As you can tell by our prior posts, we don’t pull any punches here at UnFuckIt. We are passionate about product data, and we don’t believe that sugarcoating the problem is in anyone’s best interest. So today’s blog is all about the impacts of when someone fucks up data, as shown in three easy pictures.
#1 The “Not Applicable” Fuckup
As an employee of a major discount retailer a decade ago, I discovered how badly the term “Not Applicable” fucks up product data. I spearheaded a movement to eliminate every single instance of that term possible, which took six months and a lot of planning. So today, when I was searching on a major retailer’s website (who will remain anonymous but is not the company I previously worked for) I was started to see that they are still using that terminology, and in a very visibile manner.
You will see in the faceted navigation experience to the left that the value of “Not Applicable” appears in this appliance series attribute. It’s one place out of many where I found this occurring on this site. It is also the least valuable piece of data you will ever find. If you look at the list, there are both brands and sub-brands present, but the website in question sells 50+ appliance brands. That means that most of the brands do not show up in this experience.
Is that a bad thing? Very definitely. Imagine you are Danby, who sells tons of refrigerators, mostly for smaller applications. NOBODY will ever click on “Not Applicable”, meaning nobody will EVER find Dandy products using this filter. They can filter by several large companies’ brand names, but the lack of an option to select anything by Danby limits their visibility on the website.
This hurt’s Danby’s presence on this site, but also hurts this site. Danby is potentially losing sales, but anyone using this filter is by default removing Danby products from consideration in their purchase. If the perfect product happens to be a Danby product, this can lead to a lost sale, or even worse… a return of a product that can no longer be sold as new AND costs a ton to ship.
You can argue that nobody will ever use the Appliance Series filter, but then why HAVE an Appliance Series facet in your navigation? It wastes real estate on a page that could better be used for a size facet, a color facet, or other facets that are vastly important to the refrigerator-buying experience.
How does this fuckup happen? Simple. Somebody in the data collection process set the attribute “Appliance Series” to mandatory. They didn’t want to add every single possible series, so they gave “Not Applicable” as a way to avoid filling it out. If you actually click on “Not Applicable”, 716 refrigerators appear out of 790 total refrigerators available. Knowing nobody will click on “Not Applicable”, that is 94.6% of refrigerators that will not show in this experience.
Again, you may highlight that people clicking this filter may only want to see major brands. However, refrigerators from GE, Frigidaire, Maytag, and Kitchenaid ONLY show in the “Not Applicable” experience because they were not part of a series. Imagine “Not Applicable” being available for a Maximum Torque setting on a cordless drill or on the color attribute for an appliance. Lazy people in the data collection process will select “Not Applicable” because it is easier than requesting the right value.
How do you unfuck this fuckup? Stop adding “Not Applicable” to attributes you make mandatory that don’t apply to all items that will have that attribute. It creates garbage data, which leads to fucked up experiences, which leads to bad brand experiences for your customers. “Not Applicable” isn’t data, but it is also a sign that the way you treat attributes is faulty and needs to be revisited.
#2 The Mutual Exclusivity Fuckup
I’ve talked about mutual exclusivity for decades. The common response I get when bringing the topic of mutual exclusivity into a conversation is “Yeah, we fully understand it”. It’s generally a lie. Most businesses do not understand the true implications of mutual exclusivity.
First, a definition: Mutual Exclusivity is the best practice of only allowing one category or attribute value to be selected for a single product. It also means that two attributes should not have the same definition, but that’s a story for another day. In our example today, we are going to look at what happens when you violate mutual exclusivity.
The faceted experience to the right is for gas-powered mowers. In the middle of the screen, you will see values for “Adjustable Speed” and “Variable Speed”. Adjustable-speed mowers allow you to adjust the speed. Variable-speed mowers allow you to adjust the speed. There is a highly technical difference in these values, because a variable-speed mower will allow you to vary infinitely, whereas an adjustable speed has settings for what speed you can set it to.
It’s minor, but the bigger issue is that customers don’t know the difference. Even more confusing is that the same lawn mowers show up many times in both facets, which means the manufacturer doesn’t know what the value means. If the manufacturer doesn’t know the difference and the customers don’t know the difference, what is the value of this facet? Clicking “Adjustable Speed” will get you some mowers that may be variable speed, and vice versa.
