A letter from Wells Fargo Bank pitching a debt consolidation loan to my wife was so off base it could only have been generated by a computer.

No actual banker could have looked at my wife's accounts and concluded that she needs a loan. She's as conservative as Ben Franklin, with saving and spending habits formed back when she was a commercial real estate broker and wanted her money to last through the next closing.

That letter had to have been just another data-driven marketing pitch, a reminder of how big companies try to sell based on all of the data they have amassed on us and our behavior. And in fairness to that Wells computer, there's a chance it was onto something — not that having that kind of capability should make any customer sleep better.

Ravi Bapna, a professor of information and decision science at the University of Minnesota's Carlson School of Management, said banks don't even use information that might seem logical, like a debt-to-home equity ratio, in these kinds of pitches. He suspects Wells is analyzing detailed demographic information, data from past purchases and other behavior and comparing it with data from consumers just like us.

And Wells Fargo is far from the most sophisticated in its industry, Bapna said, trailing the likes of Capital One Financial. Amazon.com is masterful at this sort of analytical work, too, and Minneapolis-based Target is also pretty sophisticated.

New York Times reporter and author Charles Duhigg described in his book, "The Power of Habit: Why We Do What We Do in Life and Business," how a Target data analyst named Andrew Pole figured out how to determine when a Target customer is pregnant.

Target carefully tracks its customers' purchases, of course, assigning each a unique guest identification number. Pole was asked if he could find out from this great mass of data when a customer was pregnant. A big life event like the birth of a child can knock consumers out of their previous buying habits, and Target could be way ahead of competitors who learn of a new baby through public birth records.

It wasn't that hard.

Women apparently buy larger quantities of unscented skin lotion about a third of the way through their pregnancy. Cotton balls are pretty common, but when somebody starts buying really big bags of them, they are very likely approaching their due date.

Target marketers soon had an analytical model that predicted pregnancy based on purchases of about 25 products.

Target found it could easily overdo any personalized marketing, though. Duhigg wrote of a dad who angrily asked Target to stop sending baby product coupons to his teenage daughter. He sheepishly admitted days later, however, that perhaps Target knew more about what was happening in his house than he did.

A Target spokeswoman said it's never been able to confirm that this dad and daughter episode ever happened. She added that Target values what its customers say and "following guest feedback to the story, we discontinued that test" and no longer tries to guess a person's "life stage."

Target's view is that it's only trying to help its customers by offering deals on products it's already figured out the customer would want, and that's entirely sensible. Yet it's easy to see how some customers would be uncomfortable with how much Target knows about them, especially since December, when Target said it lost payment account data of millions of American consumers.

Target is far from alone in being shy about its data capabilities. Those executives who do talk tend to do so to portray themselves as champions of consumer privacy.

"When our engineers work tirelessly to improve security," wrote Facebook co-founder Mark Zuckerberg in a recent post complaining about the National Security Agency, "we imagine we're protecting you against criminals, not our own government."

He doesn't discuss who is protecting us from him. The reason Facebook is even a business, one worth more than $150 billion, is its tireless work to commercially exploit the personal information of its users.

The question about what consumer behavior the big companies really track closely is an interesting one, because there's at least a chance that Wells was on to something when it pitched a debt consolidation loan.

It's true my wife is as financially sensible as they come, but it's been a bit of a rocky winter for the family's chief financial officer. A Target Redcard bill was among those that went missing for a while.

A root cause analysis of these failures identified a critical period in the process as any afternoon between 2 and 3 p.m. — when the high school kids get home. If teenagers get to the mailbox before the bill-paying adult, it can mean that weeks later the January Verizon Wireless bill suddenly reappears from underneath an AP Statistics textbook.

Paranoia can't be ruled out here, but is it possible that Wells Fargo knew of this handful of lapses and concluded a loan might be helpful? Wells did not get back with someone who could discuss it, but if its analytical ability is that finely tuned, then like Target it had better disguise what it knows before its marketing veers off into the land of the creepy.

Avoiding pitches that spook customers is one of the things Ravi Bapna said has to get better about the use of customer data, along with what he called the "stupid" ones, such as the daily ads for an event that follow the online purchase of a ticket to it.

He added that younger people fret less than middle-aged ones about what big companies know about them. Younger folks just want the data-generated offers to be personal and interesting.

Bapna said his daughter just got a new Android tablet. The first thing that the Google Play online store recommended for purchase was a recording of a song that had been searched for on YouTube the previous evening on the family computer.

"Yeah, Dad," his daughter told him, clearly pleased. "Google got it right."

Seems creepy to me.