Becoming a data-driven business

A big part of getting your company to be more data-driven is actually becoming aware of the fact that there are a lot of places where you can’t be. The best data-driven organizations actually have come to terms with this and can transfer seamlessly in conversation between those topics that can and should be matched against extensive research, measurement, and analysis, and those where they need to rely on a best guess or vaguely directional information, either because time is an issue, or information simply isn’t available.

Businesses fail when they go bare-ass data-driven, where they reject anything that can’t be tracked explicitly to cash flow. Radio ads, be damned! Display ads with poor click through rate? Get thee out! Some companies have started to mistakenly think that a lack of transparency correlates 1:1 to a lack of performance. This is wholly untrue. There are many tactics that a business may employ in their marketing, site experience, and customer service that are enormously effective, but where ROI may not be clear enough to show in a spreadsheet.

Becoming a data-driven organization does not include the outright rejection of decisions that cannot be data driven. It’s knowing when to use your tools to great effect, and having the smarts and civility to still have productive conversations where you know good, but measurement-evasive, ideas lie. Becoming data-driven, in other words, should not be a Kamikaze mission.

Get your teams together and begin bucketing your different business practices; figure out where they lie on the spectrum of analysis-enabled. On one end is a place where data can literally drive the program, like autopilot in an aircraft — things like paid search and product pricing. On the other end is a place where instinct and savvy drive decisions, like office furniture and graphic design. In the middle is a mix of the two, like display media, information architecture, etc. Once you’ve had this conversation, be comfortable with your decisions. When you talk about these programs later, don’t hold them accountable to irrational or irrelevant measures of success. You didn’t buy display media to compete with search (if you did, we need to talk), so don’t shut it down for a silly reason. On the other hand, DO redistribute that budget if you have a good reason in terms of opportunity cost, but watch carefully to see what the downside of that move is, if there is one.

The nice thing about analytics is we can apply it in surprising ways. We can see how billboards in sports stadiums impact search demand by geography. We can see how PR efforts sell the products on our web site and spike traffic. We should absolutely be trying to work toward being data driven everywhere, even in surprising or complex situations. But let’s not make enemies of good ideas because we haven’t figured out how to measure them yet.


No Trackbacks

One Comment

  1. I like the last line; Let’s not make enemies of good ideas because we haven’t figured out how to measure them yet. That and becoming a data driven company should not be a Kamikaze mission sums up what I think is difficult about analytics.

    Unlike some I don’t think it’s the technology or the discipline which is hard. I think it’s the process and the people part which is difficult to get right.

    A big issue in the road to the successful balancing act required to become truly effective at business optimization through analytics is a cultural one. Culture is defined by people. If you go around bashing peoples ideas simply because they aren’t analysts with numbers to back up their theories you’re going to find that you become unpopular.

    Once you become the unpopular analyst in the business you might as well start cleaning your desk.

    Posted December 20, 2010 at 11:35 am | Permalink