Automation is often presented as an inexorably advancing force, whether it’s ushering in a threat to jobs or a promise of increased leisure or larger profits. We’re made to imagine the robots rising, increasingly mechanised systems of production, more streamlined modes of everyday living. But the truth is that automation technology and automated systems very often fail. And even when they do, they nonetheless frequently wind up stranded in our lives.
For every automated appliance or system that actually makes performing a task easier — dishwashers, ATMs, robotic factory arms, say—there seems to be another one—self-checkout kiosks, automated phone menus, mass email marketing—that actively makes our lives worse.
I’ve taken to calling this second category, simply, shitty automation.
Shitty automation usually, but not always, comes about when new user-facing technology is adopted by a company or institution for the ostensible reason of minimising labour and cutting costs. Nobody likes wading through an interminable phone menu to try to address a suspect charge on a phone bill—literally, everyone would rather speak with a customer service rep. But that’s the system we’re stuck with because a corporation decided that the inconvenience to the user is well worth the savings in labour costs.
That’s just one example. But it gets at what makes spending some time wading through the world of shitty automation worthwhile—it often doesn’t even matter if automation improves anything at all for the customer, for the user, for anyone. If some enterprise solutions pitchman or government contractor can sell the top brass on the idea that a half-baked bit of automation will save it some money, the cashier, clerk, call center employee might be replaced by ill-functioning machinery, or see their hours cut to make space for it, the users will be made to suffer through garbage interfaces that make that waste hours of their day and make them want to hellscream into the receiver—and no one wins. Not even, sometimes, the company or organisation seeking the savings, which can suffer reputational damage.
So! Join me, as I begin a foray into the world of shitty automation. Where does it come from? What was the pitch that got it there? Why did it stick—and why does it persist, even when it’s making matters actively worse? And, maybe most importantly, what is it making us—as workers, consumers, automaton-cohabitors—do?
To start, let’s look at everyone’s favourite cluster of machinery to walk past in the grocery store with a dismissive scowl, to hold off approaching until you’ve finally, painfully decided the line you’ve been stuck is so painfully not-moving it’s worth the hassle: Self-checkout kiosks.
There are fewer better poster children for shitty automation than self-checkout. 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. And it’s not because I’m an idiot. Or not entirely, anyway. Incessant, erroneous repetitions of “please place your item in the bag” and “unknown item in the bagging area” are among the most-loathed phrases in the 21st century lexicon for a reason, and that reason is that self-checkout is categorically awful.
Hence, I turned to Alexandra Mateescu, an ethnographer and researcher at Data & Society, and a co-author, with Madeleine Clare Elish, of “AI in Context: The Labour of Integrating New Technologies,” which uses self-checkout as a case study, to find out why.
To understand how we arrived at our current self-checkout limbo, and why it’s terrible and dysfunctional in the special way that it is, it helps to understand that the technology we encounter in the grocery store is just the most recent iteration in a century-long drive to offload more of the work involved in the shopping process onto us, the shoppers.
“We as consumers have been acculturated to the concept of ‘self-service’ for decades, so doing labour that used to be done by workers has largely been naturalized,” Mateescu tells me.
For most of the first half of the 20th Century, American shoppers would show up to the market, give the clerks a list of what they wanted, and wait while the professional fetched the goods. But in 1916, a businessman named Clarence Saunders opened the first self-service supermarket: Piggly Wiggly. Staffing all those clerks were cutting into supermarket profit margins, and letting shoppers find the items themselves, Saunders figured, would bring costs down. This was sold to consumers as offering more “freedom”—Mateescu notes that one Piggly Wiggly ad, depicting a woman with a shopping bag, carried the tagline, “A nation-wide vogue in shopping that leaves women free to choose for themselves”—along with a promise of passing on some of the savings to shoppers. But it was the opportunity to significantly reduce labour costs that no doubt led supermarket chains to follow suit.
It took a couple of decades to become widely popular, but the concept did take off, driving a wave of competitors to adopt self-service and eventually standardising the industry.
“Self-service—where shoppers could actually walk around the aisles and pick items themselves—was only popularised in the 1930s and 1940s and was framed as a liberating experience,” Mateescu says, and the trend spread beyond the grocery store. “This goes beyond the supermarket context—as consumers, we do all sorts of unpaid labour for companies in the form of self-service, often replacing what we used to expect in the form of customer service.”
In the midst of that self-service boom, Saunders tried to go a step further, and introduced an early foray into shitty automation by way of a fully-automated grocery store, complete with self-checkout, in 1937. It was called Keedoozle, and it was a spectacular flop—in part because the technology rarely worked properly, and human laborers constantly had to help customers deal with it, and in part, presumably, because it was called Keedoozle.
Despite a number of further steps towards automation, like barcode scanners, it wouldn’t be until the turn of the century when major grocers would again attempt some degree of Keedoozlificaiton. Around the early 00s, prompted by improvements in computer systems and scanning technology, a number of companies began pitches the supermarket chains on self-checkout. Chief among them were the National Cash Register (NCR) Corporation, Optimal Robotics, and Productivity Solutions Inc (PSI), which would become a subsidiary of IBM.
