The politics of categorization: Say good-bye to Target’s pink/blue shelves!

It feels like this is some sort of sign! But maybe it’s just a logical step that makes better sense organizationally. Either way, library scientists sure are happy to discuss the issue!

August 14, 2015; 8:51 AM • by John Matthew Upton

Last week, Target’s PR team got a good eye roll from all sorts of commenters calling them “too PC” for deciding to group toys by a category other than which gender they think will like it most. Is the retail bohemoth just super progressive, or is this just a logical move? How does the way our brains organize things represent who we are and how we’ve changed? Well, guess what! We know a taxonomy expert, and asked him to weigh in!

The library science world is prone to neither publicity nor melodrama. In an industry where professional differences of opinion are settled with index card and quantitative analysis, a Fortune 50 company making national news with their categorization practices will live in infamy for generations. It also partially explains why I pumped out a 600-word response to an RVAnews editor’s innocent email about retail categorization practices. (The other partial explanation is that I am a pedant.)

Target shopper Abi Bechtel ignited the controversy with a single tweet:

Don't do this, @Target pic.twitter.com/cfh3cp5Nqa

— Abi Bechtel (@abianne) June 1, 2015

After the tweet went viral, Target announced last week that they would no longer use gender as an organizing principle in the children’s toys and children’s bedding categories:

“Over the past year, guests have raised important questions about a handful of signs in our stores that offer product suggestions based on gender. In some cases, like apparel, where there are fit and sizing differences, it makes sense. In others, it may not. Historically, guests have told us that sometimes–for example, when shopping for someone they don’t know well–signs that sort by brand, age or gender help them get ideas and find things faster. But we know that shopping preferences and needs change and, as guests have pointed out, in some departments like Toys, Home or Entertainment, suggesting products by gender is unnecessary.”

Most of Target’s customers approved of the change: their data capture and analysis capability has been on the bleeding edge since 2010, and they simply do not make changes without indicators that they will be successful. But that has not stopped a loud and angry consumer constituency from making their voices heard. Since the announcement, Target’s Facebook page has been inundated with comments like these:

“My family will no longer be shopping in your Politcally Correct store any longer. My wife and I have 2 ‘boys’ and 1 ‘girl’ and we are not transgender. Therefore we will no longer shop and support your PC transgender store.”

“How stupid! Really you are removing gender based signs! So afraid to make someone feel uncomfortable! We will see how uncomfortable you become when you loose long standing customers! I am going to stop supporting all of these companies that have gotten on the Bruce Jenning train! Its about to be derailed!!!”

Hence the controversy: does this change in categorization practices represent yet more evidence that liberal dark arts are hastening our fair nation’s descent into politically correct oblivion? Or does it merely reflect taxonomy best practices?

Cue your humble narrator. I may not be the foremost expert on taxonomy theory and design, but I did spend more than six years as a taxonomy consultant, with three presentations a Taxonomy Boot Camp (our annual industry conference), and spent five months on a contract revamping Target taxonomy during their replatforming initiative in 2010-2011, in which I handled the “Toys” category.

For most of the rest of this article, I will delve more deeply into the science of classification, specifically the theory and design behind retail taxonomies. If you’d rather chew glass than delve with me, you can just skip to the end. I don’t mind. For the rest of you nerds, let’s start with the tendencies of human beings as they relate to search and findability.

First things first: human beings are categorization machines. Neurologic processes take a lot of energy, so human brains have become adept at developing and consistently using mental shortcuts. Don’t believe me? Here’s an example:

Name 10 items you would find in my kitchen.

I can guarantee an 80% hit rate for this question–not because you’re spending a lot of time in my condo, oh stalker thou–but because you have a mental model of “things that belong in a kitchen.” Your brain has developed a shortcut that makes the items less energy-intensive to recall.

Taxonomies, such as those used on retail websites or inside brick-and-mortar stores, are meant to reflect the mental models that people use to discover and process information. The most important element of taxonomy design (to me, personally, but also to many other professionals) is consistency. The first thing that I look at when I’m doing a taxonomy analysis is whether they use consistent organizing principles. It’s not hard to find a retailer that performs poorly in this respect, so after 25 seconds of searching I settled on Kohl’s as a shameful example.

Here is the ubernav for kohls.com, with the corresponding organizing principle of each:

For Home : Place
Bed & Bath : Room of Use
Furniture : Object
Women : Gender
Intimates : Garment “Type”
Men : Gender
Juniors : Age
Kids : Age
Baby : Age
Shoes : Object
Jewelry & Watches : Object
Active : Adjective
Sports Fan : Customer “Type”
Clearance : Cost

That’s nine different organizing methodologies for 14 different categories, which means a user needs to browse and parse information in a different way for nearly every individual product hierarchy, and is officially in my top five worst major brand taxonomies I’ve ever seen.

Another interesting element to this debate is that brick-and-mortar stores and online stores have very different use cases. Consider a grocery store: they don’t necessarily want you to find everything you need immediately but would prefer you to browse and continue adding items to your cart, which explains the core organizing principle of “put bananas everywhere.” Consider also the case of parmesan cheese: there are plastic cans of parm in multiple locations because merchandising food relies on the semantic relationships between food items rather than a strict top-down classification. To put it in layman’s terms, grocery stores leverage “mini-groupings” that represent actual use cases to get you to buy more stuff. That’s why you can find parmesan in the cheese section but you can also find it next to the spaghetti sauce. (And only foreign transgender socialists eat spag sans queso.)

While mixing organization concepts is generally a bad practice, it’s not always avoidable. Both Target and Kohl’s have a very large range of products within their assortment, and sometimes the most common method of organizing any particular group does not fit consistently with the overall organizational scheme. Exceptions must (nearly) always be made, because a taxonomy should be built to maximize findability for the greatest number of users.

Target has embraced this element of taxonomy design, mentioning specifically in their statement that “shopping preferences and needs change and, as guests have pointed out, in some departments like Toys, Home or Entertainment, suggesting products by gender is unnecessary.”

So, the question remains: is Target’s stance the result of strong knowledge of information science best practices, or is it a stealth attack intended to wussify America like the loud people claim?

Though the organizational concept of “gender” was always less useful within the Toy category as it would be in other categories (such as “Clothing”), it was still useful while the societal concept of gender applied more directly to toys. As the zeitgeist has moved away from enforcing gender roles in children, the usefulness of gender as an organizing principle has lessened as well, which has eventually resulted in the change referenced in the article. (For reference, both the person and the taxonomist in me approves of this change.)

That’s not to say that gender can’t be a valid organizational principle for lots of things, and there are usually multiple methods of organizing any given group of products or concepts. For example, some medical conditions are sex-specific. What a taxonomist must do is gauge the benefit to findability of each grouping vs. the potential negative impact to the consistency and usability of the whole schema.

This update is in the best interest of customers and librarians both; any attempt to politicize the issue should be forgotten on the shelf between the train sets.