Numbers, they say, are a business's lifeline -- they're the only reliable way you have of monitoring your company's performance, of catching problems and spotting opportunities while there's still time to act. If you're like most business owners, however, you could well strangle on that lifeline before you have a chance to use it.

We're talking here about one of the most daunting challenges facing any manager today. How do you get control of the data that streams into your company daily, inundating you with information about everything from orders written to inventory levels to aging of receivables? Where do you find the time to put it all into perspective, and into a form you can use? How do you target the particular problems that need attention?

Large companies have a distinct advantage in this regard, since they can always dispatch a platoon of number-crunchers to figure out what's happening in any corner of the business. Small companies have to be more resourceful. On the other hand, there are some big-company techniques that small companies can use to get control of their numbers. One of most effective is something called "exception reporting."

In essence, exception reporting is a system for pulling out of the information pile those items, and only those items, that require immediate attention (the "exceptions"). Suppose, for example, that you want to monitor inventory levels of products you are selling. A standard computer report would give you average monthly levels, turn rates, and the like. That may be interesting, but it won't tell you the specific items you should be paying attention to. Alternatively, you could comb through the monthly listings of every item you have in stock -- which could take days if you carry thousands of items. An exception report, on the other hand, would short-circuit the process, identifying just those items that have unusually high or unusually low stock levels, given how much you normally sell. As a result, you'd be better able to pinpoint specific problems (or areas of opportunity) that need investigating.

To see how the technique works, consider the experience of Charles Bodenstab, president of Battery & Tire Warehouse Inc., in St. Paul, Minn., an automotive-products distributor that he bought in a leveraged buyout back in 1983. At the time, the company had sales of $7 million, but -- as Bodenstab soon discovered -- it was losing money, thanks to inadequate financial controls in a number of different areas.

One of the problem areas was accounts receivable, a field of management in which the 55-year-old Bodenstab had little experience, despite 30 years as a middle- and upper-level manager at a succession of Fortune 500 companies. "Those details," he says, "were always handled by controllers and credit managers." To be sure, Battery & Tire Warehouse had its own credit manager, but it was by no means clear that the man was up to the job. "I'd ask him basic questions, like why the average length of receivables or the percentage of past dues was up," Bodenstab recalls, "and I'd never get satisfactory answers. One month I'd hear that the problem was isolated in a few accounts. Then, the next month the same trends would continue, only for different reasons."

The credit manager was part of the problem, and he was eventually replaced. But the more Bodenstab studied the situation, the more certain he became that the company needed to do more than change credit managers. It also had to change the way it monitored accounts receivable. For one thing, Bodenstab was dissatisfied with all the aggregated data he got from computer reports -- information about such things as the average number of days of receivables outstanding. These aggregates, he found, were of little use in identifying specific problems; in fact, they often disguised them.

Nor did it help to read through the monthly account listings. "We had 3,000-odd accounts listed on 120 pages, so it was fairly time-consuming," he says. "And from the report, you had no idea which accounts were deteriorating. There was no historical perspective. The only way to interpret the results was to sit down with the previous month's report and compare."

Bodenstab decided that there were two things he needed to find out each month in order to get the situation under control. First, he wanted to know which accounts were deteriorating. Second, he wanted to know which customers were chronic deadbeats. So he began flipping through the monthly printouts by hand, making lists. "I spent hours pushing a bunch of numbers around on a piece of paper," he says. Finally it dawned on him that he could save himself a lot of trouble by getting a computer to do the work for him.

