The Devil is in the Data: Fact-Based Decision Making
Most of us like to think we conduct our affairs more or less rationally. In life, we spend a lot of our time balancing what our gut tells us to do with what we know from our life experience, our education, and maybe, in some situations, a little research. Everyday decision making is a balancing act. We weigh what we’d love to do with what it makes sense to do, make choices, and celebrate the times when it all comes together.
In business, decision making can look very much the same. Great business people have great instincts. Sometimes those instincts are backed up by data, but sometimes it’s really all about intuition (that was the genius of Steve Jobs). Good business people—which includes most of us—have good instincts. And we may not be geniuses, but we’re smart enough to know that we should be using data to verify those instincts and make the best choices we can.
So, let’s talk about data. When you’re using data to make business decisions, you need good data—I think we can all agree on that. But finding that data isn’t quite as easy as it should be. And that has consequences for business decisions. For example, say your company wants to renegotiate its contracts with its suppliers. You’d likely begin with a spend analysis and go to the negotiating table with that volume-based figure. That assumes, of course, that you’re able to calculate spend accurately. And that assumes that your data is appropriately categorized across the enterprise. But guess what? Many companies (perhaps even most) will be challenged to find that data. Sometimes the data is incorrectly categorized, or categorized differently across departments. Sometimes data quality issues get in the way. Sometimes a vendor has subsidiaries that get overlooked. And sometimes it’s just a typo. Whatever the reason, we’ve found that errors in determining spend can range from 5% to as much as 25%. When you’re negotiating a contract worth millions of dollars those numbers aren’t inconsequential.
Or suppose your company wants to identify your most profitable customer segment for an upsell campaign. “Customer” is, for most companies, a fluid category. A customer as defined by sales is probably going to look a lot different than a customer defined by marketing. And even within marketing, an online customer might mean everything to the digital team and nothing to the business development team.
The point isn’t to have “one” definition, but to be transparent about your definitions. What does that mean? To my mind, transparency is about making data—and its meanings—available to the people who need it for decision making. Transparency means being clear about where the data comes from, how it’s being defined, and, ultimately, whether it can be trusted to answer the questions you’re asking. Transparency is not perfection. I’ve seen a lot of companies get tangled up in trying to make their data perfect and honestly, that’s a fool’s game. Perfection is very expensive. And extremely time consuming. I am much more inclined to think about whether data is “fit for purpose.” Is it reliable? Can we trust it to make the business decisions we need to make? Do these sets of reports meet our threshold for accuracy? Can we tie these data points back to their sources? And do we trust those sources? If your data is detailed enough—and trustworthy enough—you can make data-informed decisions that will look like pure genius.