Raisins and endcaps: What grocery store organization can teach us about information taxonomies

David Holmberg
4 min readMar 19, 2015

Over the last year, I've been part of a project to reorganize two million pieces of UGC for a website. The content was originally loosely organized into 130 or so categories, but in order to make this even approach user friendliness, every one of those categories needed to be broken down into dozens of additional subcategories. This meant working a lot on taxonomies: making new ones, revising old ones, and revising the new ones again. During this process I've found the limitation of taxonomies, as well some possible solutions.

If you've ever hunted for raisins at the grocery store, perhaps you've encountered the same situation I have: every other item on your list is in the cart, and you’re drifting between aisles, trying to determine where the raisins might be hiding. Are they with the fruits (since they’re just dried grapes, after all)? Or are they with the cereals (since that’s the most likely place they’re eaten)? Or with the baking supplies (since that’s probably the second most likely place they’re used)? What’s so maddeningly frustrating about this search is that they’re not even an unusual food. And yet something about them defies easy categorization.

Underlying the Case of the Hard to Find Raisins is a classic UX problem: my mental model is not matching the mental model of whoever’s organized the grocery store — if it did, I wouldn't still be staring stupidly at the “Spices and Baking” sign wondering if that is where they’d hidden my raisins. And yet the grocery store stands as perhaps one of great taxonomical achievements, second only to the Dewey decimal system: outside of the occasional question to a passing employee, we successfully navigate thousands of diverse items, from dog chew toys to California Pinot Noirs.

So what’s up with the raisins?

First off, I don’t think it’s a problem with the taxonomy. Grocery stores have developed an extremely successful strategy for organizing all of these products. After all it didn't have to be this way; stores could be organized by brand, by meals, by calories, by color, by how likely my daughter is to eat it (ranging from chocolate bars to vegetables, of course). But the decision to go primarily by type of food works amazingly well.

Except for the raisins.

I will say, as well, that this isn't a problem with the raisins. After all, they’re just dried grapes. And grapes are fruit. No, it isn't raisins, because even if we reorganized the grocery store to account specifically for the raisins, it’d just be something else.

Instead, the problem is in the very idea of taxonomies. The promise of a taxonomy is that every item from a diverse set can be grouped together in a way that is meaningful and includes all members of that set. The reality, though, is that nearly every taxonomy has a remainder, an item (or two) that defies the system which otherwise works so well to categorize every other item. And even if you were to start over again, even if you were to rearrange it all and rethink the entire system, you’d still be left with a completed puzzle that somehow includes a few mysterious pieces.

These pieces are the raisins: that irreducible “other” of all taxonomies, that item or two which simply will not logically fit under any other term, either because it relates to too many or to too few.

It’s important to recognize the existence of raisins in your taxonomy. I've spent a lot of time working on taxonomies over the last year, and I've also wasted a lot of time trying to perfect them, trying to solve for that last remaining term or topic. But instead I always get raisins.

There is a lesson here, though (besides accepting the existence of these raisins), and one that we can find if we return to the grocery store. Because once you look for raisins long enough, you start to notice something startling: the raisins are everywhere. While it’s true that they don’t seem to have a specific home, they exist in small pockets in almost every place you might think: here’s a small display with the fruits, here’s another by the cereal. And most of all, here’s an endcap or two to make sure you see them. It seems that grocery stores recognized the difficulty of placing raisins within a mental model, so they decided to disperse them throughout the store. I call this “endcapping”: the strategic locating of desirable but hard-to-classify items.

In other words, what the designers of grocery stores realized — and what endcapping demonstrates — is that a perfect taxonomy is impossible, but you can still help your users find what they’re looking for by recognizing that fact and accounting for it through the frequent, strategic repetition of those problematic items. This is not a perfect system — a perfect taxonomy would be a perfect system. But if we’re to best serve our users, we need to acknowledge the limitations of any organizational model and incorporate solutions directly into the system. I wish it was different, but ultimately I do always find the raisins.