WHITE PAPER

INSIGHTS DRIVEN COMPUTING REALIZING STRATEGY

Paper discusses the methods and techniques to leverage the web analytics based services to gain insights into customer behavior to enhance user’s online experience by providing a personalized and customized experience. Enhanced user experience would translate into increased business revenues. The paper also discusses a framework to realize the analytics seamlessly from the business team without the involvement of implementation team. Finally the relevance and utility of Web analytics for the emerging markets is discussed. Introduction most popular ways to understand the user behavior on Web was done in one of the Traditional online platforms were a following ways: Sample Business Required Analytics one-way traffic wherein the key design Scenario • Market research elements including the main user interface, How e ective is my online Track clicks on campaign and conversion navigation, flow etc. were designed by the • Beta testing the web site and collecting seasonal campaign? ratio

Track site usability metrics including organization. The end-user had little or no the feedback How useful is the existing site trac, visitor prole, exit rate, online platform? visits duration. say in that. With the emergence of web 2.0, • Perform usability testing from the web is now more user-focused than How can we improve Track the recently viewed products and specialized third-party agencies. the product user interests to provide personalized ever. The complete online experience is recommendation? product recommendations However all these methods lacked one shifting towards the intuitiveness which is important factor: real-time end-user Table 1 : Web analytics sample scenarios likely to engage the end-users and keep feedback and analysis. Some of the them interested an increase the likeliness e-commerce sites need to analyze user Understand customers by their clicks : of their participation. behavior in real-time to promote and Many a times customers do not explicitly User centered experience is changing the recommend their products. They also provide feedback/rate their online way information is presented. It is in the need this information to understand any experience. However we can get a rough interest of business that the presentation seasonal patterns, emerging trends etc. so idea of their overall satisfaction by is designed which appeals to the end that they can customize their selling based “connecting the clicks” user did on the site user and the information is presented around it. and other key metrics like time spent on in its most intuitive fashion. The same site, conversion ratio etc. Web analytics In scenarios such as these, the business thing applies for navigation design as play a vital role in making sense of these needed an analytics system which well. Analytics is a broad area which deals user clicks and metrics to uncover any continuously monitored user behavior with capturing and understanding user’s hidden trends, providing customer insights and provide real-time feedback in intuitive explicit and implicit actions and trying to and to make a sense of overall online reports so that the business heads can make sense out of which. This in turn will strategy. make sense out of their customer behavior be used to optimize the business operates. and prepare their action plan around The importance of metrics also varies Web analytics mainly deals with collecting, it. It would also help them measure the across industries. For instance following measuring, analyzing, and reporting the effectiveness of their web site UI, online metrics would assume importance in data on presentation components. campaign etc. e-commerce industry [9][10] : In this paper, I introduce the key business Measuring analytics of Web is a two-way • Clickthrough rate process steps in designing a web analytics traffic: strategy for an online platform. In addition • Traffic to that, I have also narrated an intuitive • Business wants to measure and • Revenue per visit way of realizing a web analytics framework improve their intended and promoted • Bounce Rate from business team in real-time without online strategies • Page view the need of implementation or operations • Business also wants to understand team. The paper is structured as follows: the patterns emerging from their • Value per order It provides the general advantages and customers’ behavior • Repeat customer frequency relevance of web analytics in section II. Web analytics play a key role in analyzing Section III discusses the key business • Conversion rate both the intended and emerging topics in process steps required for implementing a real time. Following are some of the key web analytics strategy. Section IV discusses benefits by successfully implementing a about steps in the implementation of web web analytics strategy: analytics and Details of realizing a web analytics framework is discussed in section • Visualize the key site data (traffic, time V. Section VI provides the performance spent, exit rate etc.) in intuitive fashion. management details and Section VII • Track and measure the effectiveness of provides the relevancy of web analytics online campaigns for emerging markets and section VI • Derive patterns, emerging trends and summarizes the paper. gain valuable customer insights. Following table provides web analytics Analyzing Web usage for few typical web scenarios: Prior to the arrival of web analytics the

