019, 8, 111 10 of 23 Sustainability 2019, 11, x FOR PEER REVIEW 10 of 24

Figure 8. Seasonal plot of “Burberry” Google Trends (January 2012–December 2018). (Data Source: GoogleFigure Trends,8. Seasonal 10 February plot of “Burberry” 2019). Google Trends (January 2012–December 2018). (Data Source: Google Trends, 10 February 2019). Figure9 shows the time series plot for Burberry’s fashion consumer Google Trends. The peak in consumerThe data trends considered has remained within constant the forecast over timeevaluation and occurs was extracted in December through annualy Google coinciding Trends with and Christmas,relates to monthly but interestingly, web search consumer history interest for the insearch the brand termalways “Burberry” declines as recorded during the worldwide start of the from new year.January We 2004–February also plot the seasonal 2019. The changes chosen in time consumer period trendsis influenced over the by last the sevenavailability years of as data an example and the (Figurefact that8). longer This shows time horizons how consumer can capture interest the in hist Burberryorical trends has declined and changes each season in seasonal as recorded variations via Googlevisible Trendsin time (with series. the Analysing exception such of 2018 detailed which isinformation slightly better should than enable 2017). Moreover,forecasting upon models closer to inspection,generate more it is accurate evident parameters that there arefor constantmodelling shifts and forecasting in seasonality future in terms consumer of Burberry’s trends. However, fashion consumerit should be Google notedTrends that depending over time. on For the example, objective the of the occurrence forecasting of troughs exercise, vary the betweenuse of shorter June, time July andseries August can be resulting more useful in varying under amplitudes.certain scenarios. It will For be interestingexample, if toa fashion see how company the forecasting does not models wish performto rely on at acapturing nonparametric these forecasting minute, yet model, very important then analysing changes. shorter time series can be a better option as longerAs visible time series (Figure are9), more the data likely is highlyto be nonstationary seasonal yet fairlyover time stationary (Bradlow (there et al. was 2017). no evidence of seasonalFigure unit 9 rootsshows based the time on the series Osborn–Chui–Smith–Birchenhall plot for Burberry’s fashion consumer (Osborn Google et al. 1988 Trends.) test forThe seasonal peak in unitconsumer roots) overtrends this has time remained period. constant In terms over of the time vertical and axisoccurs in Figure in December9 and the annualy figures coinciding that follow, with the numbersChristmas, shown but interestingly, represent search consumer interest interest relative in to the highestbrand always pointon declines the chart during for the the given startregion of the andnew time. year. A We value also of plot 100 isthe the seasonal peak popularity changes forin co thensumer term. Atrends value ofover 50 meansthe last that seven the years term isas half an asexample popular. (Figure A score 8). of This 0 means shows there how was consumer not enough intere datast forin Burberry this term. has declined each season as recorded via Google Trends (with the exception of 2018 which is slightly better than 2017). Moreover, upon closer inspection, it is evident that there are constant shifts in seasonality in terms of Burberry’s fashion consumer Google Trends over time. For example, the occurrence of troughs vary between June, July and August resulting in varying amplitudes. It will be interesting to see how the forecasting models perform at capturing these minute, yet very important changes. As visible (Figure 9), the data is highly seasonal yet fairly stationary (there was no evidence of seasonal unit roots based on the Osborn–Chui–Smith–Birchenhall (Osborn et al. 1988) test for seasonal unit roots) over this time period. In terms of the vertical axis in Figure 9 and the figures that follow, the numbers shown represent search interest relative to the highest point on the chart for the Sustainability 2019, 11, x FOR PEER REVIEW 11 of 24

given region and time. A value of 100 is the peak popularity for the term. A value of 50 means that Soc. Sci. 2019, 8, 111 11 of 23 the term is half as popular. A score of 0 means there was not enough data for this term.

FigureFigure 9. 9.Google Google TrendsTrends forfor “Burberry”“Burberry” (January(January 2014–February2014–February 2019). (Data (Data Source: Source: Google Google Trends, Trends, (10(10 February February 2019). 2019). 4. Forecasting Models 4. Forecasting Models At the outset, it is noteworthy that the forecasting exercise is recursive (i.e., within each horizon At the outset, it is noteworthy that the forecasting exercise is recursive (i.e., within each horizon of interest, the models re-estimate parameters by remodelling the historical data with each new of interest, the models re-estimate parameters by remodelling the historical data with each new observation that is introduced). This process continues until all required forecasts are computed) and observation that is introduced). This process continues until all required forecasts are computed) and considers horizons of 1, 3, 6 and 12 months ahead. At each horizon, we compute N − h + 1 forecasts, considers horizons of 1, 3, 6 and 12 months ahead. At each horizon, we compute N − h + 1 forecasts, where N = 60, and h refers to the forecasting horizon of interest. The accuracy of the out-of-sample where N = 60, and h refers to the forecasting horizon of interest. The accuracy of the out-of-sample forecasts are evaluated via the Root Mean Squared Error (RMSE) and Ratio of the RMSE (RRMSE) forecasts are evaluated via the Root Mean Squared Error (RMSE) and Ratio of the RMSE (RRMSE) criteria. Those interested in the formulae for the RMSE and RRMSE are referred to Silva et al.(2019) criteria. Those interested in the formulae for the RMSE and RRMSE are referred to Silva et al. (2019) and references therein. In what follows, we provide concise introductions to the forecasting models and references therein. In what follows, we provide concise introductions to the forecasting models used in this paper. used in this paper. 4.1. Autoregressive Integrated (ARIMA) 4.1. Autoregressive Integrated Moving Average (ARIMA) The Box and Jenkins(1970) ARIMA model is one of the most popular and frequently used The Box and Jenkins (1970) ARIMA model is one of the most popular and frequently used parametric time series analysis and forecasting models (Silva et al. 2019). As a parametric model, parametric time series analysis and forecasting models (Silva et al. 2019). As a parametric model, ARIMA is bound by the assumptions of normality and stationarity of the residuals and linearity. ARIMA is bound by the assumptions of normality and stationarity of the residuals and linearity. Whilst such assumptions are unlikely to hold in the real world, for univariate forecast evaluations, it Whilst such assumptions are unlikely to hold in the real world, for univariate forecast evaluations, it is widely accepted that ARIMA should be considered as a benchmark forecasting model (Hyndman is widely accepted that ARIMA should be considered as a benchmark forecasting model (Hyndman 2010). Moreover, ARIMA has previously been adopted in studies looking at forecasting variables of 2010). Moreover, ARIMA has previously been adopted in studies looking at forecasting variables of interestinterest in in the the fashion fashion industry, industry, for for example, example, Wong Wong and and Guo Guo(2010 (2010),), Yu et Yu al. (et2012 al.) (2012) and Liu and et al.Liu(2013 et al.). However,(2013). However, in contrast in tocontrast the aforementioned to the aforementioned papers, here,papers, we here, relyon we an rely automated on an automated and optimised and algorithmoptimised for algorithm ARIMA, for which ARIMA, is popularly which is referred popularl toy as referred ‘auto.arima’, to as `auto.arima’, and is accessible and freelyis accessible via the forecastfreely via package the forecast in R. Thosepackage interested in R. Those in details interested of the in theory details underlying of the theory ‘auto.arima’ underlying are `auto.arima’ referred to Hyndmanare referred and to AthanasopoulosHyndman and Athanasopoulos(2018). (2018). 4.2. Exponential Smoothing (ETS) 4.2. Exponential Smoothing (ETS) In brief, ETS forecasts are weighted averages of past observations, where a higher weight is In brief, ETS forecasts are weighted averages of past observations, where a higher weight is assignedassigned toto thethe moremore recentrecent recordrecord withwith overalloverall exponentiallyexponentially decaying weights (Hyndman (Hyndman and and AthanasopoulosAthanasopoulos 2018 2018).). TheThe developmentdevelopment ofof thethe ETSETS technique is closely associated associated with with the the work work of of Brown(1959), Holt(1957) and Winters(1960), whilst its performance in a fashion context was previously evaluated through the work of Fumi et al.(2013). In this paper, we rely on the nonparametric, automated ETS algorithm found within the forecast package in R. Instead of replicating information Soc. Sci. 2019, 8, 111 12 of 23 surrounding the 30 ETS formulae evaluated as part of the modelling process, we refer those interested in details of the theory underlying ETS to Hyndman and Athanasopoulos(2018).

