This year, the political campaigns are turning to what's known as "micro-targeting" -- like the list brokers who know what kind of car you drive, and what you eat for dinner. The campaigns are now using that kind of information to predict how you'll vote.
When you shop, ever use a grocery discount card to get a few dollars off? The store is recording what you eat. Every time you use your credit card, more information is collected on what you buy.
"It's the kind of automobiles they own, it's whether or not they go to the movies, whether they go to the theater," said Jack Noonan from Chicago's SPSS Inc.
Noonan's company makes software that can take that consumer information and combine it with census data such as where you live, how old you are, whether you're married or single and how often you vote.
There are hundreds of pieces of information that added together form a portrait.
"All of those things are behavioral characteristics that are important to predicting what someone will do," said Noonan.
That's what Noonan's software does, it make predictions. It can't predict when you'll be in an automobile accident, but it can predict if you're lying about that accident to your insurance company.
Here's how it works: when you report an accident to your insurance company you have to answer a lot of questions.
"Where it happened, how it happened, the people involved," said Noonan.
Noonan's company collected all the information they could from accident claims that were known to be fraudulent. When they compiled all those bits of information, they found patterns -- ways in which people lie when they're trying to cheat and insurance company.
Noonan's company created computer software to spot those patterns.
Now when you submit a claim to your insurance company, there's a good chance it's scanned by Noonan's software.
"We can immediately from the data that's collected from that transaction predict whether or not that transaction is fraudulent," said Noonan.
All sorts of industries have been using this analytic predictive software, but it didn't really make it big in the political world until the start of this decade.
"A little bit in 2000, much more aggressively in 2004 and in the 2006 cycle my firm was active in more than 100 different races," said Ken Strasma, president of Washington, D.C.'s Strategic Telemetry.
Strasma's company is working for Barack Obama and using Noonan's predictive analytic software to identify undecided voters.
"And it turns out there are definitely groups that are more likely to be undecided than other groups, more likely to support a particular candidate than others," said Strasma.
Strasma says there isn't any one thing or group of things that you eat, or listen to, read, watch or drive that says how you're likely to vote. But hundreds of indicators all taken together, do show very useful patterns.
For example, a large percentage of country club Republicans supported Mitt Romney's run for the White House.
Using computer modeling, Romney's campaign was able to identify a whole new group of voters who couldn't afford to belong to the country club, but were overspending their credit cards, in an attempt to emulate the lifestyle.
The software predicted those strivers would also support Romney, and that proved to be largely true.
"So you have to look at people's fears and aspirations, not just where they currently stand," said Strasma.
Four years ago, Strasma's company was working for John Kerry.
"And we were able to focus on which precinct in the Iowa caucuses to focus on most heavily," said Strasma.
By focusing his message on undecided voters who were likely to support him, John Kerry scored a surprise upset over Howard Dean.
"Which we'd actually seen coming a fair amount ahead of time," said Strasma.
This past January, history repeated itself as Barack Obama, with help from Strasma's company, pulled another Iowa surprise -- though Strasma isn't willing to talk about that until after November.
"It's not a secret weapon if we talk about it," said Strasma.
ABC7 News asked Strasma and Noonan if they ever have second thoughts about how much of our personal information is being gathered and manipulated.
"Well you can only see what people allow you to see," said Noonan.
"People always have the option of refusing to answer," said Strasma.
But most of the time, most of us don't refuse. We sign up for the club card to save a few dollars. We don't notice the opt-out box on the computer screen, when the clerk asks for our zip code. We don't think that little bit of information means that much.