External Document © 2018 Infosys Limited External Document © 2018 Infosys Limited Analytics Driven Online key metrics and monitored. In addition, main business goals. This set would Business the business also needs to set-up generally include all key features of the appropriate real-time notification . “If you cannot measure it, you can’t and exception handling process to • Step-2 Mark and measure the improve it” holds true for managing and manage and improve these metrics. categories, flows and transactions: improving online business. Hence the For instance if the page view metrics Once the boundaries of the online primary step for a business is to layout the drop below a pre-configured threshold, business transaction is established criteria for measuring their online platform. the notification mechanism should through key categories, flows and I am providing below two broad strategies alert the relevant business team transactions they can be marked. This which follow the “measure-and-improve” to investigate this further and take is done by page tagging using a web model. appropriate action. Some of the critical analytics framework. The page tagging metrics like page load times; availability will add client-side JS event handlers A. Measure effectiveness of needs immediate attention to avoid which triggered when the event occurs. intended strategy risking customers. The event could be as small as clicking In this strategy the business wants to • Monitoring process a link to as large as completing a full • Noti cation Process Measure, measure the effectiveness of their intended • DR process Manage & Maintain order processing flow.

online strategy. A three step process in • Site usability metrics • Conversion metrics • Step-3 Identify patterns: This is the implementing this strategy is given below: • Items per order Identify Key Metrics • Page views complex step in the overall process. • Step-1 Establishing business goals: • Increase Information The business analysts have to monitor discoverability Establish Business • Maximize revenue Identify the key business success per visit Goals the graphs regularly and identify any factors for the online business. For patterns that emerge. For instance if instance a manufacturing website Figure 1 : Hierarchy of steps in a particular type of clothing product is designed with the key intent implementing web analytics intended outsells all other types in summer strategy of enabling “easy information season, business needs to perform discoverability” for its customer; B. Identify and act on emerging detailed analysis of the product an e-commerce site is designed to trends/patterns strategy attributes, seasonality factor to offer “maximize revenue per customer” more attractive packages. Similarly if It is not enough that business monitor and Step-2 Identification of metrics the bounce rate has increased after a • manage their intended business strategies; for measuring business goals: The site redesign it needs to be analyzed more often than not they need to watch business needs to identity the metrics if any of the UI elements caused this out for emerging and hidden trends from which will measure the business goals issue. their customers’ behavior. In the absence identified in step – 1. Below mentioned of such strategy, business will not be able It needs deep business and customer table provides an indicative mapping to react to seasonal trends and patterns. understanding to derive the patterns from of goals – metrics: Implementing this strategy is easier said customers’ behavior. Often a combination of multiple factors can contribute to Business Goal Measuring metrics than done as it is often difficult to arrive at establish the correct pattern. For instance, Improve information Site Usability metrics (page views, time on emerging patterns. Following are key steps discoverability site, click paths, exit rate, bounce rates, tra c source) in implementing this strategy: if there is an increase in exit rate/bounce Conversion statistics (site tra c, new visitor, returning visitors) rate in the order page there could be • Step-1 Identify the categories Increase online revenue Metrics for measuring cross-sell and various reasons. However when we per customer up-sell (revenue per visit, repeat purchase and flows: Identify the category rate, items per order, conversion rate). discover that the page load time of order duration. of consumers, transactions and page has increased off-late it helps us to flows which form the major chunk Table 2 : Indicative business goal-metrics conclude the cause-effect analysis. of business. For instance a banking mapping site can have all transactions which It is always recommended that organizations adopt both the strategies • Step – 3 Measure and manage a customer can perform ranked by to get sound customer insights and metrics: Once the business has priority into the “key transactions” understand the trends to stay ahead of established the metrics to measure group. Similarly a multi-geo site can competition. the effectiveness of intended online identify the locales and countries into business strategy. All the web analytics “key geo factors” group. frameworks provide intuitive graphs Note: These categories and flows/ which can be developed around the transactions need not be part of their