4.3. Trigonometric Box–Cox ARMA Trend Seasonal Model The TBATS model was developed by De Livera et al.(2011) for handling complex seasonal patterns. Given the earlier discussion around the importance of seasonality in fashion, it is useful to evaluate the performance of this model at forecasting fashion consumer Google Trends. Also noteworthy is that this application is the initial application of TBATS for forecasting in fashion. In brief, TBATS is an exponential smoothing state space model with Box–Cox transformation, ARMA error correction and Trend and Seasonal components, for which the detailed theoretical foundation can be found in Hyndman and Athanasopoulos(2018). Once again, we rely on the automated TBATS algorithm provided via the forecast package in R.

4.4. Neural Network Autoregression (NNAR) Neural networks represent a nonparametric forecasting model made available as an automated algorithm via the forecast package in R. The feed-forward neural network model with one hidden layer is denoted by NNAR (p, P, k)m, where p refers to lagged inputs, P takes a default value of 1 for seasonal data (as is the case here), k refers to nodes in the hidden layer and m refers to a monthly frequency. Those interested in the theory are referred to Hyndman and Athanasopoulos(2018), whilst a discussion around the impact of seasonality on neural network forecasts can be found in Silva et al.(2019). Interestingly, there have been several applications of varying Neural Network models for forecasting in fashion research (Au et al. 2008; Sun et al. 2008; Yu et al. 2011; Xia et al. 2012; Kong et al. 2014).

4.5. Denoised Neural Network Autoregression (DNNAR) The DNNAR model was introduced in Silva et al.(2019) as a solution to the issues identified with the NNAR model when faced with seasonal time series. In brief, the DNNAR model is a nonparametric hybrid model which combines the denoising capabilities of singular spectrum analysis (Sanei and Hassani 2015) with the power of NNAR forecasting to produce superior forecasts. This paper marks the introductory application of the DNNAR model as a useful and viable option for forecasting in the fashion industry. We believe the application of the DNNAR model can be useful in this context because, as Choi and Varian(2012) asserts, the sampling method used to calculate Google Trends varies somewhat from day-to-day, thus creating a sampling error which contributes to additional noise in the data. Given that noise refers to random components which cannot be forecasted, it makes sense to reduce noise levels in fashion consumer Google Trends via the use of a denoising algorithm.

5. Empirical Results In this section, we present the findings from our attempts at forecasting fashion consumer Google Trends using a variety of univariate time series analysis and forecasting models. Table1 below reports the out-of-sample forecasting results from the forecasting exercise. The first observation is that there is no single model that can forecast fashion consumer Google Trends for “Burberry” best across all horizons. We find forecasts from ARIMA outperforming all competing models at h = 1 month-ahead, whilst forecasts from the TBATS model outperforms the competing forecasts at h = 3 months-ahead. In the long run, i.e., h = 6 and 12 months-ahead, we find ETS forecasts to be more accurate than those from ARIMA, TBATS and NNAR models. The NNAR model is seen to be the worst performer at forecasting fashion consumer Google Trends for “Burberry” across all horizons. Figure 10 shows a time series plot of the best and worst performing forecasts at h = 1 month-ahead. In relation to the ARIMA model, the NNAR model fails at forecasting the peaks in search trends accurately (in addition to the problems with forecasting troughs accurately). Therefore, these initial findings indicate that if a fashion company wishes to forecast fashion consumer Google Trends for “Burberry” using univariate models, then they would Sustainability 2019, 11, x FOR PEER REVIEW 13 of 24

The NNAR model is seen to be the worst performer at forecasting fashion consumer Google Trends for “Burberry” across all horizons. Figure 10 shows a time series plot of the best and worst performing forecasts at h = 1 month-ahead. In relation to the ARIMA model, the NNAR model fails Soc.at Sci.forecasting2019, 8, 111 the peaks in search trends accurately (in addition to the problems with forecasting13 of 23 troughs accurately). Therefore, these initial findings indicate that if a fashion company wishes to forecast fashion consumer Google Trends for “Burberry” using univariate models, then they would havehave to to switch switch between between models models depending depending on on the the forecasting forecasting horizon horizon of of interest. interest. From From an an operational operational perspectiveperspective this this is is problematic problematic as as one one would would prefer prefer to to adopt adopt a a single single model model for for forecasting forecasting across across all all horizonshorizons so so that that it it gives gives more more consistency. consistency. As As such, such we, we extend extend the the modelling modelling process process further further in in search search ofof a a univariate univariate model model which which can can provide provide consistently consistently accurate accurate forecasting forecasting results results across across all all horizons. horizons.

TableTable 1. 1.Out-of-sample Out-of-sample forecasting forecasting root root mean mean square square error error (RMSE) (RMSE) results results for for “Burberry” “Burberry” fashion fashion consumerconsumer Google Google Trends. Trends.

