Self-Checkout Game Theory

As a consumer of goods, I think a lot about the nature of self-checkout, one of the most important grocery-store innovations of the last 20 years. As we close in on the most consumerist holiday of the year, now’s the right time to talk about this thing we all use, even if we don’t love it.

Simply, the vibe isn’t really working in self-checkout’s favor. For years, negative takes have been written about this technology, which takes the humans out of the purchase process. In 2019, Brian Merchant wrote a take for Gizmodo that referred to the concept as the prime example of “shitty automation.”

“I have literally never, as in not one single time, successfully completed a checkout at a self-service station in a grocery store without having to call a human employee over,” he wrote. “And it’s not because I’m an idiot. Or not entirely, anyway.”

(He later built a book around this general take, titled Blood In The Machine. I’ve yet to read, but knowing Brian’s previous work, it sounds compelling.)

More recently, Amanda Mull of The Atlantic described self-checkout as a “failed experiment,” noting the broad systemic failures to this general idea:

All is not rosy in the world of self-checkout, and some companies seem to realize it. Walmart has removed the kiosks entirely from a handful of stores, and is redesigning others to involve more employee help. Costco is stationing more staffers in its self-checkout areas. ShopRite is adding cashiers back into stores where it had trialed a self-checkout-only model, citing customer backlash. None of this is an indication that self-checkout is over, exactly. But several decades in, the kiosks as Americans have long known them are beginning to look like a failure.

In the piece, she links a CNN Business story with the headline “Nobody likes self-checkout. Here’s why it’s everywhere.” Now, apologies to Nathaniel, Amanda, and Brian, who all raise valid points about the way this technology creates chokepoints elsewhere in the system, but I’m actually a fan of self-checkout, which I think, in the right contexts, can empower the consumer.

Off the top, I can think of three big advantages to self-checkout:

• No judgment. Recently, I went to a Walgreens—a pharmacy where people often buy embarrassing goods. I bought a $1 bottle of hot sauce and yet was still the target of cashier snickers. That makes me wonder how that cashier might handle condoms or Preparation H. Being able to purchase something without a human input is sometimes a net positive.
• Sense of control. My wife hates when other people bag her groceries, because folks don’t always do it to her preferences. It can be frustrating to see a bunch of heavy things on top of the eggs or the bread getting crushed, for instance. By using self-checkout, she is able to decide how she wants to bag her goods.
• Social interactions optional. Sometimes you just don’t want to deal with people. Maybe you’re an introvert. Maybe you’re in a hurry. Maybe you’re sick and you need to get through the store with minimal human interaction. Self-checkout allows you to do all that without having to pay for delivery.

Now, I will admit there are downsides to self-checkout, one of which I frequently see at my grocery store of choice, Lidl. It’s a great store with questionably named generic breakfast cereals, but at my location, they have these things called Cheaper Stickers on items that have been manually discounted. The problem with these stickers is that they must be approved manually by an employee. And there aren’t very many employees, because like its historic sister store Aldi, Lidl is meant to be run as an extremely tight ship. So there’s usually a short wait if we buy one of these products. But if we plan our purchase order correctly, it hits at the end, so most of our groceries have been bagged already.

OK, so self-checkout doesn’t always use the latest and greatest. (Anj Simmons/Flickr)

The “tight ship” comment hints at the other problem with self-checkout, the one that all these writers have leaned into: The idea that retailers saw an opportunity to replace people with technology, a bet that largely hasn’t worked out.

Many companies do questionable things with new technology and deserve to be called out. Just today, the Federal Trade Commission banned Rite Aid from using facial recognition technology, an element of some self-checkout tools, after its technology discriminated against customers. (For what it’s worth, Fight for the Future thinks no retailer should use it.)

But on the other hand, there are situations where it might be desirable to hand off work to the robots. In some states, particularly Michigan and Oregon, it’s common to pay a deposit on beverages, only to get that deposit back upon return. I’ve written in the past about how, as a grocery store employee, this law was the bane of my existence. But it would have been even worse a few years earlier, when reverse vending machines weren’t in wide use in the United States. (Imagine how much labor corralling literally thousands of sticky, used soda bottles each day involved.) Since then, the technology has improved, and while it’s not cheap, it certainly makes managing all these bottles and cans a lot easier.

The problem I think that critics of self-checkouts are seeing is in many ways the same one that people who use in-flight entertainment see. We live in a world where we have computers and phones significantly more powerful than any onboard entertainment system, but it’s a side effect of the fact that building a machine for hundreds of people to use each day is simply more expensive and harder to upgrade than a machine designed for just one person.

These days, we fail to look at enterprise technology in enterprise terms. You can get a good tablet for $300, but a machine that scans in your goods and plugs into the data-consuming beast costs way more and needs to be more carefully managed. We keep forgetting that our innovations started out as mainframes, and in some cases, mainframe derivatives still make sense. And innovations will gradually appear. One example: Just as with trains, retailers could see big advantages by replacing barcodes with RFID on individual products, something Uniqlo is doing. Likewise, replacing price tags in aisles with electronic shelf labels has lots of advantages.

I understand the motivation for technoskeptics to want to ditch self-checkouts. On the surface, they feel like Enshittification 1.0—a design built for the bottom line. There’s a case for more human interaction in our lives—as well as a desire to balance useful with creepy.

But I think that the problem is that we think of self-checkouts as a complete replacement for humans, when in reality, they’re best embraced as a utility player. Think of it like game theory: If you have fewer than 10 items, if you’re not grabbing a ton of exotic or oddly labeled products, and if the human interaction does more harm than good at this specific moment, you’re better off with self-checkout. If you want something weird, you have three carts full of stuff, or you just want a friendly conversation, you might want to stick with a traditional checkout.

In a way, the problem isn’t that we’re replacing one with the other. The problem is that, on both the consumer and retailer side, we’re not seeing the value of both.

This is not a binary choice here—and it never should be. It’s the retail version of Mac vs. PC. Give people the option.