I’ve been at the Data On Purpose-Do Good Data conference yesterday and today and this story in particular stood out to me.

When Hyesook Chung, former executive director of the child advocacy organization DC Action for Children, looked at her data it seemed like great news. Education outcomes had improved for the children they serve. But when the data was disaggregated, separating out the details for various demographic groups, her heart sank. Outcomes for black male students had not improved at all.

What do you do when some of your data doesn’t support the story you want to tell?

Chung, who is now DC’s deputy mayor for health and human services, decided to tell the truth. It was hard, she explained, but it was the right thing to do.

I think we would all agree that honesty is the best policy, even when it is tempting to portray a set of “alternative facts” that support our desired narrative. The best practice goes a step further. In our research on data maturity in organizations, Idealware has found that organizations with mature data cultures handle this situation differently from those with less mature data cultures. They dig down, like DC Action for Children did, finding the patterns that aren’t immediately obvious. When they discover something unflattering, instead of downplaying it, they investigate to find out what it can teach them.

Chung spoke at the Data On Purpose-Do Good Data conference February 7-8 at Stanford University. For more reflections and reactions, follow #DoGoodData on Twitter.