Business Tech: The Point (of Sale)
The saloon doors swing open. The dark-haired stranger takes a mosey up to the bar. "Shot of whiskey. Leave the bottle," he says. And, through some lost and arcane magic, the bartender knows the price of a shot, the price of a bottle, how many bottles he still has, and can calculate change from a quarter. Fast forward to today... sigh.
Warning: this article will start with whining and complaining, but we will get to the tech part. Promise .
The Candy Incident
I'm not saying that Old West barkeeps had access to better education; I'm just saying that we run into a lot of cashiers who can't do math or understand pricing. Point-of-sale systems compensate by doing much of the work for the modern register operator. There is, however, a limit to technology.
For example, Syd and I walked into a <company name redacted> to buy some junk food to bring over to a friend's house. The cookies were two for six dollars. The candy was three for four dollars. Simple, right?
Syd knows it isn't. So, she tells the cashier her expectations, "These cookies are on sale. So are these candies." She even specifies the prices from the signs. That would earn you a squinty-eyed stare from the olde-timey barkeep, but for the young man in front of us, it's not uncommon.
"Cookies aren't on sale," he informs us.
We show him the sign. He was twice wrong. In addition to the sign saying the cookies were six dollars for two, was a sign saying the cookies were five dollars for two. The cookies were simultaneously on two different sales. Schrödinger would be pleased. He gets a manager's permission because direct proof is no longer enough in our technologically-advanced world of selling cookies. I miss that barkeep.
"This isn't the candy on sale," he says.
I take a picture of the sign so that he doesn't have to walk again. Where I'm taking the picture, I am in full view from the register. No chance of flim-flam. He still goes to verify. He comes back with three bags of jelly beans. Syd and I smile at each other knowingly. We've seen this before.
He's going to ring up three bags of a candy which will meet the price on the sign INSTEAD of ringing up the ones we bought.
Our tech-savvy brains work through the problem with this approach: Inventory. That point-of-sale system he's using is pretty much guaranteed as having inventory-update tech. So, he's now told the system that three bags of jelly beans are going out, not gummy bears, spearmint leaves, and spice drops. The inventory sub-system will receive incorrect details about on-hand and about buying patterns. By trying to avoid bothering his manager again, he's accidentally — training issue? — wrecking the back-end data.
Turns out, both Syd and I were wrong. He's not doing a swap to get the prices right, he's decided that if we want three for four dollars; he's going to give us the candy which matches the price, not the candy we brought up to the register. This is a different training issue.
We make him undo it. He calls the manager and we show her the sign. For those of you worried about our finances: We got all our sale prices. For those worried about customer service: We found out that the sale price of the candy was only seventeen cents a bag cheaper. We lost twenty minutes getting a fifty-one cent discount because the sale price made us think the savings was meaningful. I'm sure the deal on the cookies wasn't much better.
It took a manager and a cashier to ring us up. Those sale signs cost the company twenty minutes of two different salaries. As a bonus, the same signs had the added benefit of creating delays for the customer and a sour experience on a trip to buy sweets.
The Tech We Deserve
What if the point-of-sale software allowed staff to scan in the discount signs? The register start-up could include a list of old signs to take down and a list of new signs to put up. With the signs scanned in, the sales would all be correctly registered in the... um.. register. Less price confusion. I know this store didn't have that tech because the sales signs didn't have barcodes or QR codes.
What if the register had a motion sensor pointed at the line. When there's no one there, it could run training videos. Our befuddled operator could have learned how to up-sell. He could have learned basic customer service. He could have learned the inventory-based implications of how he scans.
Or maybe, the register software should have some DSS (Decision Support Software) installed which would alert the manager that a pattern of add-cancel-add was going on, suggesting a cashier-in-distress or a customer-near-meltdown.
That same DSS module could help with the up-sell, too. Amazon popularized "People who bought this also bought" using that sort of data-mining. And, by tying in the inventory, we can make sure the algorithm only suggests things we have in stock. Likewise, it could "predict" discounts: If you bought one more, the sale price would trigger, saving you a dollar. And, we could tailor the coupons on the back of the receipt, too.
Let's add a bonus idea: If the upper left-hand corner of the point-of-sale screen showed the front door camera, the register operator could tell, even while watching a training video, when the store was getting new shoppers. That would allow him to anticipate when he might be getting busier. Or, he might signal the manager that a second register needs to open.
It is easy to think of point-of-sale as "known territory," but there is still room for pioneers to take it further.
We used to assume that people had to be smart enough to do the job when we hire them. Using tech that covers for failings but doesn't equip the capable is a sad commentary.
Sadder still, self-serve checkout is making a strong showing in a lot of stores. The essential message of self-serve: Even an untrained customer is better at the job than a - supposedly - trained employee. If you need me, I'll be the one drinking a sarsaparilla at the bar, next Spectrum Conference.