6 Tips for Data Teams to Improve Collaboration
Everyone wants to get more out of their data, but how exactly to do that can leave you scratching your head. Our BI Best Practices demystify the analytics world and empower you with actionable how-to guidance.
In the right hands, data is the ultimate means to answer important business questions. The problem is that when data is used incorrectly, it still provides answers (just bad ones). The best way to avoid decisions made based on that bad information is to improve the relationships between the employees analyzing data and the employees acting on that information.
Today, data questions often involve someone from a line-of-business team and someone from a data team. Both people bring their own individual expertise to the collaboration, but these projects can easily fall into the trap of unclear expectations and insufficient communication.
Before building a dashboard to answer questions, data experts need to sit down with their line-of-business counterparts and have a discussion about the purpose of that dashboard. Here are a few tips for data experts to get the most out of that meeting:
Go into every dashboard with an open mind
It’s crucial to start every new data inquiry with a fresh mind. Every assumption the data experts make is an opportunity for the overall insight to lose value. On the other side, provided assumptions and business context can accelerate the data work for a faster time to value. The goal of every dashboard creator in a data request meeting should be to fully understand the data consumer’s needs and workflow. That means listening to their individual context and letting the investigation go where the data leads. If you drive the conversation or ask leading questions, you’ll end up at the resolution that you want, not necessarily the one that’s most valuable.
Get to know the individual requesting the data
Like most collaborative projects, empathy is critical to success. I like to start data request meetings by asking the data customer to walk me through their typical day. In some cases, it even helps me to shadow them for a while. What I’m looking for is a complete picture of the way that individual uses data. There might be pain points or missed opportunities for data to be used that I can help integrate. Asking someone about their typical day is also an easy way to get them communicating openly. It disarms them and puts the focus on the personal connection rather than a business problem.
Understand the business value behind a data request
An easy way for a data inquiry to get off track is for the dashboard creator to receive a request and cut straight to building a dashboard. Sometimes a request for a specific metric might miss the bigger question that data can solve. For example, a CS person might say “give me churn” but what they mean is “we need to find a way to minimize churn.” If you were to build a chart that simply listed churn over time, you’d miss all the other data points that correlate with churn. It’s often the data team’s job to connect data from different teams and maximize business value. Narrowing in on one team’s KPIs is an easy way to take your eye off the real goal. If you understand what is at the heart of the data request, it’ll be easier to work backward and find the right questions for the data to answer.
Work in pairs if possible
It’s true that two heads are generally better than one, but this concept is less about brainpower than it is about having one set of hands dedicated to note-taking. While a second data expert can definitely help understand the bigger questions in a meeting about data, it is invaluable to the flow of the conversation to have someone designated as a note-taker. Since the conversation moves at the speed of the slowest participant, it is extremely disruptive to have one person focusing on both moving the conversation forward and also documenting the important information for later. If a paired approach isn’t possible, it’s always an option to record the conversation, but I also recommend sketching out potential charts during these sessions, and that doesn’t translate well through audio.
Start general, then get specific
It’s likely that the initial dashboard request will be for a very pointed metric. This isn’t a bad way to start, but it’s the data expert’s job to guide the bring the focus to a bigger general question that this data can solve and uncover complementary data that can be included in that dashboard. Consider the example of churn from earlier. Presumably, the company’s goal is to maximize revenue and the inquiry into churn is being done to that end. It’s probably valuable to first create a variety of charts that illustrate the effect of churn on revenue. From there, it might also be useful to consider churn for individual cohorts, find factors that correlate in some way with churn or locate levers that will improve churn. Organizing the questions you ask from general to specific will also help you organize the charts on your dashboard when it’s time to create that asset.
Build an unsorted list of questions to answer with data
At the end of the meeting, the best thing to come away with is a list of questions that can be translated into data queries. An easy way to build this list is to open a document and just list the big questions that come up as you have exploratory conversations and try to understand the bigger business issue. If you can work as a pair, your note taker can do this part too. When the meeting is over (or maybe afterward if you recorded the conversation and want some time to go back and review notes), you can submit the list of questions to the dashboard requester and make sure that you have everything covered. Once that list is approved, you can start mapping each of the questions into a chart that will be organized and placed on your final dashboard.
Thinking long instead of short
Building a dashboard isn’t a straight line: there’s a lot of back and forth, questioning, editing, and iterating between the initial request and the finished product. These tips will help you build a strong foundation for this process and leave you in the best position for creating a dashboard that can really help your users make smarter decisions. Try embracing a collaborative process; you’ll be glad you took the time, asked the hard questions.
Christine Quan is a seasoned data and analytics veteran, focused on data visualization theory and building tools to empower data teams. She’s an expert at constructing SQL queries and building visualizations in R, Python, or Javascript.