As it happened, Bodenstab had designed management-analysis reports back in the 1960s, when he was at U.S. Steel Corp. That experience gave him a sense of what he was looking for, but the process, he says, was really quite simple. "All it took was sitting down and asking myself, 'What do I really want to know about?" He answered the question by drawing up a list of the important parameters for each of the two monthly reports he felt he needed -- one on adverse change, the other on chronic problems. In the adverse-change report, Bodenstab wanted a list of all accounts that exhibited a pattern of deterioration (for instance, a sudden increase in receivables more than 30 days past due). He also wanted the history of each deteriorating account. On the other hand, he didn't want that report to include accounts that were consistently past due by the same general amount. Those would turn up in the chronic-problem report. "I decided to keep the reports separate because you read them with different mind-sets," he says. "You don't handle a deteriorating account the same way you do a chronic problem. And if you're too loose about the criteria in a report, it will be longer than you want, and you'll get awfully distracted." Once he was clear about his parameters, he hired a local software house to do the programming, at a cost of $4,000.

It's been slightly more than a year now since Battery & Tire Warehouse began using exception reports to monitor accounts receivable, and Bodenstab couldn't be happier with the results. At last he has the sense that the problem is under control. That sense is reinforced shortly after the books close each month when the two computer printouts arrive on his desk.

On one recent afternoon, for instance, Bodenstab was able to scan a new adverse-change report in about half an hour. In 16 pages, the printout listed 77 accounts (out of a total of 3,167) that had deteriorated for one reason or another. It even told him which customers were making their second and third appearances on the list, suggesting that the companies were in the early stages of longer-term problems. Looking through the report, Bodenstab was able to check each customer's record of payments and account balances, and scribble notes in the margin. "After I've studied this for a few minutes," he says, "I can talk specifics with my credit manager, and within 48 hours she's expected to have credible answers."

After going through the adverse-change report, Bodenstab and his credit manager turn to the chronic-problems report and develop strategies for dealing with delinquent customers. The report does not list every deadbeat: a customer with less than $200 overdue, for example, won't appear. Even so, from 130 to 150 companies show up on the list, accounting for 99% of the overdue dollars in the 90- and 120-days-past-due category. Like the adverse-change report, the printout of chronic problems allows Bodenstab to raise questions about the most troublesome delinquents. "I can go to a salesperson," he says, "and ask why we have increased our exposure to a particular customer. It sets the agenda. You read through this list, and it's like Chinese water torture."

That may not sound like much fun, but it sure beats the kinds of surprises that await if you don't know where your problems lie. It's hard, moreover, to argue with Bodenstab's results. In the past year, he's managed to decrease the company's number of average days outstanding from 59 to 46 days, which is below the industry norm. This was possible, he says, because exception reporting has let him focus on the critical details of accounts receivable, not the trivial ones. "I'm not about to do the credit manager's job. These reports tell us both how she should spend her time and where she should direct her attention."

You don't have to be worried about credit problems, however, to find value in the technique. Exception reporting can easily be adapted to other areas of business -- indeed, to almost any area where you can measure and monitor performance against some history, goal, or prior pattern of movement. You could use it, for instance, to evaluate employee efficiency -- by comparing the actual output of individuals with daily, monthly, or yearly projections. Then again, you might use it in quality control, to figure out which pieces of machinery, or which shifts, deviate from your established goals. Or it could help you monitor monthly sales by customer: by comparing actual orders with seasonally adjusted projections, you could spot unexpected changes that might indicate problems or opportunities you had previously overlooked.

The trick is to choose the right criteria. As long as you can pinpoint and quantify what you need to know, there should be a way to design an exception report that's infinitely more useful than raw data.

This point hasn't been lost on Bodenstab, who has been busily working on the development of a series of software programs that would allow managers to apply exception reporting to a variety of problems. The programs would, of course, have different characteristics, depending on what they are designed to do. (To monitor sales, for example, you would use different data from that required for accounts receivable.) Nevertheless, Bodenstab is confident that he can come up with a broad range of applications.

"It can definitely be done," he says. "The conceptual underpinnings of all exception reporting are the same. We're just trying to mimic what a manager would be able to find out for himself if he had endless time, endless resources, and endless data." Once Bodenstab has figured out how to do it, a lot of managers across the United States can look forward to a good night's sleep.