Special delivery

Genalytics is trying to take the junk out of junk mail.

Using patented software capable of analyzing more than 500 variables per household, the Andover company predicts for its clients which homes are more or less likely to order a pizza, hire a maid, or purchase furniture.

Genalytics Inc. clients use the information to target their direct-mail advertising to those customers likely to be receptive to their message -- and bypass those who are likely to throw it in the trash.

Genalytics says the system isn't foolproof, but it cuts mailing costs significantly, while boosting customer response rates 7 to 15 percent. Company executives said it's not uncommon for Genalytics to double a company's return on its direct-mail investment.

"There's a big difference between someone who buys a pizza and someone who buys a variable annuity, but they both can be predicted," said Ray Kingman , chief executive of Genalytics.

Privately held Genalytics has financial backing from Egan Managed Capital.

For years, larger companies have used predictive modeling to tailor their marketing. But Genalytics has automated much of the process with patented genetic algorithms, making it possible for smaller companies and nonprofits to perform the same analysis in a fraction of the time and at a fraction of the cost, roughly 5 cents per name.

Kingman described genetic algorithms as formulas that breed interactions between such variables as income and age to yield accurate indicators of buying behavior.

"We're democratizing the marketing process," said Genalytics president Doug Newell. "We're allowing someone with five dry cleaners to do something that a company like American Express is doing."

Direct-mail advertising expenditures of all types hit $166.5 billion last year, according to the Direct Marketing Association. It said a $1 expenditure had an average revenue yield of $11.65.

The key to increasing direct mail's bang for the buck is reaching the right audience. Pizza coupons sent to homes where no one orders pizza go right in the trash.

Genalytics tries to eliminate waste by figuring out who is likely to be receptive to the sales pitch. Typically, the company develops a model of a business's typical customer and then finds matches among the general public.

The company said it acquires its data entirely from public sources, including telephone books, property tax bills, mortgage records, census data, real estate records, and Registry of Motor Vehicle data.

What information Genalytics doesn't have it extrapolates from data it does have. For example, Genalytics doesn't have access to credit information, but it makes educated guesses about a homeowner's income level from his or her mortgage and property tax records. Ethnic background is suggested by a name and address.

What emerges from Genalytics' analysis is a ranking of households on a scale from 1 to 100, with higher scores indicating greater likelihood that the household is interested in the product or service being marketed.

Some of the analysis is fairly easy. Someone living in an apartment is unlikely to need a new roof. Someone with a young child and a pet is more likely to need a housekeeper. Someone who can barely pay the rent is unlikely to be interested in high-end jewelry.

"We can weed out affluence from poverty pretty well," Kingman said.

Genalytics executives said their customers include smaller companies, nonprofits, and direct-mail firms looking to target customers better, as well as larger operations that buy the company's software and run their own analyses.

Genalytics said none of its smaller clients would talk to the Globe because they didn't want competitors to know about their marketing strategy.

SunTrust Banks Inc. of Atlanta, which licenses the Genalytics software for internal use, said it uses mass media to pull customers into its branches, but uses the Genalytics software to predict what products would best suit them.

Arie Goldshlager, group vice president at SunTrust, said the Genalytics software is attractive for its predictive qualities, and also because it eliminates a lot of time-consuming data analysis.

"It provides value over and above what we do," Goldshlager said.

To test the system, I asked Genalytics to run an analysis of my household, providing only my address. The company's analysis, complete with pie charts and bar graphs, included my age, number of children, income level, home value, occupation type, and length of time I've lived in my home.

The analysis couldn't guess my educational level and assumed I was single, presumably because my wife kept her last name.

Genalytics then plotted my demographic information against consumer profiles it had developed for corporate clients. I scored high as a prospect for a photo and video store (100), roofing company (96), housecleaner (91), jewelry (88), and furniture (83). I was a poor prospect for a fitness membership (10) and new appliances (18).

The results were interesting. I recently added a new roof, and the photo, housecleaning, furniture, and fitness results were on target. My interest in jewelry and my noninterest in appliances seemed backward.

"We do not have to be dead-on," Newell said. "We have to be better than any other solution."

One market Genalytics hasn't been able to crack is politics. Kingman said the company's software could be used to determine who is likely to vote and what their hot-button issues would be.

"We could definitely do a lot for them," Kingman said.

Bruce Mohl can be reached at mohl@globe.com.