Evan Schuman's StorefrontBacktalk

With the many new self-checkout offerings being introduced this week from the likes of IBM, NCR and Fujitsu, it’s not a bad idea to focus on what will truly decide whether these machines do anything to help retailers.

To state the obvious: It’s getting consumers to use them. I say it’s obvious, but one wouldn’t guess that based on what the vendors were saying this week.

From consumers I’ve watched and talked with, the hesitation comes from a lack of trust. They don’t trust themselves to run the machine correctly and they don’t trust the retailer that it’s truly a benefit (suggestions to offer consumers an X percent discount for using self-checkout have proven to be as popular as Jerry Yang at a Yahoo Daytraders Convention).

But there’s another trust issue circulating that may prove to be self-checkout’s downfall: Consumers want to trust that the store won’t have them arrested if the self-checkout machine fails or if they inadvertently forget to scan something.

Initially, these appeared to be a series of unrelated little crime stories, but they’ve been getting more numerous lately. No conspiracy theory needed, as these reports are getting more common as self-checkout deployments increase and the acceptance of these devices creeps upward.

That said, consumers are starting to hear about such stories not so infrequently. This is another of our favorite "reality versus perception" stories, where perception is far more important because it can shape reality. Without commenting on U.S. politics, it’s safe to say that perception changes reality a lot more often than reality changes perceptions.

So let’s look at the perceptions/realities behind these headlines. Perception: self-checkout is more attractive to shoplifters than a staffed aisle. Reality: The opposite. Between video cameras, spot-checks and CPUs that don’t get tired or lenient, most smart veteran shoplifters would much rather take their chances with the yawning 16-year-old working aisle 10.

But there are a lot of stupid amateur shoplifters who will try and outfox the machine. Another reality is that retailers must maintain a strict zero-tolerance policy on shoplifting.

Therein lies the conundrum. When someone sticks memory sticks or caviar in his pockets or down his pants, it’s not too hard to conclude unlawful intent. But when someone leaves a package of toilet tissue or a crate of oranges beneath the cart, was that criminal or merely unobservant or forgetful? If the intent was criminal, would a crook be that obvious?

Let’s take it one step further. What if the customer is trying to checkout, but the machine glitches? Here’s a small story that has me worried. I have no idea whether the accused consumer in this Massachusetts case is innocent. The point is that, for an IT manager responsible for self-checkout, it likely doesn’t matter.

If a customer argues that an alleged shoplifting was merely the result of a malfunctioning self-checkout system (which is hardly an outlandish scenario), there’s an awful lot at stake beyond that stolen crate of tomatoes. Is the customer offering to pay for the forgotten goods right then and there? Assume, for the moment, that the customer is telling the truth. Do you have any firm evidence (such as video surveillance) to contradict it?

A few weeks ago, I was in a Circuit City doing a little buy online/pick up in store. As I was about to leave, the alarm rang for the customer ahead of me. An employee waved the person out the door. Perplexed, I tried leaving and it did the same for me. When he waved me on through, I asked what was going on and he said, "The system does this all the time. It’s fine." I asked him if he wanted to check my bag. He said "no thanks."

These are the two extreme ends of a very wrong retail security situation. In the first, we have retailers choosing to believe a self-checkout system’s reliability and accuracy over a customer’s word, failing to seriously consider the possibility of computer error. In the second, the employee assumes computer error and concludes that every stolen merchandise alert is false.

The cost of the second extreme is obvious, but what about the first? It’s not merely a matter of losing that initial customer and their friends. It’s about losing thousands of self-checkout customers who are much less trusting of the software and who would rather be believed in the absence of clear and convincing evidence to the contrary.

This is by no means limited to a loss prevention conversation. If you want consumers to further embrace self-checkout (don’t kid yourself: You’ve thus far only gotten the low-hanging fruit), you must convince them that you’re on their side and that self-checkout isn’t frightening. The system won’t slow consumers down, it won’t charge them more and, most critically, it won’t get them falsely thrown in jail. Speaking for myself, any grocery store that doesn’t pretty aggressively avoid that last one isn’t going to get my egg purchase.