Strategies to cultivate data curiosity
How to get a quiet room engaging on data
This post is a continuation of my Learning from Metrics Review series for data leaders, where I discuss using signal from Metrics Reviews to drive your organization's data culture and data-driven decision-making forward.
The first post is here, and will be updated with links out to all other posts as they are published. Be sure to subscribe to stay up to date with future installments!
During my undergrad, I worked part time for the Math Lab at the university's Mathematics department. The Math Lab was a flexible course format for all students enrolled in an algebra course to satisfy their degree’s math requirements. It let them jump in to lectures or a study hall with tutors, as needed. Employees — mostly math majors — did the lecturing, study hall tutoring, and grading.
All told, I interacted with hundreds of students over the course of my two years working at the Math Lab. While the teaching was rewarding, there's a few students in particular whose experiences stick with me to this day. To an individual, they were smart, clever, and more than capable of mastering the material. But somewhere in their K-12 journey, a teacher told them they weren't good at math, and it stuck. Whenever it came to math, the stress took over, and they shut down.
For these students, it wasn’t that the material was too challenging. I could reliably build their familiarity with it as fast, or faster compared to other students. Overall, teaching the material was maybe 10% of my job. The other 90%: Earning their trust, so I could help them leave the voice behind that was telling them they couldn’t do it.
One thing I've learned as a Data Science leader is that there's a very wide spectrum when it comes to stakeholders' comfort with data. I put those students with a mental block for all things math on one end of that spectrum.
On the other end, I think about one of the PM partners I had when I first transitioned into a Data Science role. While he wasn't a Data Scientist, he was incredibly data savvy. His mental model for his product’s metrics was so robust, that he could absorb whatever my team threw at him, quickly pull out key takeaways, and engage in deep follow-up conversation about the data. He intrinsically knew when an request was a quick lookup vs. a lengthy project. If data he needed wasn't immediately available, he would fluidly pivot to expressing some amount of uncertainty about a hypothesis and move the discussion forward with that framing.
When I think about the business leaders in the room during a metrics review, I don't often picture someone with a deep math aversion. But the other extreme is just as uncommon. Everyone is showing up with past, invisible baggage around working with data, along with a strong desire to be seen as fully on top of their numbers. So if you notice a lack of engagement in data discussions, how do you nudge your stakeholders in a positive direction in a supportive way?
It's a tough line to walk. Lack of engagement doesn't necessarily mean lack of ability. What looks like silence could just mean there’s nothing critical to discuss, or that there’s something more important holding their attention.
So I typically avoid overtly asking about data familiarity or expertise. Instead, I focus on strategies to bolster everyone's ability, both during the review meeting, and afterwards in targeted followups.
Building a level playing field in metrics review meetings
If the goal is to create an environment where data curiosity can run rampant and data conversations can thrive, a key first step is ensuring that the data being presented is accessible to all in the room. This can seem trivial, but in the rush to prepare for the metrics review meeting and the rush of the meeting itself, it can be quickly glossed over.
One role for the data leader is to articulate and enforce best practices around sharing data that can sometimes feel too fundamental to bother articulating. For example, when it comes to data visualizations, I tell my team (and remind myself):
Don’t forget the basics: Each chart should have a title, labeled axes, and a legend to articulate the meaning of different colors / sizes / etc. used to differentiate data.
Before you discuss any insights, give a ~30 second overview on what the graph is showing, and why it’s relevant to the discussion. Depending on the flow of the meeting, give a brief pause for questions or explicitly ask if there’s more to clarify before moving on.
When you discuss your insight, highlight the part of the visualization that’s supporting the insight. If you’ve copied a chart from a dashboard into a presentation, annotate it with shapes to call viewers attention to what they need to focus on. If you’re sharing a dashboard directly, full-screen the chart you’re reviewing, and use your mouse cursor to call out what you’re referencing.
While these tips can help build a consistent foundation for sharing data in review meetings, data leaders should go a step further: When these steps are skipped, model asking those probing, clarifying questions. This may feel like a given when you personally don’t understand what’s being discussed, but it’s a also good idea even if you have prior familiarity. Depending on the org culture, it can feel scary for those in the room to ask questions — they want to look like they understand what’s being discussed. As experts, data leaders have natural cover to model asking these questions: "Can you walk me through what this graph is showing? I'm not sure I'm following yet."
Keeping the data conversations flowing after the review
The second strategy I pursue involves 1:1 conversations with targeted leaders after / between metrics reviews. Practically, this is because the 1:1 conversation isn’t suited for the review itself. But also, it helps to keep the state of metrics consistently top of mind, rather than saved for a pre-review fire drill.
If I notice a stakeholder is sitting back during data conversations and not engaging, there’s a couple tactics I’ll try to connect with them in a smaller discussion:
Connect-the-dots followup after the review meeting: I make a note of 2-3 key data takeaways from the meeting, and think about how they apply to the stakeholder’s space. Specifically, I’m looking for data that someone else presented, that I want to encourage engagement on. This might tee up our next 1:1 or a DM. "The Engineering Director shared some helpful data around the outage we had last week, and how that impacted the product. I see some signal that this could be impacting a few customers you care about in particular; shall we grab some time to discuss?"
Informal, biweekly dashboard check ins: Many data teams work hard to build dashboards and other data tooling to support teams, only to be discouraged when they aren't used. While there are many strategies to address this (topic for another day!), a key approach starts with the data leader. In addition to being responsible for data quality and availability, we also have a role in ensuring its effective use. Checking in regularly to discuss the latest dashboard trends can be a helpful way to establish this practice. "I notice a few customers had declines in their usage over the past couple of weeks, according to our usage dashboard. Here’s my theory as to why that might be. What are your thoughts?"
Each of these strategies accomplish a few things simultaneously. First, they allow me to model data interpretation by putting a stake in the ground, then inviting others to comment. Sometimes, my first interpretation might be wrong, or missing some context -- even better! It sets a helpful precedent. If a data leader can sometimes start from an incorrect interpretation, it shows that it's safe for others to follow suit.
Second, these strategies help me to express my own data curiosity, and stoke that same curiosity in others. I believe everyone has intrinsic curiosity around data, but it can get tamped down by a feeling that it’s something they should already know, or they have to go figure out by themselves. So as a data leader, I try to insert data discussion into each of my stakeholders' day when I can. Being conversational around data is a skill, and building it requires practice!
Lastly, and most importantly, it helps me catch the rare, but urgent, case where a stakeholder is lacking understanding, and anxious about expressing it. Like my math students, it takes a lot of trust to reach out and ask for help. These strategies leave the door open for a deeper conversation if someone needs it.
Post photo by Marija Zaric on Unsplash Architecture | Curiosity
Great suggestions. I recall one of your undergrad students telling me he doubts he would have graduated from college had it not been for you. Trust is key. Maybe when you are done with this series you can address what it takes to be a good stakeholder/client.