Proving the value of your data
Data teams are often met with cultural and communication challenges in their organization. We talk to data teams and data practitioners all the time and the same question keeps coming up: “How can we prove the value of our data, and the work we do?”
At first glance, this might strike you as a strange question. Isn’t it widely known that data is more valuable than ever? While that’s crystal clear for us, working within the industry, many data teams still have a lot of work to do to persuade our colleagues in Product, Marketing and other teams that we can bring value to them on a daily basis.
Even in a data-first organization, there’s often educational barriers that data teams must overcome to win their colleagues’ trust and become recognized as trusted partners. These barriers can manifest in different ways, for example:
– Often teams in the business are simply unaware of the different ways data (especially behavioral data) can aid their work;
– Some organizations still view data teams with the old mindset of slow-moving ‘Business Intelligence’ teams who struggled to iterate or keep up with company demand;
– Certain stakeholders may have struggled to find any success with data in the past, and are predisposed to a negative view of data-related projects;
–Teams hindered by restrictive or limited data tools may
On the flip side, each one of these educational barriers presents an opportunity. This is where the data team can take the lead: running workshops, upskilling their peers and raising data literacy. More broadly, data practitioners can get to the heart of key business challenges by partnering closely with colleagues to support their daily operations. The caveat here is the data team should not try to do too much.
How can I showcase the value of our data projects?
When it comes to showcasing value, the team at Hex said it best. It’s less about proving the ROI of your data projects; instead select a team within the business to make your close ally, and focus on delighting them.
Treating your data as a product and serving it to your internal ‘customers’, so they can feel the positive effects of their insights on a daily basis is the surest way to prove your value. Eventually the team you serve will evangelise for you, and other teams will be angling to work with the data team to enhance their work.
This is a journey from building initial trust, to delivering value and reaching a point where your colleagues rely on their data and can’t do their best work without it. Real success for data teams is apparent when internal teams stop running analytics projects themselves and are happy to depend on their data team to deliver insights. At this point, data is a central, strategic part of the conversation within the business, and proof of “ROI” is not required.
What about SLAs? Will these help to demonstrate value?
While Service Level Agreements (SLAs) can be useful, they should not be a defining feature of how the data team is perceived in the organization. A good place to start might be to use tickets to set expectations, for example “We’ll respond to new enquiries in X amount of time” is something the team can control to maintain communication standards.
But SLAs are not the silver bullet when it comes to showing or delivering value. Rather than being hung up on delivery times (which run the risk of the data team being seen as a cost center), it’s worth going deeper into how well data is integrated with front-line teams. It’s worth asking:
– Are we supporting this team in its daily operations?
– Are our colleagues satisfied with the reports and insights we’re delivering?
– Have we made a positive impact on their key projects in the short term?
If the answer is yes to all or any of the above, you’re likely on track for delivering data successfully. On the other hand, latency or response time SLAs (while useful) are not necessarily a more reliable way to measure value than these qualitative factors.
The answers here heavily depend on the team you’re partnering with and their main area of focus. For that reason, it’s far more efficient and effective for the data team to pick a team or group and work with them with a particular goal in mind, than to try to serve the whole company at once. In other words, you can’t deliver a data product if you don’t know your customers.
How can we identify our main data customer and initial project?
When a new data team is established in an organization, it’s likely that either of two things happen.
Scenario A: The data team is immediately swarmed with requests, and must find a way to prioritize from a tide of inquiries;
Or scenario B: Internal teams are so engrossed in their own analytics (or used to working without data) that the data team is largely ignored.
In either case, the data team must identify their main data consumer, and partner with them to establish what the first data project should look like. There are a few different ways to achieve this.
One approach is to engage with business leadership to seek out what might be low-hanging fruit where the data team can deliver immediate value. This might be a project or use case, even a small one, that will increase revenue or save money in the short term. For example, as Emilie Schario, Director of Data and Business Intelligence at Netlify found: “it came down to understanding the value of new projects – we found a use case that will save the company $10,000 a year, and that will compound over time”. For Emilie, this gave her and her team a starting point where she could quickly deliver for the company and bring data into the spotlight.
One of the best ways to decide where to get started is to look into company-wide OKRs, such as ‘product adoption’, something that’s a shared goal across marketing, product and data, for example. With this in mind, you can hone in on a particular team or segment of a team that’s working closely on this goal, and provide value to them. Once you’ve established a flow of data into that team, you can expand, iterate, and repeat the process, until the data team is supporting multiple disciplines on a host of projects.
Key to getting this right is about saying “no” as much as it is about agreeing to new projects. As Boris Jabes, CEO and founder at Census points out, this isn’t easy, but this focus is crucial to the data team staying productive.
How can we establish trust in the work we do?
When it comes to the work data teams do and how they’re perceived, it often boils down to trust. If the team you’re working with trusts their data, they’ll use it to drive success. But establishing a level of trust with data consumers is easier said than done. Where do we start?
This is where data storytelling comes in. While data reliability and lower latency matter to data teams, colleagues in product and marketing are interested in the story and the outcome, with questions like“how will this impact our work? What results can we expect? How will the data allow us to be more effective?”.
To answer these questions, we need to keep up regular, clear communication. You can achieve this by:
- Putting things in writing. Don’t just rely on hopping on a call;
- Translating the ‘ingredients’ of your project (SQL) into plain English, so people can understand what’s in their report on dashboard;
- Taking advantage of data catalogues to help people understand the data;
- Getting data as close to, or integrated with, the tools and workflows your data consumers use every day.
It’s also important to focus on delivering value, little and often, rather than trying to ship huge, long term projects that take months to complete. Taking a lead from DevOps, data teams can aim to be agile, working in sprints to ship valuable outcomes in short bursts. This will prove the data team’s effectiveness very quickly, while also removing the data bottleneck in the organization.
Finally, it’s worth considering how we as data practitioners can upskill our colleagues and improve their data literacy. With stronger knowledge around data best practices, they’ll be better equipped to make use of their data and dashboards, and better able to ask the right questions of the data team.
For a great example of this, the team at Zapier even rewards individual employees who have completed data educational programmes with ‘levels’ and are able to rank individuals on their data literacy based on the level they attained. You might not want to go to this extreme, but having some level of ‘data onboarding’ will likely strengthen the data culture and data maturity of your organization.
At Snowplow, we encourage all new employees to take an onboarding course to strengthen their knowledge of behavioral data and the wider landscape. This helps everyone on the team stay on the same page and increases understanding between technical and non-technical teams.
Whatever you decide, you’ll need to work as closely as possible with your colleagues in other teams, build empathy for their challenges and do your best to communicate with them in a language they can understand. Tools like Snowplow and others in the modern data stack can help you deliver behavioral data, but that’s just the beginning. The real work, the complex challenge of communication, is where data teams can shine.
To learn how Snowplow can help you deliver better data for your organization, get started here