External Document © 2018 Infosys Limited External Document © 2018 Infosys Limited done in the previous step: Site usability Sources Conversion statistics • Fast moving products • Seasonality trends • Page views • Referrals • New visitors • Increase critical information • Product mix and bundles Recognize • Pain points patterns • Time on site • Search engines • Returning discoverability for the site • Analyze reason for exit • Downloads • campaigns visitors • Click Map • Leads • Enhance site design • Keyword frequency Measure metrics for • Click Paths Visitor Pro le • Exit rate • Site usability metrics key transactions • Click-through • Bounce Rate • User Rate • Visit duration • Design effective campaigns for dealers, • Search Identify key segmentation • Information • Checkout • Task completion categories and • Search keywords distributors etc. discoverability Abandonment rate • Shopping cart per visit ows/transactions • Geography Rate • New Visitor • Promote targeted content for users • Items per order Performance Conversion Rate rate • Return Visitor • Make site more navigable Figure 2 : Hierarchy of steps in implementing • Landing Page • Landing Page Conversion Rate Bounce Rates Load Times • Visitor Loyalty & web analytics emergence strategy • Availability Visitor Recency • Fix issues in the steps/pages where • Speed • Geo speci c page users are exiting load times Steps in implementing web • Improve SEO to increase site traffic analytics from external search engines Few common metrics obtained during this This section provides key steps for step for an e-commerce site is given below: • Understand site performance in implementing the Web analytics. different geographies These metrics need to be carefully selected Essentially any effective web analytics based on the critical business goals and • Personalize the information and flow implementation should cover these three flows/functionalities. The list should based on user’s interest/usage history steps: also need to be refined based on the • Acquire the key metrics industry domain. For instance checkout A business framework for abandonment rate is an essential part • Analyze the information acquired automated web analytics of e-commerce site whereas it is not • Act based on the analysis applicable for information display sites. implementation Following diagram depicts these three Now we have understood the AAA model steps with sample metrics information. B. Analyzing of web analytics implementation, we can Once the required information about key now focus on implementation details. A metrics is acquired, the business analysts typical implementation process involves Acquire Analyze Act need to analyze to make sense of the following sequence of steps: obtained information. Most of the web • Site usability • Site usability • Increase • Business identifying metrics • Sources discoverability • Sources • Enhance site analytics frameworks provide intuitive • Visitor Pro le • Visitor Pro le design • Business providing detailed • Conversion • Conversion • Design eective visualizations and dashboards to provide campaigns statistics requirements to implementation team statistics • Promote targeted content for users holistic view of the captured information in • Make site more • Implementation team updates relevant navigable near real-time. Following list provides the • Fix issues in the page tags and performs testing steps/pages information that can be analyzed from the where users are exiting • Improve SEO metrics captured in the previous step: • Operations team deploys the updated • Understand site performance code to production • Personalize the Site usability Sources Conversion information statistics • Information • Campaign • Visitor stickiness As we can see that above process involves discoverability e ectiveness • Transaction multiple teams and there is definitely Figure 3 : The AAA framework of web • Issues with page • SEO e ectiveness Completion/aba design • External ad ndonment rate a lag when the business decides to its analytics implementation • Site interactivity e ectiveness • Exit rate • Path/Information steps/pages actual implementation in production architecture Visitor Prole environment. However in some mission e ectiveness • Search A. Acquiring e ectiveness critical applications the lag is not an • Geo/browser acceptable option. This step consists of acquiring the metrics visitor percentage related information. The metrics can be This section discusses methods and either obtained by mapping them to techniques to completely automate the business goals (intended strategy) or can C. Act above sequence of steps. We will also go one step ahead and eliminate the be obtained from markers at key flows and After the analysis, the business team can involvement of other teams. So when transactions (emerging strategy). come up with an action plan to “act” on business decides the changes to its the analysis. This is the crucial step in the web analytics framework it can use this model. Following is an indicative list of framework which deploys the changes actions that can be taken from the analysis automatically.

