Article FORECASTING

– should you outsource it or keep it in-house?

By Nona Fortian Corts A combination of challenges, complexities and resources [email protected] and Simon Mikkelsen, associated with forecasting motivates many businesses today to [email protected], consider whether sales forecasting should be kept in-house or Implement Consulting Group outsourced to an external service provider. Here, we will present our view on the question, and which forecast methodologies we believe to fit most current organisations the best.

Why forecasting is important? capital as well as decreased of In modern supply chains, forecasting potential lost sales, thus reduced cost plays a key role in operations by driving of obsolescence and depreciations. strategic and decisions and enabling critical business activities. Nowadays, complex supply chains Sales forecasting supports financial combined with a fast-changing business planning, raw material sourcing as well environment make forecasting a difficult as production planning and inventory task. We see businesses facing different . Having an accurate sales challenges when it comes to the sales forecast allows you to achieve a higher forecasting process such as lack of delivery performance and service level, quantitative methods or lack of tools to leading to higher customer satisfaction. analyse historical trends and calculate In addition, with accurate forecasts statistical baselines. Also, a high number you can enable optimised stock levels, of references can represent a challenge, reducing inventory costs and net working hindering the manual review as well as

implementconsultinggroup.com 1 Sales forecasting – should you outsource it or keep it in-house?

the lack of simple methods to include Based on our extensive experience bottom-up input from sales or . helping organisations design and improve Forecasting is oftentimes seen as not demand planning processes, we have only complex but also expensive, as seen that, in most cases, a simple fore- organisations utilise valuable time and casting process managed in-house can resources in this process. outperform fancy advanced . Simple forecasting consists of combining The combination of challenges, simple statistical forecasting with market complexities and resources associated insights in a simple and efficient manner. with forecasting motivates many businesses today to consider the following question: Our view on outsourcing forecasting capabilities Over the past years, it has become SHOULD FORECASTING increasingly more popular to outsource non-core business processes to external BE KEPT IN-HOUSE OR providers. Organisations opting for OUTSOURCED TO outsourcing strategies are often aiming to unlock new technological capabilities, THIRD-PARTY PROVIDERS? gaining access to new skills and, in many occasions, also cutting costs. Outsourcing In addition to the question of whether sales forecasting is not an exception. By forecasting should take place in-house outsourcing the forecasting processes, or outside the company’s boundaries, we organisations can free up time of the also see a recurring question concerning demand planners and get access to the dilemma between simple vs advanced experts in the forecasting field as well statistical forecasting methods. Typically, as cutting-edge forecasting techniques. external forecasting providers use advanced algorithms, However, there is a downside to and AI methods, but are these in fact outsourcing forecasting. It can be bringing real value to the business? argued that the quality of the forecasting process as a whole may be affected, as the decision-making that takes place in-house will be based on a forecasting Outsourcing sales forecasting result that has been created outside, thus lacking the understanding of how PRO CON the forecast result has been obtained. In other words, outsourcing forecasting can • Access to the • “Black box” be seen as a “black box” in which you most advanced • Loss of input historical sales , and you get forecasting and in forecasting statistical methods process an output which represents your • Less internal manpower • Uninformed decision-making sales, but what happens in-between is needed • Advanced statistical methods completely unknown. Needless to say, • Access to outside experts in are good at predicting the past “black box” solutions may not always forecasting techniques but not the future please everyone. • Difficult to enrich forecast with market intelligence Another consequence of outsourcing • Organisational fragmentation forecasting with a “black box” setup is • Loss of demand planning experts the difficulty of enriching the outsourced • Transfer of sensitive information forecast baseline. That is, if you are a outside the organisation company that has outsourced the generation of the forecast, you cannot

implementconsultinggroup.com 2 Sales forecasting – should you outsource it or keep it in-house?

