Recently, The Washington Post published an article about how Google's HR team uses data to optimize their selection of office snacks. And yesterday I found an article about the personas that matter most to business analytics. Not surprisingly, it's not IT, or even the power analyst. It's regular old people like you and me who want to use data to do their jobs better.
I happen to work for a company in the business intelligence space (an industry I didn't know existed four years ago as a wide-eyed and freshly minted MBA), and it has always struck me a bit funny. For years the dynamic in traditional BI has been: business users complain, IT builds a dashboard to make them go away, and business users poke around trying to find something useful. But what happens when they find something and want to do something about it? Or they come up with new questions and have no means of determining an answer? Even data discovery, which provides tools for power analysts to explore interesting data trends and derive their own insights, still relegates data analysis to a small, select group of people.
So how can a business be intelligent if people unfamiliar with the data are the ones responsible for visualizing it, or only a few people are made smart through data analysis? It just doesn't make any sense. Instead, we should try to be more like Google, where employees in all departments are running constant experiments as a matter of course. Here are several tips for how you can replicate Google's success and foster a data-driven culture at your company.
1. Keep it Simple
As an economist it pains me to admit this, but there is such a thing as too many choices. We can only process so much information at a time, and it does no one any good to create a dozen reports if you're really only going to use two. Identify your top KPIs, and create 1-2 dashboards that visualize performance against those targets. Keep track of which dashboards/reports are most often accessed, and eliminate those that aren't used.
2. Make it Visible
A data-driven culture means everyone knows what's going on and is acting in concert with the rest of the organization. That is impossible if only certain people have access to key information or insights. Instead of sharing reports via email, publish them to an internal wiki that everyone can access at any time. This also helps curb bad behavior of "sitting" on information that makes your team look poorly. While no one likes to admit failure, it's important for people to be aware of issues so you can work together to resolve them.
3. Use Data to Achieve Real Business Goals
As you identify KPIs, you'll need to define what performance levels are good and bad for each metric. For example, if the standard win rate for your industry is 25 percent and you were at 5 percent last quarter and 8 percent this quarter, you shouldn't be satisfied with staying at 8 percent. Second, you'll need to establish processes for dealing with high/low performers so everyone is on the same page and you can react appropriately. For example, one of the key performance indicators at QVC (the home shopping network) is $/minute. For each product, they're trying to maximize sales for each minute it's on the air. There are established benchmarks for each product category and time of day to evaluate performance. If a product falls below the required minimum $/minute, producers end the segment early and move on to a different product, and planners reallocate the item to a lower value time slot and/or sell the remaining quantity online. On the other hand, if a product is hot, producers lengthen the segment and buyers receive notifications to place reorders in real-time. Bottom line: set goals, and be ready to recognize when something's not working so you can stop doing it and try something else. Don't just keep doing something stupid and hope somehow things will turn around.
4. Treat Everything as a Work in Progress
One of the great things about Google is that they treat their products as perpetual works in progress. They're never finished. This kind of behavior is indicative of a data-driven culture, because it recognizes the fact that there is always room for improvement and there's always something to test, tweak, and optimize. To figure out what can you can improve, ask yourself, "What do people complain about? Where do we fall below industry standards? What does my gut tell me could be better?" Then test a few options, keep the best one, and test that against a few more options. Rinse and repeat.
5. Tie it to Employee Performance
At the end of the day, employees will do what they're incentivized to do. To drive a data-driven culture, you need to set ambitious but achievable goals, hold employees accountable for activity and performance, and reward them as they use data to make better decisions. One of my favorite examples of this is Hubspot's marketing team. In addition to daily dashboards, the team creates a monthly report that reviews activity and performance for the past 30 days. Each team is responsible for creating their own report, and each report shows names and faces so results can be tied back to individual team members. The report includes goals and metrics (e.g. generated 400 leads from blog posts), as well as activity (e.g. published 20 blog posts this month), and is posted on the internal wiki so everyone in the company can see it. Performance is tied to employee bonuses and the CMO gives out quarterly awards to team members who did something awesome to achieve one of their goals. In short, if you want something to happen, write it down, hold your employees accountable, and reward them when goals are met.
So, what other advice do you have for fostering a data-driven corporate culture? Share your tactics in the comments!