Have you ever been asked to show the impact of your data team? I have. It can be a pretty nerve-wracking question to receive as a data leader, especially when it's coming from your senior executive team (and the quality of your answer will determine whether you’ll be getting those resources you asked for).
Although this question used to be–arguably–less common for data teams (in the frenzied journey most businesses have been on to become more “data driven”), with the economic shifts of the last few years, it's quickly becoming front and center. The more you dig into this question, the more you realize that it is intrinsically tied to being able to explain what, in a quantifiable way, the value of your job is.
In the hype cycle of the “modern data stack era” (and even before this, if my memory serves) data teams have often shied away from answering this question directly. Either because we don't want to face that level of accountability, or because it’s “really hard to measure”…regardless of the reason, most times this ends up being some kind of “hand wavy” / back of the napkin math answer (that we don't really feel 100% confident about).
But, if we can all agree that at least some of the work we as data teams do plays a pretty crucial role in enabling an organization to be successful, then maybe we just need to think about how we quantify that value a bit differently.
How To Measure the ROI of Your Data Team
There are 1,000s of articles out there (a thoughtful one from Tristan Handy, a more theoretical one from Benn Stancil, and a more formulaic approach from Mikkel Dengsoe) discussing how to think about measuring the value of your data team (count this one as 1,001!), so let me throw some more ideas in the ring about it.
If we follow a basic formula for measuring return on investment (ROI), then data teams need to be able to quantify two things: the total revenue we’ve generated, and our total costs of operation:
Measuring costs directly is pretty easy. We can add up our team’s salaries, infrastructure costs (warehousing, ETL, tooling, etc.) and get to a monthly line item amount pretty quickly.
The hard part comes when you’re asked to measure the revenue you’ve generated. Quantifying the revenue impact of a data team is hard, because most of our efforts have an indirect impact on things like revenue. The work we produce often serves as a foundation for other stakeholders to leverage and derive value from data products, so we’re at least 2-3 degrees of separation away from direct revenue (even on the most successful projects).
Thinking Outside the Box–Borrowing from the “Agency” Model
But, what if we approached measuring the data team’s impact from a different perspective?
Several years ago, I was working as a data analyst at a marketing agency, and my team was facing similar challenges: we were overworked, desperately needing more resources, and the powers that be were pushing back saying we needed to justify our ROI before adding more people to the team.
So, what did we do? We built our own internal P&L report, and started recognizing the work we did as direct revenue impact. If a client requested a report or analysis as part of their campaign engagement, we separated and recognized that revenue in our team’s P&L, rather than it going to the account management team. This changed the game for us, and made it super easy to point at the impact we were having at the company. We started to standardize our offering, and take this into pitches to show potential clients the value we could offer from our work, and the associated costs.
Data Teams as Revenue Generators
So this got me thinking, what if we tried to quantify the value of internal data teams in a similar way? We could redesign our engagement model so the data team was "paid" for the work we generated (e.g. a dashboard to be built, or a feature to be added to a model).
I know. This sounds like most teams would say “hard pass”...but hear me out. This could play out in a few different ways:
Charging for Services Provided
Say a data team adopted an “internal data consulting model", where other departments or stakeholders within the organization would “pay” the data team (with internal budget) for data asset creation or support requests (i.e. stakeholders would be charged based on the number of hours it took to develop a new feature in a data visualization tool like Looker, multiplied by an hourly rate for the data team).
Charging for Access
Another method of monetizing data teams internally might be to charge teams a baseline fee to access your internal data products (kind of like an internal SaaS offering). This would help cover the cost of maintaining and scaling the data systems, as well as show which teams were willing to pay for access to the tools provided. Specific teams could be provided with access to specific features based on cost tiers, to help get a sense of interest in various parts of the data offering that could be provided.
Charging for Usage
Taking this access model a step further, data teams could also consider charging for the usage of their data products. By tracking metrics like the number of queries run by specific teams and estimating the associated cost against the data warehouse, data teams can allocate expenses more equitably across teams. This challenges the notion that data teams should be solely accountable for the costs incurred by the data product we’re creating, and pushes some cost accountability back to the primary users.
Implementing any of the models above would enable data teams to:
- Track “revenue”: The data team could more easily measure their approximate revenue impact or cost avoidance of outsourcing to a contractor for similar work
- Impact measurement: It would help data teams showcase the value of their contributions more directly
- Reduce unnecessary/low value data work: Charging stakeholders would encourage them to do a harder evaluation of the necessity of their requests, leading to a reduction in unnecessary data inquiries (which is good for the data team and the company)
Where this might need some finessing would be:
- Convincing departments to change engagement models: Adoption of this engagement model would certainly require a pretty large shift in how stakeholders interact with the data team. You would likely need to start by getting buy-in and support from your leadership team and department heads, before seeing success in rolling something like this out.
- “This just makes the data team a cost center”: Facts! But aren’t data teams already being seen as cost centers? Department heads already have many other justifiably needed costs for their operations (e.g. advertising spend for marketing). By treating data assets more like justifiable costs (i.e. “we need this attribution data to determine where to efficiently spend marketing”, or “this sales funnel dashboard is critical to be able to measure lead conversion”), departments would also be more invested in the output.
- Some teams wouldn’t have enough budget to engage us: Departments would need to make a case to the leadership team for more budget/resources to be shared from the data team (i.e. someone else is making the case to hire on your team, instead of just you as a data leader). To make things simpler on the revenue recognition side, data teams could develop a process of internal resource allocation, to avoid the need for actual funds/budget to be exchanged (i.e. through allocation of a specific number of sprint story points per department for a specific time period, and additional support would require more resources allocated to the data team).
So look, I get it, this sounds like a radical idea, and it probably is. But even just entertaining the idea helps us, as data teams, to think outside the box. Transforming the data team engagement model could help us address the challenges of measuring and demonstrating our value, and help us run our teams in a way that focuses more on justified growth and infrastructure scaling. Pushing the data team to operate more like an internal business unit that proactively focuses on its ROI may be what is needed to reach the holy grail of data-driven culture, where teams truly appreciate the importance of data teams in driving success.
Try Secoda for Free
Secoda offers a wide range of features designed to streamline data workflows, increase productivity, and maximize the value of data assets. With Secoda, data professionals can easily access and analyze relevant data, saving valuable time and effort. The platform's advanced search capabilities enable teams to quickly locate specific datasets, reducing the time spent on data discovery. Additionally, Secoda's collaborative features foster seamless knowledge sharing and teamwork, accelerating decision-making processes and driving better outcomes.