External Document © 2018 Infosys Limited External Document © 2018 Infosys Limited The following diagram depicts the business team to fully automate the end- can specify “how” the framework needs to high level working of this automation to-end management of a web analytics assign those variable names. framework: framework. There are two simple ways to assign a value Here are the detailed components of the to a tracking code variable: static wherein e nalytics Imlementation rameor framework: the business specifies a hardcoded value ransformer rameor and dynamic which depends on the page Interface A. Interface and user/session context. Static values are aacrit prefixed with “S_” followed by whereas dynamic values are capture the essential built for the business community. It’s a web Elements for automatic prefixed by “D_”. The framework has code page tagging including based interface wherein the business team URL, element class, event to populate the values prefixed with D_ at name Etc. can enter specific details of web analytics ulisin o runtime. For instance in the above example that need to be tracked. Let’s consider a values like URL, locale etc. are dynamic. The simplistic scenario for discussion purposes: usiness eam framework assigns the appropriate values Deloyment business wants to track the traffic and to these variables at runtime and provides page visit data for their product landing information the meaning of these tracking Figure 4 : Web analytics framework page (www.ecommerce.com/products). code variable names and how the dynamic components Normally a page visit tracking requires value would be calculated at runtime. minimal tracking. Let’s say that we require Before the explanation of the web analytics following variables for tracking page visit: Another feature of the framework is implementation framework a brief that it “inherits” the tracking code. In the • URL discussion of how most of web analytics above example, all pages which matches frameworks work is given below: • Browser details products/* which are subsequent product 1. Typically web analytics frameworks • User details pages will inherit all the values specified for products landing page. In addition it can need the supporting JavaScript files • Geo which can be optimally placed in the also override and/or add new values. In this • Locale footer section of the page. For instance example we are adding a dynamic value for Google Analytics requires ga.js file[6] • Campaign tracking product name. 2. Upon page load the required tracking Business team has to specify that the Let’s consider another scenario wherein the code/JS variables are populated. products landing page wants a “page visit” business wants to track the click of button. tracking and also the values for the above The business team wants to specify the 3. The tracking code will be sent to the variables. They can specify that details in values as given below: web analytics server as an image the interface: request. The code is sent when the • Page: The URL of the page appropriate trigger happens (page load ae nalytics racin ode • Section/Module: The module which ecommercecomroducts LD or event) roser Droser contains the button ser Duser 4. Web analytics server builds a near eo Deo • Element: Id or class name of the button Locale Dlocale amain 12 real time reports based on the data it • Event name: Click or mouseover receives. ecommercecomroducts roductame Drodname • Tracking code: Tracking code as Web analytics implementation framework Table 3 : Framework interface details specified in the previous table. is built on top of these steps and provides The interface also offers highly intuitive following additional capabilities: Let’s examine what the above table does: features to visualize the sections while The “page” column indicates to the web • Automatic injection of tracking code specifying markers. For instance they analytics framework that the URL specified values. For instance population of can preview the section and component needs to be tracked. In the “Analytics values related to locale, user attributes (button) for which the tagging is specified Tracking code” column the business team etc. to ensure that the marking is done for the can also specify the values for tracking • Automatic addition of the events correct component. code. that need to be tracked. For instance Specifying tracking code in the interface Note: The details of tracking code is tracking the click of a new button. is the only key step that business needs to technical; to abstract the inner details of • Automatic re-deployment of the perform. Once it is done they can submit this tracking code from business team, the updates to all publishing servers. the details to the framework and it takes interface pre-populates the tracking code care of handling all subsequent steps till These steps would essentially equip a names required for tracking. Business team

