The numbers don't lie -- or do they?

In the case of annual U.S. ATM shipment numbers, it all depends upon whom you ask.

ATM&Debit News, one of two industry trade publications that produces annual shipment numbers, just published its 2002 survey of manufacturers, which showed a 3.5 percent increase from the 2001 numbers.

Ann All

(ATM&Debit News generously allowed us to reprint the survey on ATMmarketplace)

David Gosnell, the senior editor of ATM&Debit News and the man responsible for gathering the numbers, said he largely relies on the honor system.

"I often tell people I am not the ATM police. I'm depending on the companies to represent their ATM shipments accurately," he said.

Truth, justice and the ATM way

Therein lies the problem, if you listen to some industry pundits.

"People lie," said a market analyst for one manufacturer.

"Just between you and me, we think the numbers (reported by a competitor) are $%?*," said a director of marketing for another manufacturer.

To listen to one executive at yet another manufacturer (all of whom wished to remain unnamed), any numbers ending in zero are suspect. "Oh come on, how often is it going to be a nice, round number like that?" he said.

All of those quoted above insist that their own numbers are pristine, of course. And despite their skepticism, they seemingly can't wait for the numbers to come out every year -- if for no other reason than to level their scorn at some new targets as well as the old favorites.

In my opinion, manufacturers are about as exaggeration-prone as ATM deployers. Generally speaking, that means I take 10 percent to 30 percent off the top when someone gives me a number. The subtraction occurs only in my head, however. Like Gosnell, I go with the figures I'm given -- unless I can substantiate claims to the contrary.

I suspect this won't come as a shock to the majority of those in the ATM business. At a recent trade show I attended, some sages in the hotel bar were offering their advice to neophyte deployers. My favorite bit of wisdom: "Multiply the number of machines you have by five and tell everyone you know." (This earned an appreciative round of laughter and, I believe, drinks.)

Test of time

While it would be unrealistic to expect the numbers to be exact, Gosnell said he believes they provide an accurate portrayal of the state of the U.S. ATM industry -- especially when they are reviewed over a period of several years.

"I think most of the companies represent themselves accurately. For those that don't, I think it catches up with them in time," he said. "In the long run, I think it helps give everyone an independent look at where the ATM industry stands and where it is going."

I agree with Gosnell. In fact, I think the aberrations themselves can be quite telling.

While everyone is bemoaning the "flat" numbers of the last few years, the truth is the market began leveling out before then. It was at least partially obscured by the inflated shipment numbers reported by manufacturers doing business with Credit Card Center (CCC), the ISO that went belly up in June of 2001 after taking delivery of thousands of ATMs that never got installed.

In fact, it's obvious that the numbers for 2000 would have been very different if "the BPI factor" -- bought, paid for and installed -- had been somehow accounted for in that survey.

Gosnell's job is made even more difficult because several companies, including Diebold and NCR, no longer release their numbers. In those cases, Gosnell looks for clues in their financial statements and interviews financial analysts who cover those companies to come up with his estimates. Then he runs those numbers by the manufacturers and, in some cases, what they tell him leads him to revise his estimates.

Both Diebold and NCR have told Gosnell that because their businesses are increasingly global, the U.S. numbers alone don't provide an accurate "big picture." Similarly, they've told me that part of the reason they stopped providing shipments is because they felt it was misleading if sales of software, services and upgrades were not included.

While those arguments have some merit, I tend to agree with an executive (again, who prefers to remain nameless) at a software company, who left his marketing post at a major manufacturer in 2002. "(The numbers) are really all we have to go on," he said. "They do at least give us some kind of a benchmark."