know how much information, such example, Marketing, Sales and Finance as promotions, capacity constraints, will no longer talk to Operations in the sales strategies etc., has already been same way as if a collaborative demand incorporated in the calculation of the forecasting process would take place resulting baseline. Therefore, enriching in-house. We have seen that when fore- forecasts with market intelligence can casting is outsourced, each organisational become a big challenge for demand area does its own for future planners and could even worsen the sales, that is, Finance will do a value quality of the final forecast. forecast, and Operations will do a volume forecast. We all agree that the quality of the forecast is the key driver when When outsourcing forecasting, there considering different forecasting is not only a risk of losing control of strategies. Generally, companies business processes, but there is also providing forecasting services use a risk of losing people with specific very advanced and quite complex knowledge and expertise in demand forecasting algorithms. Even though planning. Outsourcing will naturally lead methods like machine learning or to a reduction of the number of demand advanced statistical models may sound planners needed, which poses a risk of interesting, scientific literature shows losing valuable knowledge which is not that using complex algorithms does easy to reacquire. not create better results than simple techniques. Advanced forecasting Finally, the last point that needs to be algorithms have shown to be good at considered here is the willingness to predicting the past, but what about transfer critical information outside the future? Are these methods able to the organisation. Security concerns capture the impact in the forecast of may arise when considering outsourcing future events such as capacity issues, forecasting, as some of the inputs increased sales, promotions or changes required for generating a demand in market conditions? The answer is, forecast consist of sensitive information possibly, no. These types of events are such as product launches, future trends extremely difficult to predict simply by or expansion strategies. using advanced history-based statistical models. Spyros Makridakis, one of the world’s leading experts on forecasting, Keep it simple – keep it in-house argues that complex statistical models At this point you may be wondering – why fit past data well but do not necessarily not keep forecasting activities in-house? predict the future accurately, whereas simple models do not necessarily fit past We see many arguments to decide to data well but predict the future better. In keep forecasting capabilities within short, no advanced statistical models will organisations. be accurate because the future, in fact, cannot be written. First, by maintaining the planning process READ MORE in-house, not only do we ensure a higher about simple forecasting Keeping supply chain parts and degree of control of operations and in Spyros Makridakis’ partners working together is already a processes, but also a higher degree of book: Forecasting: big challenge in modern organisations, flexibility and reactivity. Imagine, for Methods and Applications and outsourcing forecasting may not example, that your company is facing influence in the desired direction. In an unforeseen event such as a supply particular, outsourcing may lead to a loss shortage that will have an important of control in the forecasting process, impact on the forecast in the future which can result in a loss of control in months. In an outsourcing scenario, the overall planning processes as well you will need to wait until the next as organisational fragmentation. For forecasting cycle to incorporate this

implementconsultinggroup.com 3 Sales forecasting – should you outsource it or keep it in-house?

Keeping in-house sales forecasting

PRO CON

• A higher degree • Need for time of control of and resources operations and • Need for a robust processes forecasting process • More flexibility and reactivity • Relevant company-specific knowledge by the demand planners • Fully informed decision- making • More visibility and transparency across the supply chain • Simple and easy to understand

market knowledge in the forecast to predict your future demand. Simple calculation. However, in the case where statistical forecasting methods are, at forecasting is done in-house, you could least for repetitive decisions, typically exceptionally recalculate your forecast superior to expert judgements. They are to see the impact on your baseline right also easy to understand and are therefore away and adjust your supply chain easier to add input to. Thus, proactive and operations to mitigate it. robust sales forecasts can be created by combining simple and easy-to-under- Second, forecasting is also about stand statistical forecasting methods incorporating market knowledge to with the market knowledge of the people your demand prediction, and who, if not in the organisation. With simple statistical your internal demand planning team, methods, higher stability can be achieved, can do this task better? Over time, enabling better decision-making. We demand planners have acquired relevant have seen that with outsourcing, forecast company- specific knowledge that should results vary significantly from one release not be ignored. Having the knowledge of to the next, making it difficult to make why the forecast values look like they do consistent decisions over time. However, enables fully informed decision-making, when forecasting is done in-house using increasing the degree of visibility and simple methods and applying simple transparency across the supply chain criteria and rules, we can achieve a more network. stable forecast signal, thus allowing a superior decision process. Third, in-house forecasting means being in control of the company’s intellectual Simple forecasting brings robust property, as sensitive data is not shared with external providers. In addition, the accuracy and control to your process becomes less complex because forecasting process no data loads are required, which we have Based on our experience, we only need observed in some companies to be quite a few simple elements to get a solid tedious to manage. demand forecast.

And last but not least, forecasting can Simple forecasting consists of generating be kept in-house and done in a simple a statistical baseline with a simple way! There is no need for fancy algorithms statistical method, such as the single

implementconsultinggroup.com 4 Sales forecasting – should you outsource it or keep it in-house?

