"How much will this data project make us?" It is the question any CFO asks — and the one many technical teams struggle to answer. Calculating the ROI of a data project is not magic; it is connecting the investment to concrete gains.
The base formula
ROI (Return on Investment) is simple: (gain − cost) ÷ cost, as a percentage. If you invest 50k and generate 150k in value, ROI is 200%. The hard part is not the math — it is estimating each side well.
The cost side (the easiest to forget)
It is not just the license or the consulting. Also count: internal team hours, cloud infrastructure, training, and maintenance after go-live. A project that is cheap to install can be expensive to maintain — include that.
The gain side (the easiest to overstate)
- Time savings: reports that took 2 days now take 10 minutes × cost/hour × frequency.
- Additional revenue: better segmentation, less churn, faster decisions that capture opportunities.
- Avoided costs: fewer errors, less dead stock, fewer compliance fines.
- "Soft" gains: better decisions and trust — real, but assume them prudently.
The common mistake: only counting what is easy to count
Many projects look low-ROI because only the hours saved are measured, ignoring the revenue captured by deciding better and faster. Others inflate "strategic" gains impossible to verify. The balance is assuming conservative but complete gains.
In practice
Before you start, write the ROI hypothesis in one paragraph: what it costs, what gain you expect and how you will measure it. After go-live, go back and confirm. That discipline separates data projects that survive from those cut in the next budget. Can you write today the ROI hypothesis of your next project?
