< blog
11 min read

Listen, You Should Measure Data ROI at the Data Product Level. If You Don’t, You Can Fix It. I Did.

Measure Data ROI at the Data Product Level

Data used to be seen as an expense, its connection to business results unclear. But in today’s data-driven world, this view is shortsighted. Companies now need to measure the return on investment (ROI) of their data. But without looking at the data product level they can only understand the investment in data and not the actual ROI. 

The modern data stack (MDS) has made it very easy to create data pipelines and products. It gives us great flexibility to leverage data. But because the MDS gives us so many use cases, so many users, so many pipelines — it also brings high levels of complexity. This makes it difficult to measure the returns of your data outcomes. You can now manage data more dynamically and efficiently, which bumps the cost up, but the MDS makes it very complicated to reveal ROI. 

With organizations making a shift from viewing data as a cost center to a strategic asset, ensuring resources are directed towards products that deliver real business value has never been more important. So, measuring data ROI at the data product level, despite its difficulty, is becoming crucial for optimizing resources, maximizing the impact of data initiatives, and, ultimately, achieving business success.

So How Exactly Do You Define a Data Product?

Pinning down a watertight definition of a data product is tricky. Ask six data engineers and you might even get seven answers. Okay, so we know it’s complicated. But how can we apply an ROI if we don’t know how to define a data product, calculate its true cost or define the value it delivers the organization? The answer is, not very accurately. 

So, let’s start with a solid definition. Data products are applications or assets, such as dashboards, reports, applications, or ML models, that deliver information or services to data consumers. They empower users with informed answers to complex questions. This goes beyond simple data delivery, with data products tailored to specific user needs and use cases. Think business-focused interfaces, not raw data dumps, constantly evolving to serve your strategic imperatives.


The Simplistic (Well, Let’s Call it Naive) Approach to Calculating the Value of Data Products: Usage Divided by Costs

So, it seems like a no-brainer to classify a Tableau dashboard as a data product. From there, breaking down the dashboard’s usage into numbers of views and distinct users seems like a logical way to gauge ROI. The cost, however, is deeper than just the dashboard. There are complex workflows, data pipes and data sources feeding into every dashboard. This is why this simplistic naive approach falls short of truly understanding the cost, ROI and real value. For example, what if someone utilized that Tableau dashboard to provide a game-changing insight to the CEO, leading directly to a significant investment decision and subsequent revenue increase?

This raises a fundamental question: How can the head of data measure the true ROI and value of data products? Unlike R&D teams, who can easily track the revenue generated by a specific feature, data teams find it incredibly challenging to measure impact and value. Yet doing so is vital for data teams to demonstrate their value, justify their budgets and secure the resources they need to become a true strategic partner within the organization.

The Lesson I Learned — Stop Looking at Costs and Start Explaining Value

Year after year, I constantly confronted this dilemma. Starting with a modest $50k budget for Snowflake, my costs skyrocketed to nearly $500k. To justify this expenditure, I began measuring the number of active users in Snowflake and Tableau. They grew from 50 to 300. Then I measured the number of pipelines, which grew from 100 to almost 1000. But as costs increased I began to question — what value did this truly bring?

Reflecting on my journey, I realized a critical shift needed to occur. Instead of solely focusing on explaining costs, it’s imperative for data managers to articulate their value. It is important that this value is connected to the strategic initiatives of the company. Data teams need to make sure they invest the time and budget to serve these strategic initiatives versus the organization’s day-to-day requirements, which also need to be served but in the right proportion. Therefore as data leaders, we must seek benchmarks and systems to uncover and measure the impact our data has on strategic initiatives. 

This requires a shift in focus from the get-go. So, let’s start with what most heads of data do in their few weeks in a new job. They instinctively spend the first week looking at the technology. That’s what I did.  I would wonder why they are still using SQL servers and why they have not switched to Snowflake. But I was wrong. 

When starting a new job, data leaders should first of all focus on understanding the problems and objectives of C-level managers. Just like I learned to do, they need to curb their initial instinct to delve into technology. Data leaders need to start by understanding the problems and objectives of C-level managers. Experience has taught me to always start from what creates value and then find the best tech solutions needed to solve them. 

The problem is that if you look at it from the engineering approach first you might try to optimize a job in Snowflake, even though it feeds a report that you later discover is only viewed once every quarter. Why optimize a process that no one is using? The engineering standpoint looks at the how, not the why, instead of focusing on what matters. So start with how the data is consumed and then go back to figure out the optimal process. 

Discovering the True ROI of Data Products: Key Takeaways

Through learned experience, making bad decisions and taking corrective action, I created a framework for calculating data product ROI. This is based on the following key points.

  • Focus on What Brings Most Value 

Measuring data ROI at the data product level requires a comprehensive understanding of value creation, performance, and quality. By shifting focus from costs to value, data teams can truly become strategic partners within the organization, driving impactful projects and fostering innovation.

  • Prioritization

When there is a problem with a specific data product, the ability to see the usage of that data product is critical when assessing the urgency of resolving the issue. If 100 people are using it you know it is urgent. This allows you to better prioritize resources. 

  • Look Beyond Dashboards

When evaluating a data product, it’s essential to look beyond the surface. The dashboard is merely the tip of the iceberg, there’s a complex web of workflows behind the scenes. When you measure the investment in a data product it is not only the last ETL that creates this data product but the whole workflow behind the scene that needs to be understood. This requires explainability and root-cause analysis in order to uncover and attribute these costs. 

Think of Data as a Value Driver, not a Cost Center

Traditionally, data has been viewed as a cost center, with its value only loosely connected to financial outcomes. However, in the age of data-driven business decisions, this approach is outdated. Today, data managers are under pressure to try and understand the return on investment (ROI) of their data products. This is needed to resolve one of the most complicated questions in the industry — how do you justify your data team. 

So, measuring data ROI at the data product level should now be considered as an absolutely crucial goal for optimizing resources. This realization led me to starting Seemore Data. I wanted to help organizations use data ROI to maximize the impact of data initiatives, prioritize them and ultimately achieve business success. 

This is a key difference between what we are spearheading at Seemore Data compared to other players in the sector. Their approach is to try to optimize the data workloads, which is the largest chunk of the process. They only impact ROI through a focus on cost reduction. And maybe using these tools means you are trying to optimize a workload that doesn’t bring any value at all. 

The bottom line is data leaders need to start looking beyond their data cloud. They need to think of the value of their data products. So, don’t stop at optimizing Snowflake and start looking at the value your data products are bringing. That’s why you perceive yourself as a cost center and not as a driver of value. The ability to look at ROI at the data product level will make you more connected to the value.  Always think — value, value, value.

Seemore Data enables data leaders to achieve this by revealing how their data is being used and the value it creates. Understanding the impact of their data will help them make wiser investment decisions in data initiatives that deliver value and benefit. 

If data leaders want to reduce data costs and to understand the impact their team has on the business, they need to understand Data ROI at the data product level and not at the solution level. They need to stop viewing data as a cost center and constantly focus on the value and impact it delivers. Helping you achieve this is our mission at Seemore Data. 

— Are you interested in continuing this discussion directly with Guy? You can message him at guy@seemoredata.io to delve deeper into the benefits of measuring data ROI at the data product level and how you can achieve this.

Snowflake Cost Optimization
7 min read

Snowflake Cost Optimization Best Practices

6 min read

Comprehensive Guide to Mastering the Snowflake Query Profile

Snowflake Clustering for Effective Data Organization
17 min read

How to Master Snowflake Clustering for Effective Data Organization

Ready to start seeing more data ROI?

data ROI