HorizonHorizon ARIMA ARIMA ETSETS TBATS TBATS NNAR NNAR 11 2.30 2.30 2.522.52 2.42 2.42 3.08 3.08 3 3 2.762.76 2.902.90 2.752.75 4.13 4.13 6 6 2.99 2.99 2.982.98 3.073.07 4.06 4.06 12 12 3.53 3.53 3.303.30 3.503.50 3.79 3.79 Note: Shown in bold font is the best performing univariate model at each horizon. Note: Shown in bold font is the best performing univariate model at each horizon. Even though the NNAR model was the worst performer for this data, we find it pertinent to give it furtherEven thoughconsideration; the NNAR especially model wasas in the an worstera of performerbig data the for importance this data, we of finddata it mining pertinent techniques to give itsuch further as neural consideration; networks especially for the future as in of an fashion era of big analytics data the should importance not be ignored. of data mining Moreover, techniques existing suchfashion as neural trend networksforecasting for platforms, the future such of fashion as Edited analytics (2019), should also overly not be rely ignored. on neural Moreover, network-based existing fashionmodels trend for its forecasting data analytics. platforms, Accordingly, such as weEdited call( upon2019), a also recently overly published rely on neural hybrid network-based neural network modelsmodel forwhich its datais referred analytics. to as Accordingly, the DNNAR we model call upon(Silva a et recently al. 2019). published hybrid neural network modelThe which first is step referred of the to DNNAR as the DNNAR model involves model ( Silvathe application et al. 2019 ).of SSA for denoising the “Burberry” fashionThe firstconsumer step of Google the DNNAR Trends model series involves and extracti the applicationng signals ofto SSAcreate for a denoising less noisy, the reconstructed “Burberry” fashionseries. consumerThis reconstructed Google Trends series series is then and used extracting as input signals data to createfor generating a less noisy, NNAR reconstructed forecasts—in series. a Thisnutshell reconstructed this summarises series is the then DNNA used asR model. input data Figure for generating11 below shows NNAR the forecasts—in SSA signal extractions. a nutshell this The summarisesextracted components the DNNAR themselves model. Figure will 11 relate below to shows a trend, the SSAperiodic signal components, extractions. Thequasi-periodic extracted componentscomponents themselves and noise (Hassani will relate et to al. a 2016b). trend, periodic The combination components, of these quasi-periodic signals provides components us with and the noisereconstructed, (Hassani etsmoothed al. 2016b ).“Burberry” The combination fashion ofconsumer these signals Google provides Trends usseries with for the forecasting reconstructed, with smoothedNNAR. “Burberry” fashion consumer Google Trends series for forecasting with NNAR.

Figure 10. Out-of-sample forecasts for “Burberry” fashion consumer Google Trends at h = 1 month-ahead from ARIMA and NNAR models. Soc. Sci. 2019, 8, 111 14 of 23 Sustainability 2019, 11, x FOR PEER REVIEW 15 of 24

Figure 11. Figure 11. SingularSingular spectrumspectrum analysisanalysis (SSA)(SSA) signalsignal extractionsextractions forfor “Burberry”“Burberry” fashionfashion consumerconsumer Google Trends. Google Trends. The trend extraction in Figure9 shows the long run behaviour of online consumer trends for Finally, Figure 12 provides a graphical representation of the out-of-sample forecasts at h = 1 “Burberry”. Accordingly, it indicates the necessity for rethinking Burberry’s online marketing strategies month-ahead from the DNNAR and NNAR models. Upon close inspection, it is visible that denoising as following the peak in March 2012, online consumer interest in the brand indicates an ongoing with SSA has enabled the DNNAR model to accurately forecast 4/5 peaks in the out-of-sample data, declining trend over time. The upward sloping trend between 2005 and 2012 is attributable to whilst also providing comparatively better forecasts for 4/5 troughs in relation to the NNAR model. Burberry’s successful initiatives in driving online consumer interest in the brand through launching Forecasting Burberry’s fashion consumer Google Trends with DNNAR can enable the prediction of its UK transactional website in 2006, to the first ever live-streamed Burberry in 2009, to change points in consumer trends more accurately, and strategists can then make use of this launching Burberry.com in 2011 (Burberry 2019). In terms of seasonality, Figure 11 also clearly indicates information to inform their resource allocations and decision-making. For example, more efficient that Burberry’s fashion consumer Google Trends are strongly influenced by seasonal factors. Moreover, decisions on when to start targeting online marketing campaigns to uplift consumer interest in the the seasonality underlying this time series is of varying amplitude depending on the periodicity in brand and vice versa could be made well in advance. The ability to forecast when consumer trends question. These seasonal patterns add further to the difficulty associated with forecasting Burberry’s would rise and fall, and the time period over which a contraction could last, or a recovery could take fashion consumer Google Trends. It is noteworthy that the use of SSA for denoising fashion data also can be of utmost importance for strategic fashion management and marketing decision-making. enables fashion companies to forecast seasonal variations alone to determine how consumer trends can vary over selected periodic cycles. Table2 below reports the out-of-sample forecasting RRMSEs for “Burberry” fashion consumer Google Trends. Here, we consider the DNNAR model as the benchmark and report all results relative to same. As such, where the RRMSE is less than 1, it indicates the DNNAR model is more accurate than a competing model. Moreover, the RRMSE criterion enables us to quantify the accuracy gain further as a percentage, such that it equals 1-RRMSE%. First and foremost, it is evident that the DNNAR model outperforms ARIMA, ETS, TBATS and NNAR forecasts for “Burberry” fashion consumer Google Trends and takes over as the best univariate forecasting model across all horizons. In addition, the majority of the RRMSEs reported here represent statistically significant differences between forecasts (except in the very long run at h = 12 months-ahead). The results in Table2 show that when forecasting at h = 1 month-ahead, DNNAR forecasts are significantly better than forecasts from ARIMA, ETS, TBATS and NNAR models by 28%, 35%, 32% and 47%, respectively. Likewise, at h = 3 months-ahead, the DNNAR forecasts are 29%, 33%, 29% and 53% significantly better than ARIMA, ETS, TBATS and NNAR forecasts, respectively. At h = 6 months-ahead, the DDNAR model outperforms the competing forecasts with statistically significant accuracy gains of 27%, 27%, 29% and 46%, respectively. Finally, at h = 12 months-ahead, the DNNAR model outperforms all competing models but we fail to find evidence of statistically significant

Sustainability 2019, 11, x FOR PEER REVIEW 15 of 24

Soc. Sci. 2019, 8, 111 15 of 23 differences between forecasts, suggesting that these long-term forecasting gains could be a result of chance occurrences.

Table 2. Out-of-sample forecasting RRMSE results for “Burberry” fashion consumer Google Trends.

Horizon ARIMA ETS TBATS NNAR 1 0.72 * 0.65 * 0.68 * 0.53 * 3 0.71 * 0.67 * 0.71 * 0.47 * 6 0.73 * 0.73 * 0.71 * 0.54 *

12 0.80 0.86 0.81 0.75 Note:Figure * indicates 11. Singular the differences spectrum between analysis forecasts (SSA) are signal statistically extractions significant for “Burberry” based on the fashion Hassani-Silva consumer test (HassaniGoogle and Trends. Silva 2015a) at the 10% significance level.

Finally,Finally, Figure Figure 12 12 provides provides a a graphical graphical representation representation of of the the out-of-sample out-of-sample forecasts forecasts at ath h= = 1 1 month-aheadmonth-ahead from from the the DNNAR DNNAR and and NNAR NNAR models. models. Upon Upon close close inspection, inspection, it it is is visible visible that that denoising denoising withwith SSA SSA has has enabled enabled the the DNNAR DNNAR model model to to accurately accurately forecast forecast 4/5 4/5 peaks peaks in in the the out-of-sample out-of-sample data, data, whilstwhilst also also providing providing comparatively comparatively better better forecasts forecasts for for 4/5 4/5 troughstroughs in in relation relation to to the the NNAR NNAR model. model. ForecastingForecasting Burberry’s Burberry’s fashion fashion consumer consumer Google Google Trends Trends with with DNNAR DNNAR can can enable enable the the prediction prediction of of changechange points points in consumerin consumer trends trends more accurately,more accurate andly, strategists and strategists can then makecan then use of make this information use of this toinformation inform their to resource inform their allocations resource and allocations decision-making. and decision-making. For example, moreFor example, efficient more decisions efficient on whendecisions to start on targeting when to onlinestart targeting marketing online campaigns marketing to uplift campaigns consumer to uplift interest consumer in the brand interest and in vice the versabrand could and bevice made versa well could in advance.be made Thewell abilityin advance. to forecast The ability when to consumer forecast trendswhen consumer would rise trends and fall,would and rise the timeand fall, period and over the whichtime period a contraction over which could a contraction last, or a recovery could last, could or take a recovery can be ofcould utmost take importancecan be of utmost for strategic importance fashion for management strategic fash andion marketingmanagement decision-making. and marketing decision-making.

