The age of discounts, sadly, is coming to an end. For e-retailers, VC money is getting harder to come by and investors are, more and more, demanding metrics such as revenues and conversions. There is an increased focus on driving more revenues and hence profits.
Even the government, a few weeks ago, released a set of rules that directly affect the deep discounting models that were driving eCommerce in India until now. The new guidelines state: “eCommerce entities providing marketplace will not directly or indirectly influence the sale price of goods or services and shall maintain a level playing field”. So what does one do in times like these, when the playing field is changing daily, and a big monstrous competitor with bottomless pockets for funds is eating away into your market share? Or how does a new eCommerce entrant in a really crowded market, grab market share and increase revenues?
One does it in two steps:
Step 1: Achieve a level playing field with the market leader
Amazon is the clear market leader in eCommerce. Amazon invested in technology long before technology became the backbone of e-retail. Amazon invested in user data and data analytics before big data analytics became the buzzword. Amazon personalized its eCommerce offering almost 10 years before everyone else realized the importance of personalization.
Amazon collected user behavioral data for the last 15 years, building data models and refining them, continuously A/B tests the models and driving user engagement and satisfaction using these models. Even now, Amazon constantly A/B tests features in personalization that help it retain customers, increase engagement and hence, its revenues.
Over 86 per cent of eCommerce customers in a recent study by Forrester Research said that when retailers personalise, and personalise well, they are influenced to buy more. This is perhaps the reason why more and more people prefer Amazon, to other retailers (all other things being equal).
So how does one achieve this level playing field? One way to level this playing field is to quickly analyze consumer behavior, with data from multiple sources, other than just transactional and clickstream behavioral data. By using external data sources, one can effectively do away with years and years of data. Using real time sources such as social data, financial data, and other open sources of data, one can arrive at great insights about the customer like Spending Potential, Brand Affinities, Media Habits, etc. Using these insights, it is relatively easier to personalize and acquire that level playing field.
Step 2: Get ahead of the market leader
Having achieved a level playing field, one needs to get ahead of the market leader. One of the huge problems in retail today, is the lack of a structured Product Catalog. This is especially true in case of online marketplaces – vendors upload products, with scant regard to information about their products, resulting in an extremely unstructured, sparse product catalog.
The first thing that drives better product search and discovery is a well-structured Product Catalog, with detailed information about the products. One way to do this would be to add more information; 50-100 attributes about the product, as meta-data in the Product Catalog. Machine Learning, Image Recognition and Natural Language Processing play a huge role here. Semantic Technologies will help you define relationships between products in the catalog. When you have 500 products in the catalog, manually feeding the relationships is do-able. But when you have lakhs and perhaps millions of products in the catalog, which gets updated every hour, semantic technologies coupled with machine learning really comes in handy here.
A retailer in San Francisco, Stitch Fix, uses 100 to 150 product attributes, along with user behavioral data to analyze what products would work best for a given customer. The retailer already shows a 30per cent increase in conversions over their counterparts, who do not use product insights based analytics.
If one looks at Amazon’s personalisation, you will see that the products that are recommended are usually based on past transactions. If one has purchased a TV, 15 days ago, and is now back at Amazon, it does not mean that the user is looking to purchase a TV again. By using analytics that use both user behavior and product insights, one can understand exactly what the user is looking to purchase, in real time, and provide the user the right product, when the user’s wallet is open, so to speak.
This is evidenced on how many number of transactions now occur due to personalization. Following is a visualization of the jump in the number of transactions through Infinite Analytics’ recommendation engine, over the last one year period.
By just using these technologies, one can get a huge jump in conversions, very shortly. Some of our customers have seen a huge jump, close to 217per cent increase in conversions, within 3 months of using our product. Another customer has seen an impact of $ 125 mm on their annual revenue, after using our product. This proves that the approach that is outlined above really is the right approach to take, to level that playing field with the industry leader.
Authored By: Akash Bhatia, Co-founder & CEO, Infinite Analytics