As long as there has been data, there have been errors. But at least in the good old days (way back in the 1980s, say), there was less data to manage. "The main source of customer data was basically the address to ship an item and an address to bill the customer," says Paul Kirby, a research director at AMR Research Inc. Today, data is considerably more complex. In addition to keeping track of vendors, suppliers, inventory and financial records, companies now track customers' buying habits, preferences and a host of other bits of data that make marketers salivate.
All of that drooling, of course, presumes that the numbers are accurate. But what if they're not? This May, Gartner Inc. released a startling statistic: More than 25 percent of critical data used in large corporations is flawed, due to human data-entry error, customer profile changes (such as a change of address) and a lack of proper corporate data standards. The result: soiled statistics, fallacious forecasting and sagging sales. What's more, the research firm says that through 2007 "more than 50 percent of data-warehouse projects will experience limited acceptance, if not outright failure, because they will not proactively address data-quality issues."
In other words, that major data-integration project that you championed for monthsand on which you staked your reputationcould ultimately fail, damaging your credibility and giving upper management one more reason to blame the IT department.
While it may seem ironic that companies spend far more time and money analyzing data than ensuring it's accurate, Gartner analyst Ted Friedman isn't surprised. "People become enamored with data-analysis tools and make assumptions that the data is readily available and in good shape," he says. "They gloss over the data issues, and that's a recipe for failure."
Gartner isn't the only research company tracking the data-quality problem. The Data Warehousing Institute estimates that companies lose more than $600 billion every year thanks to poor data. Forty billion of that can be attributed to the consumer packaged goods industry and retail supply chain alone, says Kosin Huang, a senior analyst at The Yankee Group. According to a recent study by Forrester Research Inc., 37 percent of companies cite duplicate and overlapping files as significant data-management problems.
Dirty data can damage every aspect of your business. On the customer-service side, bad and out-of-date information can mean failed marketing promotions and angry customers. In the supply chain, poor product data can cause production bottlenecks and slow down delivery orders to retailers. And if you're a company that has to meet with, say, Wal-Mart Store Inc.'s RFID mandate, flawed data increases the likelihood of glitches in that rollout, too. "Sure, you can track cases and different pallets," says Huang, "but you may be tracking the wrong thing."
The benefits of squeaky clean data are numerous (and, in many ways, obvious): Better customer data means improved customer service and a decreased risk of failed marketing promotions, not to mention new opportunities to cross- and up-sell. It also increases the accuracy of forecasting, makes your supply chain more efficient, and helps companies comply with federal regulations such as Sarbanes-Oxley.
"There's no doubt that a quality-checking process cannot be skimped," says J. Wadsworth, vice president of data services at MarketTouch, an Alpharetta, Ga.-based direct-marketing company. "If you're not doing it, you're throwing money to the wind."