Figure 12. Out-of-sample forecasts for “Burberry” fashion consumer Google Trends at h = 1 month-ahead.

6. Discussion We begin our discussion with a comparison of our findings with those from previous research. As Nenni et al.(2013) note, the importance of seasonality in fashion demand continues to be visible within Burberry’s fashion consumer Google Trends, whilst the seasonal plot in Figure8 further evidences Smith’s (2018) assertion that there is a growing need to understand the shifting seasonality in fashion products. Soc. Sci. 2019, 8, 111 16 of 23

Overall, the forecast evaluation undertaken in this study saw the introduction of several new time series analysis models for forecasting fashion data and this combination of new methods applied to new data has resulted in new research insights as suggested by Bradlow et al.(2017). First and foremost, in line with Jun et al.’s (2018) assertion that the purpose of big data utilisation shifting from monitoring to forecasting, our forecasting evaluation of fashion consumer Google Trends illustrates the possibility of accurately forecasting fashion consumer Google Trends using time series analysis models. Secondly, we evidence how Hassani and Silva’s (2015a) assertion that noise reduction is important for big data forecasting is reasonable and extremely useful in helping generate forecasts, which are more consistent in accuracy across both short and long run horizons. The introductory application of the DNNAR model for fashion forecasting yields similar findings to those reported in Silva et al.(2019) where the authors found the DNNAR model providing significantly better forecasts when faced with highly seasonal data. We also extend the comparison of the DNNAR model’s performance to include the TBATS model in addition to the ARIMA and ETS models which were considered in Silva et al.(2019). The findings here indicate that forecasts from the DNNAR model can outperform those from the TBATS model too. Finally, as this is the first attempt at forecasting fashion consumer Google Trends, we are unable to compare our findings with directly comparable studies. Nevertheless, the superior forecasting performance of the DNNAR model is in line with the findings in Yu et al.(2012), where the authors found that artificial neural network models outperformed ARIMA at forecasting fashion colour trends; Wong and Guo(2010) found neural network forecasts outperformed ARIMA in terms of forecasting fashion retail supply chains. Au et al.(2008) found neural networks to be a good forecasting model for fashion retail data with weak seasonal trends. In contrast, our findings show that the DNNAR model which denoises seasonal data can produce significantly better forecasts for highly seasonal fashion data.

6.1. Are Web Searches for Burberry Predominantly Generated by Online Fashion Consumers who are Looking to Shop? The simple answer is that we do not know for certain. Such analysis would require access to microlevel data around Burberry’s monthly sales (both online and offline) which could then be evaluated in detail via causality tests. However, it is important to remember that fashion brands load millions of Stock Keeping Units (SKUs) relating to different sizes and different colour ways online regardless of whether consumers buy or not. As such, Google Trends can still be useful for brands to identify the consumer attitudes towards its product range, and to identify which SKUs are most popular online (regardless of whether these trends translate into sales or not). The ensuing analytics can ensure brands are providing the options demanded by their consumer and if brands can accurately predict future consumer trends about these SKUs, then they can make more efficient managerial and marketing decisions leading to better resource allocation. Nonetheless, Google Trends also report Google Search patterns from consumers accessing Google Shopping. We analysed “Burberry” web search patterns and Google shopping patterns. Figure 13 shows the time series and scatterplots used at the beginning of this analysis. As both graphs indicated the possibility of a strong correlation, we evaluated for same via Pearson’s correlation and found a statistically significant strong, positive, linear correlation of 72% between web search patterns and Google shopping patterns for the search term “Burberry”. Whilst appreciating that correlation does not imply causation, these findings indicate that a 72% increase in web search patterns can signify a 72% increase in Google shopping patterns related to “Burberry”. This justifies the consideration of web searches as a proxy for online consumer shopping behaviour to a certain extent and appears to be in line with Hastreiter(2016) and Boone et al.(2017) assertions that Google Trends can identify purchase decisions. Moreover, Bloomberg(2017) reported that Google Trends was able to predict a slowdown in Salvatore Ferragamo SpA sales six to nine months before it happened in 2015, and that Prada reports one of the highest correlations between Google web searches and revenue growth. Sustainability 2019, 11, x FOR PEER REVIEW 17 of 24

found a statistically significant strong, positive, linear correlation of 72% between web search patterns and Google shopping patterns for the search term “Burberry”. Whilst appreciating that correlation does not imply causation, these findings indicate that a 72% increase in web search patterns can signify a 72% increase in Google shopping patterns related to “Burberry”. This justifies the consideration of web searches as a proxy for online consumer shopping behaviour to a certain extent and appears to be in line with Hastreiter (2016) and Boone et al. (2017) assertions that Google Trends can identify purchase decisions. Moreover, Bloomberg (2017) reported that Google Trends was able to predict a slowdown in Salvatore Ferragamo SpA sales six to nine months before it happened in 2015, and that Prada reports one of the highest correlations between Google web searches and Soc.revenue Sci. 2019 growth., 8, 111 17 of 23