Simple forecasting approach

INPUT OUTPUT

Historical sales Forecast Statistical Adjust with Track and KPIs Segmentation baseline market insights performance principles

, and adjusting it That is, according to the different pre- based on seasonality, trend and market defined simple segment rules, the input intelligence from Sales and Marketing. givers provide market insights into the In essence, simple forecasting is about forecast focusing on impact. For example, leveraging the benefits from statistical a high-runner item with a quite stable forecasting and the human judgement sales pattern will be easier to predict with combined in a simple and efficient statistical methods and will require less structure. This can be done through attention than a high runner with more segmentation principles, making it easy volatility in sales. for demand planners and other input givers to focus their effort in the fore- Finally, the last step in the simple casting process where it matters the forecasting approach is to monitor the most. forecast performance with the use of standard reports and KPIs. These analysis A typical simple forecasting approach tools will help planners identify potential that we recommend for the monthly issues in specific products or for specific demand planning process consists of time periods, focusing their efforts three simple steps. even more to achieve higher forecast performance. These tools can also help Historical sales data is used to calculate monitor the forecast accuracy and drive the forecast baseline generated with a future behaviour in the demand planning simple statistical method such as the process. single exponential smoothing technique and applying seasonality and trends on All in all, the most accurate forecast top. with the lowest bias and the highest stability is therefore a combination of very Historical sales data is also used to define simple and understandable statistical a segmentation of the product portfolio forecasting, market intelligence, including based on their planning characteristics opportunities and based on simple (typically, impact and of rules that enable incorporation of market sales). The resulting clusters or segments insights into the sales forecast. are used to focus the efforts on enriching the forecast by demand planners (in step 2) as well as enabling a common language Conclusion for communicating across supply chain To conclude this reflection and answer operations. to the initial question posed: “Should forecasting be kept in-house or The output of the statistical forecasting outsourced to third-party providers?”, generation is then reviewed and adjusted our answer is clear: Keep forecasting supported by the segmentation principles. in-house and do it in a simple manner!

implementconsultinggroup.com 5 Sales forecasting is not about complex algorithms

We have seen how enriching an On the other hand, if you lean towards outsourced forecast coming from a using advanced statistical methods with “black box” calculation with market machine learning or AI, these “black box” intelligence is incredibly challenging. calculations need to be owned in-house to Therefore, a simple approach is needed. ensure a better understanding and enable There is scientific evidence that the best dialogue supporting the decision-making forecasting models are simple – and if process. you simplify your forecast process, there is no reason for outsourcing it. Good luck with your demand planning improvement journey!

Case example:

How did LEO Pharma improve the data, it was proven that a simple As LEO Pharma’s regional supply forecasting process by keeping it forecasting approach would chain manager for Europe says: in-house and utilising the simple have provided a more accurate forecasting approach? result than the historical manually ”I see the implementation of generated forecasts. The simple statistical forecasting in LEO Pharma LEO Pharma, a Danish leading forecasting approach was a com- as a key part of the overall journey to company in the pharmaceutical bination of segmentation concepts improve forecasting processes and sector with sales in more than with simple statistical methods accuracy. The transition has been a 100 countries worldwide, was in a and was supported by the tools very good experience for us.” situation where demand planning and systems already in place. was in focus, seeking to improve their sales forecasting process. The conclusion resulting from this In this context, two very different assessment was clear: LEO Pharma approaches were considered; should keep their forecasting outsourcing forecasting capabilities in-house, utilising the already LEO Pharma is a pharmaceutical to a third-party provider or imple- existing resources and expertise, company with a global team of 6,000 menting an improved statistical and simple forecasting should be people serving 92 million patients forecasting process in-house. implemented. An implementation in 130 countries. Headquartered in approach was designed and put Denmark, LEO Pharma specialises A project took place supported by a in place in a short time. The tran- in pharmaceutical products for joint team of LEO Pharma demand sition towards simple statistical dermatology and thrombosis. For planning experts and Implement forecasting is in the process of more than 100 years, LEO Pharma consultants aiming at defining being rolled out, and LEO Pharma has been using patient insights to which forecasting strategy LEO is already starting to benefit from research and develop new, safe and Pharma should pursue. Based on the first improvements in forecast effective medicines to treat skin an extensive analysis consisting quality and stakeholder satisfaction. diseases. of based on actual

FAST FACTS ABOUT IMPLEMENT

Founded: 1996 Number of employees: 900 Offices: Copenhagen, Aarhus, Stockholm, Malmo, Oslo, Zurich and Munich implementconsultinggroup.com

6