According to a 2009 study by Christopher K. Andrews, the author of The Overworked Consumer: Self-Checkouts, Supermarkets, and the Do-It-Yourself Economy and an assistant professor of sociology at Drew University, NCR claimed that each of its self-checkout lanes, which cost around $US30,000 ($42,571) each, would “pay for itself in twelve to eighteen months.” Similarly, PSI “advertises that its machines save businesses up to $US225,000 ($319,285) a year.” The cost savings, of course, were to come from automating people out of work.
“By automating labour, self-checkouts may allow businesses to replace cashiers with machines, and thus shed significant labour costs... Most of the self-checkout manufacturers’ websites readily acknowledge their savings in labour-related costs; a report on NCR’s website states that ‘self-checkout…allows stores to cut labour costs, which account for more than ninety per cent of the costs associated with running the front end of a retail store.’ Likewise, Optimal Robotics notes that a four-station, one-attendant configuration would require approximately one hundred and fifty fewer labour hours a week compared to the regular checkouts…”
Just to reemphasize, the pitch from the automated checkout makers was and is all about labour savings and much less about any perceptible improvements for customers. The automated checkout companies do try to nod to some consumer benefits, but it’s a distant and deeply secondary tier of the sell. “While the manufacturers also promote consumer advantages that may indirectly affect businesses‚ most commonly shorter lines and faster checkouts—the main selling point is lower labour costs,” Andrews writes.
“This is the pretty straightforward ‘sell’ laid out in all the marketing material from vendors—‘invest in these machines, and get the savings back and more by trimming payroll,’” Mateescu notes. And, as Andrews adds in another paper, “growing pressure from low-wage nonunion competitors (e.g., WalMart) may cause supermarkets to consider expanding automation beyond current levels, thereby significantly affecting current employment patterns.”
So, the retailers and grocery chains bit. Kmart, Walmart, and a number of other major supermarkets began installing self-checkout systems in the late 90s and early 00s, to mixed success. At best. Actually, people mostly ignored or hated them, for the same reasons you probably still hate them today—items wouldn’t scan, sensors went haywire, and a cashier or supervisor had to be called over to reset the machine or enter some arcane code every other item. Home Depot famously overcommitted to self-checkout and pulled them out after losing business to Lowe’s.
Since then, things have improved, but barely.
“The number of self-checkout machines in stores largely leveled out, and since then different retailers have periodically rolled out and scaled back on their use,” Mateescu tells me. IKEA, for instance deployed self-checkout but got rid of them in 2012.
They’re just kind of there in most stores—the average chain that has self-checkout aisles has four of them—taking up space and occasionally being endured by customers who don’t want to brave a longer line with a cashier operated by a human.
So what actually happens when automated checkout hits a grocery store? First, the store’s management removes some of the old human-operated checkout aisles to make room for the new automated machinery, and workers, often more experienced cashiers, are assigned to oversee the area.
Then, they experience the stress of having to oversee (and try to get people to use) a cumbersome and unpopular technology. As Mateescu tells me:
In our fieldwork, we found that frontline cashiers often felt understaffed and overwhelmed by long lines. But instead of increasing staffing, store management kept increasing the quotas for how many shoppers they wanted to go through self-checkout. This puts workers in a bind, because now they are pressured to funnel more and more reluctant shoppers towards the self-checkout machines. So there’s a lot of human effort to facilitate the “success” of these systems, what we call “human infrastructures” in our report.
At the same time, self-checkout machines can’t be left alone to shoppers. An attendant plays a role similar to that of a traffic officer at a busy intersection, multi-tasking across as many as six machines to help confused shoppers navigate the point of sales system, answer questions, and spot potential theft… So overall the impact is work intensification often exacerbated by understaffing, and not much acknowledgement that these reconfigurations often require new skills and responsibilities. These aren’t inevitable outcomes of these systems, though—they’re shaped by management decisions.
Now, whenever a shift like this (from a human-operated to an automated system) happens, and a job threatens to be made redundant, companies tend to say that no jobs will be lost, and instead, the move is beneficial to human workers. “In public statements about retail tech, companies typically say that systems like self-checkout or inventory-scanning robots free up workers’ time, which can then be redirected to do other necessary tasks in stores,” Mateescu says. “However, that’s not an inevitable outcome, since employers can choose instead to simply cut hours and make fewer employees do more of the work.”
This applies not just to self-checkout, but to nearly every conceivable automated system that threatens to disrupt retail jobs. The executive in charge of Amazon’s cashier-less Go stores, for instance, told the Times that the stores don’t eliminate jobs, just *change* them. “We’ve just put associates on different kinds of tasks where we think it adds to the customer experience,” she said. Yet each store requires only 3 to a maximum of 10 employees. When Walmart was displeased with me for portraying its calls for government assistance for workers it automated out of jobs as corporate welfare, their spokesperson repeatedly assured me that the nature of jobs was merely changing, and that employees would be given opportunities to do work that required a “human touch.” Yet they are firing all of their greeters, the only job they offer that is entirely reliant on said human touch.