Figure 13. Time series plot and scatterplot for Burberry fashion consumer Google Trends: web search Figure 13. Time series plot and scatterplot for Burberry fashion consumer Google Trends: web search vs. Google shopping (Data Source: Google Trends, 14 March 2019). vs. Google shopping (Data Source: Google Trends, 14 March 2019). 6.2. Google Trends vs. Trend Forecasting Giants (Edited & WGSN) 6.2. Google Trends vs. Trend Forecasting Giants (Edited & WGSN) To the best of our knowledge, Edited does not exploit Google Trends data to inform its current analyticsTo the and best there of isour no knowledge, publicly available Edited informationdoes not exploit which Google indicates Trends same. data Accordingly, to inform its first current and foremost,analytics inand contrast there is to no the publicly analytics available made possible inform viaation Edited, which Google indicates Trends same. can Accordingly, help users determine first and theforemost, effectiveness in contrast of marketing to the campaignsanalytics made from a possible very broad via sense Edited, (as explained Google Trends in detail can in Sectionhelp users 2.2). Secondly,determine Edited the effectiveness and WGSN of are mark botheting subscription-based campaigns from services a very thatbroa ared sense purely (as focused explained on productin detail listings,in Section and 2.2). therefore Secondly, fail toEdited provide and a WGSN complete are picture, both subscription-based whereas Google Trends services is entirelythat are freepurely to accessfocused and on can product be utilised listings, to analyseand therefore online fail consumer to provide interest a complete on any picture, aspect of whereas fashion. Google Trends is entirelyThirdly, free it isto notaccess possible and can to exportbe utilised raw to data analyse from online Edited consumer (unless you interest pay for on subscription),any aspect of whichfashion. hinders the ability to conduct independent, raw analysis using new methods, and could potentiallyThirdly, hinder it is not the possible discovery to ofexport new raw research data from insights Edited (Bradlow (unless et you al. 2017 pay ).for In subscription), contrast, users which can freelyhinders download the ability the to aggregated conduct independent, data capturing raw onlineanalysis fashion using consumernew methods, Google and Trends could potentially over time. Inhinder addition, the thediscovery founder of of new WGSN research states insights that consumers (Bradlow complain et al. 2017). about everythingIn contrast, looking users can the samefreely today,download but that the itaggregated is inevitable data as thousands capturing ofonline fashion fashion companies consumer are signed Google up Trends for trend over forecasting time. In servicesaddition, and the looking founder at of the WGSN same states colour, that material consum anders silhouettecomplain about forecasts everything (The Fashion looking Law the 2017 same). Thus,today, there but that is the it questionis inevitable of a as decline thousands in creativity of fashion which companies is the backbone are signed of fashion.up for trend It is ourforecasting notion thatservices the inclusion and looking and at consideration the same colour, of fashion material consumer and silhouette Google Trends forecasts can (The add moreFashion variety Law to2017). the eventualThus, there product is the offering.question of a decline in creativity which is the backbone of fashion. It is our notion that Finally,the inclusion Edited and mines consideration the websites of offashion brands consumer and retailers Google around Trends the can globe add and more generates variety trends to the basedeventual on theproduct patterns offering. it identifies within the existing product offering (i.e., what is being stocked and whatFinally, is sold). Edited In contrast, mines Googlethe websites Trends of brands focus on and consumer retailers searcharound interest the globe and and can generates offer insights trends whichbased areon the more patterns directly it identifies comparable within with the consumer existing needsproduct and offering wants. (i.e., For what example, is being Edited stocked predicts and thatwhat in is terms sold). of In colour contrast, ways; Google yellow, Trends green, focus pink andon consumer neon would search be the interest trending and colour can offer options insights for men’swhich clothing are more in directly 2019 (Yau comparable 2019). However, with consumer if the trend needs forecasts and wants. also considered For example, fashion Edited consumer predicts Googlethat in terms Trends, of thencolour it mightways; haveyellow, given green, a different pink and and neon more would relevant be the colour trending way colour forecast options for 2019. for Figuremen’s 14clothing below in shows 2019 that (Yau even 2019). in 2018However, yellow if and the pinktrend were forecasts not popular also considered colours in fashion relation consumer to men’s clothingGoogle Trends, web searches then it whilst might the have fashion given consumer a different Google and more Trends relevant for these colour colours way inforecast 2019 follows for 2019. a similarFigure trend.14 below If Edited shows considered that even correlatingin 2018 yellow their an trendsd pink with were fashion not popular consumer colours Google in Trends relation and to forecasting same into the future, then they would be able to incorporate this new information into their decision models and improve the accuracy of their trend forecasts further. As such, we believe there is scope for the development of more efficient fashion trend forecasts by combining fashion consumer Google Trends data with the trend forecasts from Edited and WGSN. Sustainability 2019, 11, x FOR PEER REVIEW 18 of 24

men’s clothing web searches whilst the fashion consumer Google Trends for these colours in 2019 follows a similar trend. If Edited considered correlating their trends with fashion consumer Google Trends and forecasting same into the future, then they would be able to incorporate this new information into their decision models and improve the accuracy of their trend forecasts further. As such, we believe there is scope for the development of more efficient fashion trend forecasts by Soc.combining Sci. 2019, 8 fashion, 111 consumer Google Trends data with the trend forecasts from Edited and WGSN.18 of 23

Figure 14. Time series plots for fashion consumer Google Trends on Edited’s 2019 trending colour waysFigure for 14. men’s Time clothing series plots (Data for Source: fashion Google consumer Trends, Google 24 March Trends 2019). on Edited’s 2019 trending colour ways for men’s clothing (Data Source: Google Trends, 24 March 2019). 7. Conclusions 7. Conclusions In terms of the uses of Google Trends in fashion, we find evidence of the fashion industry seeking to exploitIn terms Google of the Trends uses of to Google benefit Trends the consumer in fashion, experience. we find evidence However, of the there fashion is little industry evidence seeking (and noneto exploit from anGoogle academic Trends perspective) to benefit the of its consumer exploitation experience. as an analytical However, tool there for betteris little decision-making evidence (and andnone forecasting from an academic in fashion. perspective) We present of its examples exploitation of howas an fashionanalytical consumer tool for better Google decision-making Trends can be usefuland forecasting for fashion in brands fashion. to identifyWe present seasonal examples patterns of how in demand, fashion conduct consumer competitor Google Trends analysis, can brand be extensionuseful for opportunitiesfashion brands and to identify better marketing seasonal patterns terms. in demand, conduct competitor analysis, brand extensionRegarding opportunities the overarching and better aim marketing of this paper terms. to determine the existence of a single univariate forecastingRegarding model the that overarching can predict aim fashion of this consumer paper Googleto determine Trends the accurately existence across of a allsingle horizons, univariate using Burberryforecasting as anmodel example, that can the predict forecast fashion evaluation consumer failed Google to find aTrends single accurately univariate across model all which horizons, could provideusing Burberry the best forecastas an example, across allthe horizons. forecast evaluati The studyon consideredfailed to find both a single parametric univariate and nonparametricmodel which forecasting models such as ARIMA, ETS, TBATS and NNAR. Whilst ARIMA, ETS and TBATS were successful in providing the best forecast at least at one horizon of interest, the NNAR model was the worst performer. The failure of any single model at providing the best forecast across all horizons could be a result of the complex seasonal variations with varying amplitudes underlying the Burberry fashion consumer Google Trends series. Soc. Sci. 2019, 8, 111 19 of 23

Given that existing trend forecasting platforms such as Edited too relies on Neural Networks for its modelling and forecasts (Edited 2019), we were motivated to evaluate the performance of the DNNAR model when applied to a fashion context. Overall, this application resulted in providing overwhelming support indicating the importance of signal extraction and denoising for fashion analytics. This is because the application of the DNNAR model, which includes denoising and signal extraction with SSA prior to forecasting with NNAR, resulted in a model which was successful at providing the best forecast for Burberry’s fashion consumer Google Trends across all horizons with statistically significant outcomes in most cases. Thus, in addition to identifying a hybrid univariate model which can be used by Burberry for forecasting its fashion consumer Google Trends across all horizons, we also show the usefulness of applying denoising and signal extraction techniques for fashion data analytics in an era of big data. Like all research, our study and its findings are not without its limitations. As discussed previously in our paper, it is well documented that using only search information for analytics without complementing it with other sources of big data and news has its limitations (Jun et al. 2018). Thus, brands should always consider using different types of information to support any strategic fashion management or marketing decisions they wish to pursue based on fashion consumer Google Trends. Moreover, in this paper, the DNNAR model is applied for forecasting a highly seasonal example of fashion consumer Google Trends. It is likely that these results will not hold for fashion consumer Google Trends which do not display such high levels of seasonality. Nevertheless, our study opens several research avenues. Firstly, there is a need for more research that can provide conclusive evidence on whether fashion consumer Google Trends can help predict fashion purchases. Such research would require the fashion industry to agree on liberalising its practices when it comes to making data available for research purposes. Secondly, a more thorough forecast evaluation which considers a wider range of univariate and multivariate models at predicting a range of fashion consumer Google Trends for a variety of brands at higher frequencies should be undertaken. Thirdly, it would be interesting to collaborate with Edited and/or WGSN to determine how their trend forecasts can be further improved upon by incorporating fashion consumer Google Trends whilst assessing the existence of any correlation or causality between Google Trends-based fashion forecasts and Edited/WGSN trend forecasts.

