A.I. Puts Top Grocery Holiday Merchandising Fallacies to Rest

When it comes to certain categories of retailers such as the grocers, the weekly sales promotions remain the heart and soul of marketing efforts across all channels—whether through direct mail, newspaper inserts, in-store or online mediums.

During holiday seasons, promotions become even more essential as supermarkets compete to win holiday baskets that comprise of a myriad of products, ingredients and accessories that draw store traffic and drive purchases.

But how do grocers select the right products to display in their weekly sales promotional circular that will not only bring in shoppers but encourage them to fill their basket with profit-driving product combinations? As grocery merchandisers rush to get their holiday circulars ready for distribution, it’s important to realize these decisions have historically been guided by a central decision-making myth: That to win holiday basket, grocers must promote all holiday items, and only holiday items, prominently in the circular. This fallacy often leads, unfortunately, to needlessly lower margins.

By way of example, consider the Thanksgiving turkey and everything that goes along with it. Yes, turkeys by themselves boost business big-time—in theory at least. According to Nielsen, shoppers who buy a turkey spend 1.5x more on average than non-turkey buyers across the store. Turkeys also have a very specific “halo effect.” When it comes to the big turkey dinner, there is an assumption that the buying of the bird will lift sales for the rest of the Thanksgiving holiday basket, i.e. aluminum tins, aluminum foil, potatoes, stuffing, cranberry sauce and gravy, pumpkin pie, etc.

Related:Thanksgiving Is Over!

So, the promotional decision is often made by merchandising executives to give away margin on the bird to capture the big Thanksgiving basket. However, more often than not, that same grocer extents the promotional discount beyond just the turkey to the rest of the holiday basket – discounting the stuffing, cranberries, gravy, potatoes, pies, etc. - thus giving away margin on the whole basket by promoting them all in the circular. The result? A double-negative margin effect.

This is where the fallacy lies. When it comes to that Thanksgiving promotions, the problem is that a lot of grocers fall into the trap of winning the holiday basket and then discounting every item in this basket, thus negating any overall positive effect to their business.

For example, research from Nielsen further found that at many grocers, one-third of all pumpkin and pecan pies sold in a year from the in- store bakery are purchased during the week of Thanksgiving. There’s simply no need to use up valuable promotional space and margin dollars to sell more pies on Thanksgiving. If the retailer wins the turkey, they will more than likely win the pie…and everything else in the holiday basket.

This thinking may have made sense in the days when data analytics were limited or non-existent, and promotional decisions were mainly based on “gut” feelings, vendor deals, assumptions and fallacies. After all, measuring product sales, associated sales, cannibalization, cadence and elasticity of 40,000 products every week is beyond human capacity.

But now, artificial intelligence (A.I.) is changing the promotional game in grocery…and across the retail landscape. With vast computational power and complex, automated, data-driven capabilities that go far beyond what humans can do, A.I. can run through massive amounts of TLOG data and help grocers understand what products would have sold regardless of the promotion — and thus make recommendations as to what grocers should promote instead.

A.I. in Grocery Merchandising: Busting Myths for More Effective Outcomes

Artificial Intelligence (A.I.) will play a tremendous role in the transformation of the grocery industry as a whole, significantly influencing every facet of every chain from huge multinational corporations to the small niche retailers. AI is replacing traditional predictive analytics and is now starting to fundamentally change promotional planning processes by reducing the amount of time required for merchandisers to plan promotions and creating more effective and predictable sales outcomes.

A.I.-powered solutions take massive amounts of retail data and consider thousands, even millions of variables to make better  promotion recommendations—looking at everything from halo effects, cannibalization, marketing channel choices, pantry loading, promotional timing, price elasticity, promotion elasticity, competitive effects, etc.

For example, promoting the Thanksgiving turkey and all its associated fixings before the holiday ignores all the customers who do not eat turkey at Thanksgiving.  A lost opportunity!

An A.I.-powered approach to product selection for a retailer looking to drive incremental revenue and profits from a holiday circular may recommend eliminating some of the turkey’s halo items from the promotion to create more promotional space for the grocer to speak to all its customers by including specials on other items that drive different non-turkey basket. For example, an A.I.-driven approach may recommend frozen pizza or convenience meals as a good alternative sought by customers who work during the holiday or may recommend the promotion of a different meat choice reflecting the buying data from certain ethnic shoppers during the holidays.

A.I. understands the difference and separately measures promotional elasticity and price elasticity for every single product.

The promotion of a pie, potatoes and tin foil might be enough without discounting, capturing more margin and allowing for even more discounting on the basket driver: the turkey.

The bottom line is A.I. allows grocers to use valuable sales and customer data to guide their promotional decisions and better meet customer needs in ways never before possible. Grocery retailers need to allow themselves to get beyond the traditional methods by which they have managed and measured promotional success and move towards a holistic approach driven by the right data-driven technology.

Editor's Note: Gary Saarenvirta is founder and CEO of Daisy Intelligence. He can be reached at [email protected].