External Document © 2018 Infosys Limited External Document © 2018 Infosys Limited the live deployment. Details are specified in Performance Management • Challenges in infrastructure including the subsequent sections. low bandwidths Though traditionally the web analytics • Price sensitivity B. Transformer frameworks are used for tracking key parameters, another facet is related to • Multi device enablement The transformer component collects all performance. It is possible to understand Let’s look at the implications for each of the details submitted and converts it into a the following key performance parameters these factors on web analytics framework JSON object. JSON object is subsequently [8] using the same techniques and implementation. used by the framework JS files to inject • Complete Page load times across the values on to the page. A sample JSON geographies A. Infrastructure challenges object for the table III is given below: • Perceived page load time across As many locations of emerging markets { geographies have low bandwidth following web page: ‘www.ecommerce.com/products’, analytics metrics gain prime importance URL:’D_URL’, • Load time for assets when implementing the web analytics Browser:’D_Browser’, • Availability across geographies User:’D_user’, strategy: Geo:’D_Geo, • Total transaction/process time across • Perceived page load time is important Locale:’D_Locale’, geographies. Campaign:’S_1234’ as the landing pages need to be light } Cross channel/device performance weight and load quickly. These parameters would help the business • Availability Once the JSON is created, the transformer understand how their online platform • Site usability metrics to enable easier sends it to the framework JS files. is performing across geographies. As information discovery performance is becoming the key success C. Framework JS factor for the online strategy it is important B. Price Senstivity The Framework JS files are the core to track the metrics related to these for a E-commerce sites need to identify the components of this framework. They are successful online campaign. fast moving products and seasonal trends developed in JS and will perform following Performance needs special consideration and provide price-sensitive product tasks: as it is one of the critical factors for recommendations. Following metrics need business continuity. In this regard the • Read the values from the JSON file to be closely monitored: business has to adopt following process: • Parse the page DOM after page load • Average order value • Setup appropriate real time and cross- • Inject the tracking code to the tracking geography monitoring process. This • Exit rate/bounce rate code variables can be achieved through web analytics • Inject the tracking code to the specific and other monitoring tools. C. Multi device enablement components and the events. • Setup an appropriate notification Enterprises also need to look out for These JS files need to be included in the process. For instance if the site multi-channel avenues for their online footer section of the page. response time falls below a pre- strategy. Following are the key metrics in configured threshold limit, it should this context: D. Publishing Job send automatic notifications to the • Multi-device load time site admins. Once the JSON and framework JS files • Click path analysis are ready, the publishing job will push/ • Setup a disaster recovery (DR) • Downloads deploy these two assets to the appropriate environment to handle business publishing locations. continuity Note: In addition to identifying the key metrics, it is always recommended that Post deployment all the user requests firms also have a strong strategy for would get the updated assets once their Relevancy of web analtyics identifying emerging trends. local cache is cleared. for the emerging markets As we can notice that above four Emerging markets challenges in components would eliminate the need for infrastructure and are developing development team and operations teams. economic countries. Following are the It promotes a self-service model wherein key factors that need consideration business can deploy their changes on- while adopting web analytics strategy in demand. emerging economies:

External Document © 2018 Infosys Limited External Document © 2018 Infosys Limited Summary effectiveness of intended strategies and There was a discussion of automating the to discover the emerging patterns along web analytics realization framework from In this paper we started the discussion with various metrics relevant for these two business alone. We saw how web analytics with a brief introduction of web analytics strategies. A sound web analytics strategy can be used for performance management concepts and its importance. We discussed for a business adopts both of these and how the discussion points are relevant the utilities and few sample scenarios of strategies. for emerging markets. web analytics. Then we also discussed two broad strategies to measure the We then touch based AAA model of implementing a web analytics framework.

References 1. Eric T. Peterson, “Web Analytics Demystified: A Marketer’s Guide to Understanding how Your Web,”

2. Avinash Kaushik, Web Analytics 2.0: The Art of Online Accountability and Science of Customer.

3. Jason Burby and Shane Atchison, “Actionable Web Analytics: Using Data to Make Smart Business Decisions”

4. Bernard J. Jansen, “Understanding User-Web Interactions Via Web Analytics”

5. Dennis R. Mortensen, “Yahoo! Web Analytics: Tracking, Reporting, and Analyzing for Data-Driven”

6. Brian Clifton, “Advanced Web Metrics with Google Analytics”

7. Alistair Croll and Sean Power “Complete Web Monitoring: Watching Performance, Users, and Communities”

8. 4 Key metrics for performance

9. What are the most important metrics for e-commerce companies?

10. Seven Key Performance Metrics for Any E-Commerce Business

11. Empowering customer centricity

About the Author

Shailesh Kumar Shivakumar

Shailesh Kumar Shivakumar is a Technology Architect with the Consulting and Systems Integration division at Infosys. He has over 11 years of industry experience. His areas of expertise include Java Enterprise technologies, portal technologies, User interface components and performance optimization. He can be reached at [email protected]

External Document © 2018 Infosys Limited External Document © 2018 Infosys Limited For more information, contact [email protected]

© 2018 Infosys Limited, Bengaluru, India. All Rights Reserved. Infosys believes the information in this document is accurate as of its publication date; such information is subject to change without notice. Infosys acknowledges the proprietary rights of other companies to the trademarks, product names and such other intellectual property rights mentioned in this document. Except as expressly permitted, neither this documentation nor any part of it may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, printing, photocopying, recording or otherwise, without the prior permission of Infosys Limited and/ or any named intellectual property rights holders under this document.

Infosys.com | NYSE: INFY Stay Connected