Author Contributions: Conceptualization, E.S.S.; Data curation, H.H.; Formal analysis, E.S.S. and H.H.; Resources, L.G.; Supervision, D.Ø.M.; Writing—original draft, D.Ø.M. Funding: This research received no external funding. Acknowledgments: We would like to thank two anonymous referees for many helpful comments. However, any remaining errors are solely ours. Conflicts of Interest: The authors declare no conflicts of interest.

References

Armstrong, Paul. 2016. Why Your Business Should Be Using Google Trends. Available online: https://www.forbes.com/sites/paularmstrongtech/2016/04/01/why-your-business-should-be- using-google-trends/#4495c66d8268 (accessed on 22 July 2018). Au, Kin-Fan, Tsan-Ming Choi, and Yong Yu. 2008. Fashion retail forecasting by evolutionary neural networks. International Journal of Production Economics 114: 615–30. [CrossRef] Bae, Su Yun, Nancy Rudd, and Anil Bilgihan. 2015. Offensive advertising in the fashion industry: Sexual objectification and ethical judgments of consumers. Journal of Global Fashion Marketing 6: 236–49. [CrossRef] Bain, Marc. 2016. Google Already Knows the Fashion Trends that Will Be Popular this Fall. Available online: https://qz.com/767643/google-already-knows-the-fashion-trends-thatll-be-popular-this-fall/ (accessed on 22 July 2018). Bain, Marc. 2018. Burberry Won’t Burn Unsold Goods Anymore, but What about the Brands that didn’t Get Caught? Available online: https://qz.com/quartzy/1381150/burberry-wont-burn-unsold-goods-anymore- but-what-about-everyone-else/ (accessed on 24 March 2019). Soc. Sci. 2019, 8, 111 20 of 23

Bangwayo-Skeete, Prosper F., and Ryan W. Skeete. 2015. Can Google data improve the forecasting performance of tourist arrivals? Mixed-data sampling approach. Tourism Management 46: 454–64. [CrossRef] BBC. 2018. Burberry Burns Bags, Clothes and Perfume Worth Millions. Available online: https://www.bbc.co.uk/ news/business-44885983 (accessed on 11 March 2019). BBC. 2019. Boohoo Remains in Fashion as Sales Surge. Available online: https://www.bbc.co.uk/news/business- 46874367 (accessed on 19 February 2019). Bloomberg. 2017. Google Searches Seen as Luxury Investors’ Must-Have Tool in Fashion. Available online: https://www.businessoffashion.com/articles/news-analysis/google-searches-seen-as-luxury- investors-must-have-tool-in-fashion (accessed on 17 March 2019). Bobila, Maria. 2017. Chinese Retail giant Alibaba Launches ‘Big Data Anti-Counterfeiting Alliance’ with Louis Vuitton, Swarovski and More. Available online: https://fashionista.com/2017/01/alibaba-anti-counterfeit- alliance (accessed on 3 February 2019). Boone, Torrence. 2016. Fashion Trends 2016: Google Data Shows What Shoppers Want. Available online: https://www.thinkwithgoogle.com/advertising-channels/search/fashion-trends-2016-google-data- consumer-insights/ (accessed on 14 February 2019). Boone, Tonya, Ram Ganeshan, Robert L. Hicks, and Nada R. Sanders. 2017. Can Google Trends Improve your Sales Forecast? Production and Operations Management 27: 1770–74. [CrossRef] Box, George Edward Pelham, and Gwilym Jenkins. 1970. Time Series Analysis: Forecasting and Control. San Francisco: Holden-Day. Bradlow, Eric Thomas, Manish Gangwar, Praveen Kopalle, and Sudhir Voleti. 2017. The Role of Big Data and Predictive Analytics in Retailing. Journal of Retailing 93: 79–95. [CrossRef] Broomhead, David S., and Gregory P. King. 1986a. Extracting qualitative dynamics from experimental data. Physica D: Nonlinear Phenomena 20: 217–36. [CrossRef] Broomhead, David S., and Gregory P. King. 1986b. On the qualitative analysis of experimental dynamical systems. In Nonlinear Phenomena and Chaos. Edited by Sarben Sarkar. Bristol: Adam Hilger, pp. 113–44. Brown, Robert Goodell. 1959. Statistical Forecasting for Inventory Control. New York City: McGraw/Hill. Brown, Indya. 2016. Watch Burberry’s First See-Now, Buy-Now Collection Live. Available online: https: //www.thecut.com/2016/09/watch-burberrys-first-see-now-buy-now-collection-live.html (accessed on 11 March 2019). Bulut, Levent. 2018. Google Trends and the forecasting performance of exchange rate models. Journal of Forecasting 37: 303–15. [CrossRef] Burberry. 2018a. Strategic Report. Available online: https://www.burberryplc.com/content/dam/ burberry/corporate/Investors/Results_Reports/2018/Burberry_AnnualReport_FY17-18.pdf (accessed on 11 March 2019). Burberry. 2018b. Burberry Ends Practice of Destroying Unsaleable Products. Available online: https://www.burberryplc.com/en/news-and-media/press-releases/corporate/2018/burberry-ends- practice-of-destroying-unsaleable-products.html (accessed on 11 March 2019). Burberry. 2019. History. Available online: https://www.burberryplc.com/en/company/history.html (accessed on 16 March 2019). Choi, Hyunyoung, and . 2012. Predicting the present with Google Trends. Economic Record 88: 2–9. [CrossRef] Cusick, Ashley. 2016. Burberry Breaks All the Rules at London Fashion Week. Available online: http://www. essentialjournal.co.uk/burberry-breaks-all-the-rules-at-london-fashion-week/ (accessed on 11 March 2019). D’Arpizio, Claudia, and Federica Levato. 2017. The Millennial State of Mind. Available online: https://farfetch. prezly.com/bain--company-media-pack (accessed on 11 March 2019). De Livera, Alysha M., Robin John Hyndman, and Ralph David Snyder. 2011. Forecasting time series with complex seasonal patterns using exponential smoothing. Journal of the American Statistical Association 106: 1513–27. [CrossRef] Deloitte. 2018. Global Powers of Luxury Goods 2018: Shaping the Future of the Luxury Industry. Available online: https://www2.deloitte.com/content/dam/Deloitte/global/Documents/Consumer-Business/cb- global-powers-luxury-goods-2018.pdf (accessed on 11 March 2019). Edited. 2019. More Data. With More Data Science Behind It. Available online: https://edited.com/data/ (accessed on 16 March 2019). Soc. Sci. 2019, 8, 111 21 of 23