This difference might seem petty if you’re not buying a lawn mower, but the same site lists a facet under power tools for Corded/Cordless that includes the values “Corded”, “Cordless”, “Hybrid”, and “Both”. A hybrid tool is a tool that can use either a cord or a battery. That is the EXACT SAME definition of the value of “Both”. Sure, maybe one or two of those tools may have a hand crank to power them, but otherwise, these values are violating mutual exclusivity. The customer is going to pick one of the two values, and get different experiences depending on which value they choose, where both values should be selected. Violating mutual exclusivity can cause missed sales because the product the customer should have bought is never presented to them.
To unfuck this problem, make sure your values that can be selected in a drop down list such as these are mutually exclusive, meaning they have unique definitions per value. Otherwise, you may be selling your customers the wrong product or forcing them to buy elsewhere to find the right product.
#3 – The Generic Categories Fuckup
The final way I’m going to show today to fuck up your data is to empower very generic classifications of products. Unfortunately, I can’t show you the mess I found on the retailer I’ve been showing without directly exposing them, which I won’t do. But suffice it to say, it’s a mess.
I clicked on the category “Storage and Organization”. In the Product Listing Page (PLP) experience I was greeted with, my first six products were some Gorilla Tape, an outdoor gazebo, a furniture dolly, an over-the-toilet cabinet, some industrial plastic wrap, and a 10′ x 14′ metal shed. Perhaps if I wanted to move my toilet into a new shed on a furniture dolly, tie it down with Gorilla tape, wrap the seat with plastic wrap as a practical joke, and finish the ensemble with a nice gazebo this would have been helpful. I’m guessing you can already see how it’s not helpful.
The bigger problem is the facets that appeared based on this choice. I had price, which started at $0-$10 and went to $5000+, with an extra value called “Special Values”. “Special Values” is not a price, and when you have products that range from $1 to $5000 in the same experience, you have a problem. I also saw Brand as a facet, which included ALMOST EVERY BRAND THAT RETAILER SELLS! I had a category filter, which included some wonderful values like “Hooks”…. I get you can use hooks for organization, but hooks are hardware!!!! They also had three different shelving categories based on room, but a generic shelving category as well. Then there was the Bundle facet, which let me pick if something was part of a bundle.
None of these facets help me solve the problem I’m trying to solve by coming to that website, and it’s because the website tried to put too many things into a single experience. The data required to make this experience occur involves putting attributes on categories that don’t make sense, and perverting the definition of a facet to include values that don’t meet the definition of an attribute. I don’t care if my Gorilla tape or my furniture dolly is part of a bundle, but someone had to collect the data to make that work. I don’t care if my gazebo is a special value if I’m trying to find a gazebo for under $1000.
Here’s the truth of this situation – It’s not garbage-in, garbage-out. It’s a web hierarchy driving a data collection experience that is ridiculously stupid. When the display of products in any experience has an outsized weight on the data collected for that experience, the experience drives the garbage. It’s Garbage-Expectation, Garbage-Data Collection. Vendors hate it, customers hate it, but companies continue to do it.
The unfucking of this experience requires a deep look into how these experiences come together BEFORE making decisions on how collect data. If all you’re going to do is generize your experiences at the cost of collecting ridiculously unintelligable data, your data is fucked. Have a strategy for your data presentation that is supported by data, not a data strategy driven by unrealistic (or moronic) experience expectations.
The Fucking Summary
I would like to be able to say this was limited to a single retailer. It wasn’t. I just used one retailer as an example. I’d like to say that I spent hours finding these issues. I didn’t. They were everywhere, and I chose the ones I wanted to choose. Bad experiences drive bad data collection, especially when the people setting up the experiences and the people setting up the data don’t communicate. It leads to bad customer experiences, lost sales, outsized returns, and a failed brand promise.
Want the UnFuckiteers at UnFuckIt to analyze your data to show you how you’re fucking it up? Click our “Contact Us” page for details on how we can help.