The key point is that installing self-checkout is an opportunity for employers to reduce employee hours—these are rarely salaried workers—regardless of whether the system “works.” This is a phenomenon that Astra Taylor dubbed “fauxtomation”—whereupon the logic of automation is deployed to offer companies to slash wages, hours, benefits, or entire jobs, regardless of whether the system replacing them is actually up for the task (it often isn’t)—and it often arrives hand-in-hand with shitty automation.
So grocery store and retail employees are scrambling to make the “labour-saving” technologies “work,” at behest of management, which is well aware that many tens of thousands of dollars have been sunk into this technology, which everyone hates.
Employees who had been hired as cashiers have to learn how to do things that are not at all in their original job description, like repair malfunctioning machines, how to walk shoppers through the system, and how to soothe them when they became frustrated, lest the worker face management’s displeasure. They’re also subjected to experimentation with confusing and freshly shitty automated systems. In 2018, Walmart rolled out Scan and Go systems in 150 stores, forced employees to learn it, then dumped it a few months later.
All this, Mateescu says, is invisible labour; the kind of work that’s absolutely necessary, but isn’t recognised by management—and employees are rarely trained or compensated for.
“I’m in a lot of supermarkets around the country,” John Karolefski, who runs the blog Grocery Stories, told Vox’s Kaitlyn Tiffany last year in a funny piece about the myriad failings of self-checkout. “I watch people. I can tell you that I’ve been in stores where the lines that have cashiers are very, very long, and people are a little upset, and there are three or four self-checkout units open and nobody is using them… Wouldn’t the shopper be better served, customer service improved, if those weren’t there?... ‘Why do I want to scan my own groceries? Why do I want to bag my own groceries?’”
The answer is simple: We don’t. And consumers are starting to actively get fed up with the accumulating self-service labour, or the extra “shadow work,” that’s filling up their days. But if automated checkout is bad for consumers, bad for employees, and expensive to adopt, then it must be good for the corporations paying for the technology, right?
That’s the million dollar question. Beyond the steep cost of installation, there’s also the matter of licensing the software or paying to have in-house software developed. Then maintaining it. On top of that, theft rates at self-checkout is a lot higher than regular checkout. It’s so widespread that an entire subculture and attendant vernacular has developed around ripping them off. One study found that a truly crazy 20 per cent of people who’d used self-checkout admitted to stealing from them. Another found that theft led to 4 per cent losses, double its nonautomated counterpart. Which, of course, is why employees have to be brought over to keep a watchful eye over the “self” services, and do more work keeping tabs on shoppers.
If supermarket chains properly tabulated the theft and the invisible labour hours it takes to ensure self-checkout isn’t constantly malfunctioning, it might be a flat-out money-loser for corporations, on top of being enormously derided and making life worse for everyone that encounters them. So why do supermarkets keep them around?
“All in all they do de facto ‘work,’” Mateescu says. “Self-checkout is a good example of the difficulty of defining the “success” or “failure” of a technology without specifying for whom. They work for supermarket retailers up to a point, if they are able to achieve the aim of cost-cutting. Sunk costs might also be factor, since supermarket retail in particular does not have huge funds to invest in new technology to replace the old.”
In short, so as long as they allow the chain to justify reducing hours of its part-time employees, it’s probably worth it. “Supermarket retailers operate on extremely thin profit margins, so at the end of the day it can turn out that even after you tabulate all of these many downsides, there is a very small profit advantage that justifies their existence for retailers,” she says.
She also worries about the next wave of automated shopping—a la Amazon Go, which really might actually stand to “work” seamlessly—just at the cost of shoppers and workers subjecting themselves to corporate surveillance on a much wider scale. And going cashless, as Amazon Go mandates, disadvantages poorer customers who might not have credit cards or digital payment accounts. “So some of the clunkiness of automated checkout could, in theory, be solved with better technology,” Mateescu says, “but with so many tradeoffs it’s worth asking where it begins to be a solution in search of a problem.”
And that’s what this thoroughly shitty example of shitty automation boils down to. This is a service that was never intended to serve customers, but rather specifically (though clumsily) designed to cut or trim jobs—as such, it’s little surprise that it serves no one. It’s a cumbersome stain—if relatively benign—stain on our routine quest to obtain cereal and cucumbers. It is pointless-to-malignant in impact to us. It remains lodged in the far aisle of Ralph’s because it saves the chain just enough money to justify not taking it out. Yet it winnows employees’ hours who need them, justifies attrition, and serves to amplify the erosion of yet another precarious job in the retail sector.
The real rub is that self-checkout probably doesn’t *have* to be shitty.
“Self-checkout could have been better integrated into supermarkets in ways that augmented and better facilitated interactions between frontline workers and shoppers, but that design goal goes against the very reason they were adopted by retailers in the first place—to cut labour costs,” Mateescu says. “So the tool has a self-defeating logic to it from the get-go.”