How To Get One Version Of The Truth

How do you get "one version of the truth"--that one source that everyone in a company agrees is the real, trusted number for some operating data?

Without one version of the truth, people spend a lot of time arguing about whether there's a problem, how big it might be, and whose fault it could be. Notice what's not being discussed: how to solve the problem.

Yet one version of the truth is hard to come by at a lot of companies.

We're not talking about audited, public numbers like revenue and net income. We're talking about operating data, the kind that tells if you have product in stock, fulfilled an order, or met a customer service appointment. With today's growing use of dashboards by large groups of employees, and the focus on managing through key metrics, having one version of the truth is becoming more important.

Jeanne Ross, director of MIT Sloan School’s Center for Information Systems Research, has some elegant advice for how to create one version of the truth: just declare it. Pick the source, and declare that this is now the one version executives and employees will use to make decisions.

Getting to one version of the truth "doesn’t have anything to do with accuracy, it has everything to do with declaring it," Ross said. "Once you tell everyone 'This is our single source,' they work pretty hard to make it more accurate." Ross spoke at last month's TechTomorrow conference in Columbus, Ohio, about how companies can wring more value out of their installed IT to compete in the digital age. Establishing one version of the truth is one key element.

Does Ross's "declare it" approach sound rash? Well, the alternative is that teams spend far too long collecting, cleansing, and perfecting what they think is the right data, all while trying to convince people it's worth their while to embrace this one view--people who might prefer to keep the truth fuzzy, debating the numbers rather than addressing a business problem.

[ Want more on CIO strategies? Read 9 Critical Trends For Innovative IT: InformationWeek 500 . ]

My question for Ross: How do you get started? How do you get through the initial patch when you have weak data without steering the company onto the rocks and losing all credibility? "Be modest," Ross said. Start with an area where you know the limits of the data, admit the limits, while still making clear those are the numbers you're using.

The idea resonates with what I’ve heard from some CIOs.

Procter & Gamble CIO Filippo Passerini did something very similar. Passerini's IT team created an executive meeting room with data at its core, with 8-foot-tall screens projecting data that execs use to analyze how the business is doing. The IT team also is giving dashboards to tens of thousands of employees.

But Passerini didn't wait until he had perfectly cleansed data to get this analytics effort started. "We intentionally put the cart before the horse, because it is a way to force change," he said when we named him InformationWeek's 2010 Chief of the Year. If a data point is getting discussed on those screens by the executive team, the pressure grows to acquire that data.

The problem with waiting until the data is perfect to make it the "one source" is that getting the data perfect isn't a business goal. It reeks of data for data's sake. But if people can see how a single source of the truth relates to gaining market share, or not ticking off customers, then they know why it's important. That's a truth they'll find worth seeking.

Read here for more on P&G's tech strategy led by Filippo Passerini.