Retail Resolutions for 2018: A five-point agenda for US retailers to win with their data

DECEMBER 1, 2017

Prashant Agrawal, CEO

A growing economy, “sky high” consumer confidence and low unemployment should spell good times for the retail sector. But the retail sector is suffering. Gloomy news abounds – store closures, capture of market share by e-tailers led by Amazon and shrinking margins. Does it have to be this way? Not necessarily. What is playing out in the US retail landscape is an evolutionary game of the survival of the fittest. We believe that some retailers will lose and bow out, while the winners will win big. Efficient use of data needs to be at the heart of the winners’ game plan. We have laid down a five-point agenda for US retailers to use their data wisely in 2018.

Retail Resolutions

Target 1: Reward your loyal customers by knowing them really well

Modern retail trends and technology have made every US shopper behave like an ‘entitled’ one. Shoppers are no longer content with merely the product, they crave experiences – it is this realization that partly fuels the online-to-offline move. Loyalty award systems are now table stakes[2] – customers demand personalized offers and expect the ability to control their offers.

To effectively personalize and provide an enhanced customer experience, retailers need to mine their data to know their customers well. They need to curate descriptors for each customer by cleverly analyzing their data. Advanced machine learning algorithms like clustering, multi-class classifiers and recommender systems need to be employed to precisely tune personalization strategies.

Once a retailer has set up a personalization solution that provides marketers easy access to 360-degree customer behavior data and lets them easily experiment with it, a floodgate of opportunities for personalization opens up

Retail Resolutions

Target 2: Compete on price with online players, but be smart about it. Don’t blindly match online pricing.

Way back during the holiday season of 2013, Amazon was expected to make price changes to 80 million or more items. In 2017, they probably change prices even more frequently. Amazon uses a wise strategy to drive low price perception by discounting only certain items. Take Amazon’s acquisition of Whole Foods Market. Amazon won the day by decreasing prices on the price of a few products (every day grocery items such as eggs, milk, apple types – technically known as KVIs – Key Value Items) and slowly changed the price perception of Whole Foods Market. To compete with Amazon, retailers should not react by putting their entire items on sale. They must dive deep into their data to identify price elasticities for each product, and sometimes at each location. They must estimate all indirect effects from promos. This then should inform their pricing and promotions strategy to achieve the following:

  • Raise prices of inelastic items, at inelastic locations
  • Optimize prices on elastic Key Value Items
  • Avoid toxic promotions that are depleting margins
  • Leverage profitable promotions through tuning for maximum impact

Data is the key to profitable pricing and promos and all retailers should be understanding the value they bring to the customers.

Once a retailer has set up a personalization solution that provides marketers easy access to 360-degree customer behavior data and lets them easily experiment with it, a floodgate of opportunities for personalization opens up.

Retail Resolutions

Target 3: Every store is different – so have a specific strategy for each one

Many legacy US retailers have closed stores in 2017[4] and many more plan to do so in 2018. However, physical retail is nowhere close to being dead, as online-to-offline moves by several players suggest. Retailers need to seize the opportunity offered by an experience-craving clientele by having a differentiated strategy for each of their stores.

This can be facilitated by synthesizing multiple disparate sources of data, which often do not stay together within an organization, by creating 360-degree store scorecards. Modern A.I technology can then help make sense of the data to suggest strategies for each store.

Sam Walton famously used to allow store managers to price 100 items as they saw fit. Store and district managers need to be empowered to make decisions and be rewarded for doing so. 360-degree store scorecards can help in achieving th

Target 4: Be prepared to sell, every time and anywhere

Chinese retailers have been pioneers in offering same day deliveries and now US shoppers too expect rapid deliveries[5]. US retailers lose more than $100 B annually[6] due to stock outs, and this is an extremely costly error. At the same time, with rising real estate and financing costs,[7] retailers cannot also afford to be saddled with excess inventory.

The need to be able to fulfill orders via multiple delivery channels and always be well stocked, while not storing excess inventory makes it necessary for retailers to be able to forecast their demand with a high accuracy and manage their inventory right.

The forecasting problem is indeed daunting with multiple challenges being posed by the need for granular forecasts and shifting trends, but high-powered machine learning ensembles can be employed to tackle this.

Stores can be assets with great benefits and should be treated such. Walmart is now charging more for some items at its store. The myth that the consumer is checking the phone and pricing for each item is not true. Amazon is not competing on price anymore but convenience.

Target 5: Start training the bots now before your competitor does

The use of chatbots in retail is an emerging trend with countries like Singapore and China leading the way. Several US retailers have also taken an early lead with chatbots. Retailers need to understand two realities about bots immediately and frame a chatbot strategy before the time runs out.

First, chatbots are not mere replacements for human agents but can be more effective than them – “Superagents” – and this makes them inevitable. Chatbots can actually offer a much more engaging experience to customers with their always-perfect personalized recommendations derived from AI, their ability to present limitless information and by being available always at the convenience of the customer. This means that chatbots will start ruling retail very soon – the lack of ‘human feel’ is what is stopping that from happening right now, but the advances in NLP and cognitive technology will soon surmount that obstacle.

Second, there is no overnight path for retailers to get a fully functional chatbot for themselves. This is because off-the- shelf chatbots come trained on general conversational data but need to train on specific retailers’ contextual data. It is impossible to collect this data without first having a basic rule based chatbot interacting with your customers. So, while the early trial bots might not appear to be very sophisticated at this point, they are collecting invaluable real context data and will give the companies pioneering them several years of advantage.

Others need to start trialing bots without further delay and go through iterations over 2018 to be able to deploy fully functional bots in 2019. They are easy and cheap to start experimenting with as social media giants such as Facebook provide an easy platform to quickly deploy bots.

Retail Resolutions