(Reposted from timberry.bplans.com. In the summer of 1979, after we had moved from Mexico to Escondido Village on campus at Stanford, while waiting to start the MBA program, I worked with Creative Strategies International as a contract consultant. My last Mexico employer, Business International, owned Creative Strategies and helped me get the contract work.)
In 1979 Creative Strategies International, a high-tech market research company, assigned me the job of forecasting the market for automated teller machines in the U.S.
There were a very few of them already in place at the time. Maybe a few hundred.
I got the numbers as best I could. I found out how many banks there were, and how many branches. I got data for growth in banking customers and growth in branches. I got data for employees in banks.
I talked to dozens of experts. Product managers in companies making ATMs, or companies that could possibly be making them. Lots of bankers, lots of consultants to bankers, several journalists involved in bank-specific trade magazines.
Not all of them wanted to talk to me, but an older and more experienced vice president in the same firm (Tom Arnett was his name) had some good advice (and I’m paraphrasing here, it’s been 30 years):
They have to be interested in what they do, the market they’re in, or they couldn’t get up in the morning.
So hook them in fast. Tell them you’re forecasting the market for Creative Strategies. Tell them you’re thinking the market is going to grow at some percent — it doesn’t matter — but most experts disagree. Make it clear as quickly as possible that you’re going to offer opinions and information as part of the conversation, if they have time to talk to you.
Most of them will. They want to feel like experts. They want to be asked. And they want to know what you’re thinking too.
I used Tom’s advice a lot, and talked to a few dozen experts.
In the end, though, there was not technical or mathematical way to forecast ATMs. At least I was able to relate the projection to numbers of branches to give me some sense of error check, but it wasn’t clear that ATMs would all be in bank branches.
What made the biggest difference to that forecast was the placement of two ATMs at Stanford Shopping Center on the side of the Bank of America branch there. We lived in grad student family housing at the time, and we would ride our bikes over to the shopping center. The ATM was very convenient. It gave me cash fast. It gave me cash after banking hours. I used it a lot.
Most of the bankers told me people would never accept doing business with a machine. They’ll never warm up to that.
Happily, I believed what I saw instead of what the bankers told me. I projected a very fast growth rate for ATMs. As the years ticked off, it turned out I was very close. The forecast I made in 1979 gave a relatively accurate view of the future, given how much uncertainty was there in the system.
Take note, however: it was a human educated guess. The math helped me to compare my projection to the numbers of bank branches, but that was just a reality check. I was guessing.