How much do you know about your customers? If you're like most businesses the answer is "not enough," according to Alan Trefler, founder and CEO of business software provider Pegasystems and author of Build for Change. Trefler is on a one-man mission to get businesses to stop collecting data and make better use of the data they already have.

"Data only tells you who the customers are and what they did in the past," he says. In order for data to be useful, you need to understand customer intent--on why they do or don't buy from you. And for that, he says, you likely already have all the data you need.

One big mistake many business leaders make is to draw conclusions based on correlations alone. That is, if you put an ad in the local paper and sales went up, it may be because of the ad. But it could also be that your product is one that's most used in fall, and the change in weather brought customers to you. Without understanding customer intent, correlations can lead you astray. "The divorce rate is nearly perfectly correlated with consumption of margarine," he notes.

To get the best information out of your data, he says, follow these three steps:

Step 1: Create customer personas.

Who, in general terms, are your customers? Are they suburban dads who tinker on the weekends? Career-driven mothers? High-school athletes? Perhaps all three?

Your customer base can't be everyone, so Trefler recommends creating "personas"--hypothetical characters who represent a typical customer for you. "See if you can group your entire customer base into five or six fictional people," Trefler says. "Tom, the weekend athlete business man. Sue, the super-aggressive executive. For each of these, envision an actual made-up person, like a character in a play. What types of things would drive that person to make a purchasing decision? What things would they find insulting or demeaning?"

This is a thought exercise you should be continually revising, he says. Just as important--are there personas who aren't buying from you, but should be?

Step 2: Form some hypotheses.

Once you've created your personas, make some educated guesses about how each one is likely to react to different things you might try. For instance, if one of your personas lives in the suburbs and commutes for work, you might think that radio advertising during morning and evening drive times is an effective way to reach that person. If another persona is very focused on family, then tying your marketing efforts to holidays might be a smart approach.

Step 3: Test your hypotheses.

Begin by comparing your hypothesis against existing data. You likely have detailed sales records and know which times you did better and worse. You also should have detailed records of promotions you tried and what advertising you ran when. You should know when you adjusted prices up or down and what impact that had on sales. "Most small companies do have a lot of data at their disposal," Trefler says. "It's important to tie that data to what you're trying to accomplish."

Next, test your hypothesis in the real world. Does your data suggest that a slight price drop will lead to a sharp sales upturn? Then try dropping prices in a limited-time offer without changing anything else, and see if that idea holds true. If it does, you've learned a valuable lesson about how price-sensitive your market is.

The more you create and test hypotheses this way, the better you'll understand your customers and how to reach them, Trefler says. And you can get to a pretty sophisticated understanding of your market without using any analysis tool beyond a simple spreadsheet, he adds.

"The big mistake most companies make is they focus on data without first forming a hypothesis," he says. "You need a different way of thinking."