Clever Hans returns: winning in e-comm with Decision Paths

Michael Thompson and Lindsey Leikhim

Jul 5, 2022

Remember Clever Hans? The horse that tried to make people happy by solving math problems? We last discussed him in the context of showing how purely algorithmic approaches to retailer shelf layouts are a self-fulfilling prophecy, the retail equivalent of a horse stomping his foot until his owner looks satisfied enough that he can stop.

As silly as the mania over Clever Hans appears in retrospect, we can appreciate it was an understandable misinterpretation of a real observation. After all, Hans usually answered correctly when his owner knew the answer. But, just like purely algorithmic approaches to determining shelf layout, Clever Hans couldn’t tell people anything they didn’t already know.

What is less understandable is how certain channels continue to arrange products seemingly without any regard to how consumers make decisions. In particular, major e-comm platforms’ CPG assortment looks like it was arranged by a horse. And yet, online shopping is at a critical juncture given pandemic-induced behavioral shifts.

COVID has accelerated adoption of e-commerce CPG shopping. We've fielded research throughout the pandemic and found that around 20% of shoppers have tried a new e-commerce platform for CPG shopping during COVID, and most of those consumers expect to continue to use that platform moving forward. The stakes have never been higher to build the e-comm experience in a way that helps consumers engage with brands, locate the ideal product for their needs and explore new offerings.

It may come as a surprise, then, that sophisticated e-comm platforms don't do a better job assorting products in a way that makes sense to consumers. We observe three common e-comm shelf arrangement issues:

MIXED METAPHORS. Search for “corn” and you get corn tortillas. Search for “tortillas” and you get tortilla chips. Search for “chips” and you get chocolate chips.


SELL OUTS. Stock-out products remain at the top of search results, so you know for sure that someone else is eating your pretzel sticks right now.

JUNK DRAWER ASSORTMENT. A category like Frozen Pizza is frequently presented as a dog's breakfast of disparate options. A classic thin pizza may appear next to an indulgent rising crust option, next to next to a pack of five frozen cauliflower crusts, next to Pepperoni Pizza Hot Pockets.

Though none of these are ideal from a shopper’s perspective, Mixed Metaphors at least provide a backstop for difficult search phrasing, and Sell Outs help a shopper move on to alternatives. Junk Drawer Assortment is the least defensible: it’s frustrating from a shopper's perspective and discourages engagement with the category. Brick and Mortar stores have long applied rigor to their shelf layout decisions (albeit often misguided, as Clever Hans previously illustrated), which makes the current state of e-comm platforms seem behind the times in comparison.

But hope is not lost for e-comm platforms. With the right guidance, category assortment can be more intuitively arranged to better meet shoppers’ needs. The Cambridge Group’s Decision Paths framework can help e-comm platforms and manufacturers achieve joint growth.

As we discussed in our previous Clever Hans article, consumers balance multiple distinct drivers in choosing any product. Some drivers will be of higher importance to consumers, and the ideal shelf layout can be defined by the hierarchy of drivers’ importance. For example, let’s consider an illustrative example from the frozen treat category:

In this simplified frozen treat example, we can easily see the value in assorting ice cream and popsicles separately, with the opportunity to further subdivide those categories by secondary choice drivers specific to each category (as well as possible additional choice drivers from there).

The typical Decision Path for a food category can have anywhere from two to five hierarchy levels spanning elements of flavor, health, value, pack size, brand and experience. In each category, the Decision Path represents the ideal approach to guiding consumers through the aisle. Furthermore, virtual shelf sets are free from many of the constraints that in-store aisles are subject to, making the stakes that much higher for application of Decision Paths to guide online shoppers through decision-making in an e-comm context.

Exploring Decision Paths is a bespoke process that leverages custom consumer research to uncover the true drivers of choice in your category. Once defined, Decision Paths offer a multitude of near-in, practical improvements to the shopping experience, which in the case of e-comm could include:

Each of these applications relies upon partnership with e-comm platforms to implement changes, though none represents a technical challenge beyond the capabilities of a typical e-comm retailer of CPG products. Improvements will derive much more from understanding demand than from any technical changes. To put it another way, unique insight into how consumers make decisions in your category is the true “analytic horsepower” at play here.

Talk to us about what makes the most sense for your category and your e-comm partners. Given the rapid pace of change in the e-comm space, the stakes for providing a better shopper experience are massive and time is of the essence. Don’t spare the horses.


Michael Thompson  [email protected]

Lindsey Leikhim  [email protected]