How Predictive Analytics drives Competitive Advantage in Online Retailing

A sneak peek into the new budget of the Modi Government reveals plans to open up India’s $500 billion retail sector to global online retailers such as Wal-Mart Stores and Amazon. With this the competition in Indian online retail industry heats up as online retailers jostle the e-commerce space to engage customer activity.

Online retailing in India took off in the last 7 years growing at a compounded annual growth rate (CAGR) of over 56 per cent from around Rs 15 billion revenues in 2007-08 to Rs 139 billion in 2012-13. This was driven by increasing internet penetration, changing lifestyles, improved middle class purchasing power and robust use of credit cards. While it is further estimated that Indian online retail industry will reach Rs 504 billion by 2015-16, one can definitely expect a further revision in the estimate with implementation of post-budgetary reforms.

The concept of purchasing online has caught the fancy of Indians because of the convenience of online product research, comparison shopping and competitive pricing, besides the comfort of home delivery. Flipkart, Jabong,Quikr and more – India’s Rs 18,000 crore online retail industry is growing at a rapid pace. With a diverse range of listings and loads of data on the buying and selling patterns, Big Data analytics is deeply ingrained in the DNA of these online platform.

As a student of analytics learning the SAS / SPSS , you’ll be excited to know that SAS & IBM-SPSS products are in popular use by online retailers. These drill through multiple levels of data – visitor navigation, search trends, buying and selling patterns – to understand customer/ visitor profile, his needs or expectations,  for a competitive advantage.

Big Data analytics is leveraged for actionable insights, to build descriptive and  predictive modelling algorithms (decision trees, bagging and boosting, linear and logistic regression, neural networks, memory based reasoning, partial least squares, heirarchical clustering, self-organised maps, sequence and Web path analysis) and strategise implementations.

Product offers, advertising strategies, SoMoLo (Social, Mobile, Location) engagement, identifying profitable customer segments, and assessing credit worthiness are some of the actions based on the insights.

Some typical cases of analytics-in-action in online retailing:

Targeted advertising  – A customer looking for children’s clothing is grouped as a ‘parent’ / ‘relative’ and offered targeted advertising based on the search criteria. From clothing sizes browsed, an insight is obtained into the child’s age group – based on which predictive ads / listings are put up to cater to the customer preferences.

Predictions – By analysing time series and correlating it with external variables like weather, festivals and competition, predictions are made to understand various metrics like sale volumes, fast-selling items, visitor profiles, and so on.

Competitive Benchmarking – It is used for comparing the online retail operations against those of the competitor’s to gain information about market trends, customer habits, vendor practices,  promotional campaigns,  and geographic / location penetration.

Behavioural targeting – Analytics helps to leverage real-time contextual advertising based on customer browsing behaviour, buying history and inferred psychographics. So if a visitor is looking for travel bags in the luxury category, he is targeted with advertisements of travel insurance plans.

Geo-behavioural targeting makes use of many variables like location, referring URL, device used and so on, to target content based on location. For instance, a browser located in Sikkim with a buying history of apparel will be targeted with listings / offers of warm jackets and similar products.

So now you know why if you are a student of social science, MBA or IT, additionally equipped with analytics training you would be a preferred candidate in analytics hiring!

Leave a Reply

Your email address will not be published. Required fields are marked *