Eren-Erdogmus, Irem, Ilker Akgun, and Esin Arda. 2018. Drivers of successful luxury fashion brand extensions: Cases of complement and transfer extensions. Journal of Fashion Marketing and Management: An International Journal 22: 476–93. [CrossRef] Fashion United. 2019. Global Fashion Industry Statistics—International Apparel. Available online: https: //fashionunited.com/global-fashion-industry-statistics/ (accessed on 19 September 2019). Forni, Christine. 2018. Using Google Trends to Get an Edge in Digital Marketing. Available online: https: //growthpilots.com/using-google-trends-edge-digital-marketing/ (accessed on 11 March 2019). Fumi, Andrea, Arianna Pepe, Laura Scarabotti, and Massimiliano M. Schiraldi. 2013. Fourier Analysis for Demand Forecasting in a Fashion Company. International Journal of Engineering Business Management 5: 1–10. [CrossRef] Gallhagher, Kevin. 2017. Luxury Fashion Brands Shift Budgets to Digital. Available online: https://www. businessinsider.com/luxury-fashion-brands-shift-budgets-to-digital-2017-6?r=US&IR=T (accessed on 11 March 2019). Google. 2019. How Trends Data is Adjusted. Available online: https://support.google.com/trends/answer/ 4365533?hl=en (accessed on 11 March 2019). Gordon, Steven. 2017. Google Trends: A Metric of Consumer Behavior. White Paper. Sphere Quantitative Insights. Available online: https://sphereqi.com/wp-content/uploads/2017/09/whitepaper-6-30-17-Using-Google- Trends.pdf?pdf=GoogleTrends-A-Metric-Of-Consumer-Behavior (accessed on 23 March 2019). Guzman, Giselle. 2011. Internet search behavior as an economic forecasting tool: The case of inflation expectations. Journal of Economic and Social Measurement 36: 119–67. [CrossRef] Hall, Casey, and Zoe Suen. 2018. Dolce & Gabbana China Show Cancelled Amid Racism Outcry. Available online: https://www.businessoffashion.com/articles/news-analysis/racism-accusations-force- dolce-gabbana-to-cancel-chinese-fashion-show (accessed on 11 March 2019). Hamid, Alain, and Moritz Heiden. 2015. Forecasting volatility with empirical similarity and Google Trends. Journal of Economic Behavior & Organization 117: 62–81. Hanbury, Mary. 2018. Burberry is Changing Its Policy of Burning Millions of Dollars of Unsold Goods after Backlash Erupts Online. Available online: https://www.businessinsider.com/burberry-responds-to- backlash-stops-burning-unsold-goods-2018-9?r=US&IR=T (accessed on 24 March 2019). Hand, Chris, and Guy Judge. 2011. Searching for the picture: Forecasting UK cinema admissions using Google Trends data. Applied Economics Letters 19: 1051–55. [CrossRef] Hassani, Hossein, and Emmanuel Sirimal Silva. 2015a. A Kolmogorov-Smirnov Based Test for Comparing the Predictive Accuracy of Two Sets of Forecasts. Econometrics 3: 590–609. [CrossRef] Hassani, Hossein, and Emmanuel Sirimal Silva. 2015b. Forecasting with Big Data: A Review. Annals of Data Science 2: 5–19. [CrossRef] Hassani, Hossein, and Emmanuel Sirimal Silva. 2018. Big Data: A big opportunity for the petroleum and petrochemical industry. OPEC Energy Review 42: 74–89. [CrossRef] Hassani, Hossein, Gilbert Saporta, and Emmanuel Sirimal Silva. 2014. Data Mining and Official Statistics: The Past, The Present and The Future. Big Data 2: 34–43. [CrossRef] Hassani, Hossein, Xu Huang, and Emmanuel Sirimal Silva. 2016a. A Review of Data Mining Applications in Crime. Statistical Analysis and Data Mining: The ASA Data Science Journal 9: 139–54. [CrossRef] Hassani, Hossein, Zara Ghodsi, Emmanuel Sirimal Silva, and Saeed Heravi. 2016b. From nature to maths: Improving forecasting performance in subspace-based methods using genetics Colonial Theory. Digital Signal Processing 51: 101–9. [CrossRef] Hassani, Hossein, Xu Huang, and Emmanuel Sirimal Silva. 2018. Banking with blockchain-ed big data. Journal of Management Analytics 5: 256–75. [CrossRef] Hastreiter, Nick. 2016. 4 Ways Big Data Is Going to Revolutionize the Fashion Industry. Available online: https:// www.futureofeverything.io/4-ways-big-data-is-going-to-revolutionize-the-fashion-industry-2/ (accessed on 2 May 2018). Holt, Charles C. 1957. Forecasting Seasonals and Trends by Exponentially Weighted Averages. (O.N.R. Memorandum No. 52). Pittsburgh: Carnegie Institute of Technology. Hyndman, Robin John. 2010. Benchmarks for Forecasting. Available online: https://robjhyndman.com/ hyndsight/benchmarks/ (accessed on 13 March 2019). Soc. Sci. 2019, 8, 111 22 of 23

