Data teams tasked with measuring ROI embody both an illuminating and critical aspect of driving positive business outcomes; though applicable to a variety of moving parts operating within an enterprise, measuring the ROI of the data team itself can provide a strong foundation for how data manifests in the rest of the company.
John Steinmetz, VP of data and analytics at Shiftkey, and Shane Murray, CTO at Monte Carlo, joined DBTA’s webinar, “Virtual Data Panel: Measuring Data Team ROI,” to discuss the strategies and tools data teams can implement as they look inward to measure their own ROI—whether it be to improve their data quality, or to effectively vocalize their organizational value.
Though data teams excel at measuring ROI in other areas of an organization, it seemingly becomes challenging to measure this component internally for data teams.
According to Murray, there are places within a business where measuring project impact and other variables is relatively straightforward.
For example, ML teams tend to be running constant experiments, already measuring the impact of their recommendation systems or targeting algorithms, making its ROI more forthright. In contrast, for data teams who may be, for example, enhancing the velocity or quality of a decision, it’s rather challenging to put a dollar sign to signify the value a data team is adding to a piece of analysis.
Overall, Murray explained that often data teams’ tasks are “one step removed” from creating positive business value.
Data teams operating within data platform projects, whether they are building a shared data set for lifetime value or migrating to a new data stack to support faster analytics, tend to be operations that don’t directly add value to the business—making it difficult to measure ROI.
Though difficult, it’s not impossible; in terms of implementing strategies to make ROI measurable for data teams, Steinmetz remarked that, “This is also one of those very tricky things, because you have to get your whole team to buy in to these processes. I preach to my team daily about business value. Every morning I message them and talk to them regarding what they are going to do today that’s going to actually impact the business.”
ROI is efficiency, according to Steinmetz. If you have a set of dashboards that no one utilizes, yet you must maintain, is an example of inefficiency that drives down enterprise ROI. A constant and continuous evaluation of what and how investments are impacting the business is a method in ensuring that a data teams’ ROI is measurable.
In terms of the literal calculation of ROI, Murray explained that a way to prove and enhance this number for the business is to reduce time to insight or time to maintain data products.
Proving return can also be proved through ML products, as they can be a vital tool in representing the impact from the data. Once these ML products are successful, they can effectively compensate for an enterprise’s entire data investment. Looking at metrics to aid in reducing data downtime, whether it be the number of data incidents seen or their impact, can also drive ROI visualization for data teams.
Setting and measuring SLAs is another mode in recognizing data team ROI, though it comes with its challenges.
Steinmetz explained that getting people to understand the difference between an SLA for acknowledgement versus an SLA for solution is a large issue in SLA implementation. They are inevitably subject to change as an enterprise grows in maturity—often needing to be relegated to a backlog as they increase in quantity. Communicating that an SLA is being recognized, and then providing an estimated timestamp of when a solution will be provided, must be consistent and as accurate as possible across an entire organization, according to Steinmetz.
Driving stakeholder alignment is another facet of materializing data team ROI; stressing outcome over output, Steinmetz argued, can be a simple difference in getting stakeholders to understand data team value. Prioritizing issues that will provide continuous value, as opposed to more temporal, short-lived issues, can help drive the visualization of data team ROI to stakeholders.
Aside from quantifying data team ROI, an essential ingredient in realizing a data team’s positive impact is data quality and reliability. How to get started in investing in data quality, as well as what point in enterprise maturity is an ideal place to begin, is a significant question.
Once an organization delves into BI and automation, Steinmetz explained, data quality becomes a priority.
Murray further explicated that data governance is a critical piece toward ensuring data quality.
“Whether it’s financial data or operational data—data that needs to be high quality in order for your product to work or to deliver accurate financial reports to Well Street—are the reasons why teams need to invest in quality and reliability as a foundational piece to their platform,” said Murray.
“The reason for governance or quality is ultimately about building trust in data so they can drive wider adoption, better decisions, and eventually more value from that data, across the company,” he concluded.
For an in-depth discussion of enabling measurable ROI for data teams, you can view an archived version of the webinar here.