In an earlier post I reviewed Eva Murray’s Empowered by Data: How to Build Inspired Analytics Communities, a great read which really drove home the importance of establishing informal and community-driven initiatives to cultivate data literacy.
Daylong training events are great, but not enough; a culture of data confidence and inquiry is built by the everyday interactions of teams working with and helping each other work with data. As Eva puts it in the book:
Documentation… and reusable content should be easy to access and as openly available to community members as possible. … this results in a community that can help itself and analysts who can take responsibility for their own learning and skill development.
A common way for teams to collaborate on documentation is a wiki. They’re relatively easy to set up and, thanks to Wikipedia, everyone’s familiar with the idea. But opening a wiki is one thing; people using it, and it facilitating data literacy, takes more effort.
To gather insights on the topic, I spoke with Michael Zelenetz, Director of Data Management and Analytics at White Plains Hospital in New York, who successfully implemented an analytics wiki at a previous role.
Designing for your audience
Just like with any analytics project, the guiding principle of building a wiki is to “know your audience.” After all, the wiki has to prove its benefit to them, and you need a plan for adoption.
In particular, Michael found that the way to design the wiki depends on how comfortable the team or organization currently is with data. During lower stages of data maturity, something more like a business glossary (related to data governance) may be a better fit than a full-on wiki. This puts everyone on equal footing to discuss processes and documentation.
As the team becomes more comfortable working with well-governed data, they are in a better spot to contribute to a wiki. But how do you facilitate that adoption?
It’s a good way to have a discussion … so the conversation is captured over time.
Facilitating adoption
When asked about what holds people back from contributing to a wiki, Michael recalls: “In my experience, people have been nervous about having it too democratized in the same way that people are nervous about Wikipedia.” There’s commonly a fear of making unsubstantiated changes to the content, of being wrong in public.
To break the logjam here, Michael found that adding up- and down-vote features was a quick way to begin engaging with the wiki. Comments and questions also provided a way to contribute without the trepidation of changing the source documentation.
These records became just as valuable as the source itself for building data literacy in a team: “It’s a good way to have a discussion and use it almost as a message board, so the conversation is captured over time,” Michael advises. And, when working with data, there’s a lot to have discussions about.
Benefiting from the wiki
Anyone who’s worked with data knows there are nearly as many exceptions as there are rules. The wiki became a place for Michael’s team to flag these cases. It “allowed people to add in caveats or findings they encountered along the way, interesting exceptions or danger zones,” Michael added.
This speaks to an interesting point about knowledge sharing: that if you’ve had a problem, chances are someone in your team has had it too. The wiki offers an efficient medium for documenting and opening up a dialog for dealing with these problems. But this is only possible when there’s a shared language to begin with – and that’s why getting data governance in order comes before opening a wiki.
Evaluating success
I am a data analyst, after all, so I had to ask: What sorts of KPIs or ROI figures indicate a wiki’s success? There are quantitative ways to evaluate, such as number of backlinks, comments, and so forth. But to Michael, the “success was in its adoption… the fact that it continued to be referenced and updated.” After all, a wiki doesn’t require much physical capital to set up.
Wiki wiki un-wild data
The success was in its adoption… it continued to be referenced and updated.
It does, however, take real cultural buy-in to adopt and derive value from. So start small by encouraging users to vote or comment first, then change documentation when they’re ready. Make it clear that there’s value in documenting steps taken and lessons learned, and that there’s a conversation worth having among colleagues with the help of the wiki.
Data literacy is not a one-and-done initiative; it takes teams coming together to share and grow their knowledge through day-to-day operations. A wiki is a great medium for facilitating and capturing this progress.
As organizations become more distributed and operate more frequently through asynchronous channels I see the value of wikis rising, not just for data teams. Thanks again to Michael Zelenetz for sharing his expertise on building an analytics wiki; I hope it can inform your data literacy efforts.
Leave a Reply