Hyndman, Robin John, and George Athanasopoulos. 2018. Forecasting: Principles and Practice, 2nd ed. OTexts: Available online: https://otexts.com/fpp2/ (accessed on 19 February 2019). Jahshan, Elias. 2018. UK Fashion Brands Continue to Attract Overseas Online Shoppers. Available online: https://www.retailgazette.co.uk/blog/2018/05/uk-fashion-retailers-continue-catch-eye-overseas- online-shoppers/ (accessed on 22 July 2018). Jun, Seung-Pyo, Hyoung Sun Yoo, and San Choi. 2018. Ten years of research change using Google Trends: From the perspective of big data utilizations and applications. Technological Forecasting and Social Change 130: 69–87. [CrossRef] Kong, Xiaohong, Lixia Chang, and Junpeng Xu. 2014. Fashion colour forecasting based on BP neural network. Computer Modelling & New Technologies 18: 584–91. Laney, Doug. 2001. 3D Data Management: Controlling Data Volume, Velocity, and Variety. META Group Research Note 6: 1. LaValle, Steve, Eric Lesser, Rebecca Shockley, Michael S. Hopkins, and Nina Kruschwitz. 2011. Big Data, Analytics and the Path from Insights to Value. MIT Sloan Management Review 52: 20–31. Liu, Na, Shuyun Ren, Tsan-Ming Choi, Chi-Leung Hui, and Sau-Fun Ng. 2013. Sales Forecasting for Fashion Retailing Service Industry: A Review. Mathematical Problems in Engineering.[CrossRef] Madsen, Dag Øivind. 2016. Using Google Trends in Management Fashion Research: A Short Note. Available online: https://hal.archives-ouvertes.fr/hal-01343880/ (accessed on 2 March 2019). Marr, Bernard. 2015. Why only One of the 5 vs. of Big Data Really Matters. Available online: https://www. ibmbigdatahub.com/blog/why-only-one-5-vs-big-data-really-matters (accessed on 11 March 2019). McDowell, Maghan. 2019. What Google Search can Reveal about Trends. Available online: https://www. voguebusiness.com/technology/big-data-trends-spate-google (accessed on 23 March 2019). Meena, Satish. 2018. Forrester Analytics: Online Fashion Retail Forecast, 2017 To 2022 (Global). Available online: https://www.forrester.com/report/Forrester+Analytics+Online+Fashion+Retail+Forecast+2017+ To+2022+Global/-/E-RES145235 (accessed on 19 February 2019). Nenni, Maria Elena, Luca Giustiniano, and Luca Pirolo. 2013. Demand Forecasting in the Fashion Industry: A Review. International Journal of Engineering Business Management 5. [CrossRef] Net Market Share. 2019. Desktop Search Engine Market Share. Available online: https://goo.gl/ESpbRp (accessed on 23 March 2019). Osborn, Denise R., A. P. L. Chui, Jeremy Smith, and Chris R. Birchenhall. 1988. Seasonality and the order of integration for consumption. Oxford Bulletin of Economics and Statistics 50: 361–77. [CrossRef] Perez, Sarah. 2016. New Project Muze Proves Machines Aren’t that Great at . Available online: https://techcrunch.com/2016/09/02/googles-new-project-muse-proves-machines-arent- that-great-at-fashion-design/ (accessed on 22 July 2018). Reuters. 2018. Fashion Brands Slow to See Now, Buy Now. Available online: https://www.businessoffashion. com/articles/news-analysis/fashion-labels-dither-over-see-now-buy-now (accessed on 11 March 2019). Rodulfo, Kristina. 2016. Burberry Introduces the “Seasonless” Fashion Calendar. Available online: https://www. elle.com/fashion/news/a33859/burberry-cuts-down-four-show-calendar-to-two-shows-a-year/ (accessed on 11 March 2019). Rogers, Simon. 2016. What is Google Trends Data—And What Does it Mean? Available online: https://medium. com/google-news-lab/what-is-google-trends-data-and-what-does-it-mean-b48f07342ee8 (accessed on 11 March 2019). Rovnick, Naomi. 2019. Burberry Warns on Impact of No-Deal Brexit. Available online: https://www.ft.com/ content/073cc80a-1edb-11e9-b126-46fc3ad87c65 (accessed on 11 March 2019). Sanei, Saeid, and Hossein Hassani. 2015. Singular Spectrum Analysis of Biomedical Signals. Boca Raton: CRC Press. Siliverstovs, Boriss, and Daniel S. Wochner. 2018. Google Trends and reality: Do the proportions match?: Appraising the informational value of online search behavior: Evidence from Swiss tourism regions. Journal of Economic Behavious & Organization 145: 1–23. Silva, Emmanuel Sirimal, Hossein Hassani, Saeed Heravi, and Xu Huang. 2019. Forecasting with denoised neural networks. Annals of Tourism Research 74: 134–54. [CrossRef] Smith, Katie. 2018. Understanding the Shifting Seasonality of Apparel. Available online: https://edited.com/ blog/2018/08/shifting-seasonality-apparel/ (accessed on 2 May 2018). Soc. Sci. 2019, 8, 111 23 of 23

Solanki, Hiten. 2018. COMMENT: Is Plus-Size a Growing Sector of Fashion? Available online: https: //www.retailgazette.co.uk/blog/2018/11/comment-plus-size-growing-sector-fashion/ (accessed on 19 February 2019). Soubra, Diya. 2012. The 3Vs that Define Big Data. Available online: https://www.datasciencecentral.com/forum/ topics/the-3vs-that-define-big-data (accessed on 11 March 2019). Sun, Zhan-Li, Tsan-Ming Choi, Kin-Fan Au, and Yong Yu. 2008. Sales forecasting using extreme learning machine with applications in fashion retailing. Decision Support Systems 46: 411–19. [CrossRef] The . 2017. Is Trend Forecasting Doing More Harm to Fashion Than Good? Available online: http://www.thefashionlaw.com/home/is-trend-forecasting-doing-more-harm-to-the-fashion- industry-than-good (accessed on 14 March 2019). Think with Google. 2017. Zalandos Project Muze: Fashion Inspired by You, Designed by Code. Available online: https://www.thinkwithgoogle.com/intl/en-cee/success-stories/global-case-studies/zalandos- project-muze-fashion-inspired-you-designed-code/ (accessed on 22 July 2018). Thomassey, Sébastien. 2014. Sales Forecasting in Apparel and Fashion Industry: A Review. In Intelligent Fashion Forecasting Systems: Models and Applications. Edited by Yong Yu, Tsan-Ming Choi and Chi-Leung Hui. Berlin and Heidelberg: Springer, pp. 9–27. Varian, Hal. 2014. Big data: New tricks for econometrics. Journal of Economic Perspectives 28: 3–28. [CrossRef] Verhoef, Peter C., Scott A. Neslin, and Björn Vroomen. 2007. Multichannel Customer Management: Understanding the Research-shopper Phenomenon. International Journal of Research in Marketing 24: 129–48. [CrossRef] Wedel, Michel, and P. K. Kannan. 2016. Marketing analytics for data-rich environments. Journal of Marketing 80: 97–121. [CrossRef] Willner, Richard. 2017. Competitor Analysis: The Essential Guide for Marketing Managers. Available online: https://www.further.co.uk/blog/competitor-analysis-essential-guide/ (accessed on 11 March 2019). Winters, Peter R. 1960. Forecasting sales by exponentially weighted moving averages. Management Science 6: 324–42. [CrossRef] Wong, W. K., and Z. X. Guo. 2010. A hybrid intelligent model for medium-term sales forecasting in fashion retail supply chains using extreme learning machine and harmony search algorithm. International Journal of Production Economics 128: 614–24. [CrossRef] Wu, Lynn, and Erik Brynjolfsson. 2015. The future of prediction: How Google searches foreshadow housing prices and sales. In Economic Analysis of the Digital Economy. Chicago: University of Chicago Press, pp. 89–118. Xia, Min, Yingchao Zhang, Liguo Weng, and Xiaoling Ye. 2012. Fashion retailing forecasting based on extreme learning machine with adaptive metrics of inputs. Knowledge-Based Systems 36: 253–59. [CrossRef] Yau, Charlotte. 2019. 2019 key Trends: See Them First Right Here. Available online: https://edited.com/blog/ 2018/12/2019-key-trends/ (accessed on 24 March 2019). Young, Elewisa. 2018. What Does the Future Hold for Vegan Fashion? Available online: https://www. plantbasednews.org/post/future-hold-vegan-fashion (accessed on 19 February 2019). Yu, Yong, Tsan-Ming Choi, and Chi-Leung Hui. 2011. An intelligent fast sales forecasting model for fashion products. Expert Systems with Applications 38: 7373–79. [CrossRef] Yu, Yong, Chi-Leung Hui, and Tsan-Ming Choi. 2012. An empirical study of intelligent expert systems on forecasting of fashion color trend. Expert Systems with Applications 39: 4383–89. [CrossRef] Yu, Lean, Yaqing Zhao, Ling Tang, and Zebin Yang. 2019. Online big data-driven oil consumption forecasting with Google trends. International Journal of Forecasting 35: 213–23. [CrossRef] Zhang, Zuochao, Yongjie Zhang, Dehua Shen, and Wei Zhang. 2018. The cross-correlations between online sentiment proxies: Evidence from Google Trends and Twitter. Physica A: Statistical Mechanics and Its Applications 508: 67–75. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).