Predicting Customer Behavior
I was watching a couple of talking heads hotly debate ObamaCare the other night on a news program. What struck me more than anything else was that neither speaker had much of a clue about healthcare, much of a command of the issues, or, frankly, much concern for the facts.
I know that
facts can be stubborn things. And it can become pretty obvious that some people
don’t like them because they interfere with one narrative or another.
Years ago, in my
last stint in a newsroom, I was covering a pretty contentious issue. Seems the
local power company wanted to divert water from a nearby river to its nuclear
power plant about 20 miles away in order to provide water for the plant’s
cooling towers. The plan would have allowed the utility to produce and sell more
nuclear energy from the plant.
Supporters and
opponents of the plan lined up along the usual fault line: on one side, college
kids, liberals, greens, and suburbanites opposing the plan; on the other,
businesses, the Chamber of Commerce, and lower-income people hoping for lower
heating bills.
The plan’s
opponents were able to get a referendum on the issue on the November ballot.
The station’s GM asked me to conduct a survey of voters so we could predict whether
the referendum would succeed or not. Most people thought the vote would support
the water diversion plan.
With little more
sophistication than pencil and some copy paper I conducted a survey from the
newsroom. My numbers showed that the voters would kill the power plan by a
pretty comfortable margin. We ran with the story.
Because of the
predicted margin of victory for the plan’s opponents, we had other news
organizations contact us. The survey had become
the story.
Some reporters ran
with the story: Local survey shows
residents oppose power plan. But others walked away from it because it
contradicted what had been their story line—that the opposition to the power
plan was being fed largely by out-of-town celebrity protesters. And that was
good enough for them.
By 8:00 on
election night it wasn’t good enough. Voters approved the referendum that killed
the water diversion project. Unstated anywhere outside the newsroom was the
fact that the large percentage of voters turning thumbs down on the power plan
closely matched the percentage of surveyed voters who said they would approve killing
the project.
We went from a
small, backwater broadcasting outlet to one that had outfoxed and outworked our
city slicker cousins. It was a lesson I never forgot.
Today in
business we use surveying to uncover the facts on which we base our conclusions
for clients. For example, we survey retailers on their level of electronic
payments sophistication in order to predict their ability and willingness to
adopt new forms of electronic payments like electronic benefits for WIC, a
government-sponsored nutrition program for young families.
We’ve also
survey state agencies to determine what effect, if any, pending regulation will
have on our customer’s clients. Those regulations could have severe adverse
consequences for companies that supply those agencies with technology.
In all these
cases, advance surveying yields tactical, actionable data that can be filtered,
modeled, and hypothesized in order to predict behavior.
The next time
you’re developing a business model regardless of what it is, consider
surveying. Without it you’re an attack column without scouts.
Don’t take my
word for it. Take the word of your target customers!