Growth-Stage SaaS Platform · Technology & SaaS · San Francisco, CA · 8 weeks
Growth-Stage SaaS Platform Data Foundation case study
A growth-stage SaaS company faced confusion with five different ARR numbers used by finance, sales, product, and the board, causing a $2.1M gap. We created a certified revenue metric layer that standardized definitions and aligned all teams on one reliable ARR figure for board reporting.
Challenge
Finance, Sales, and Product each reported ARR using different data sources: billing, CRM, and usage logs. Their methods made sense individually, but the numbers didn’t match. This inconsistency confused the board and left the CFO unable to provide clear answers. We had to unify the approach and create a single, reliable ARR calculation.
Approach
We worked directly with finance, sales, and product leaders to agree on clear definitions for ARR, NRR, GRR, and churn. Then, we created a metric layer in dbt that pulls from one source of truth to calculate these metrics consistently. This layer feeds into Looker, making the numbers available and reliable for all teams. We also documented each metric and the exact business rules used.
Outcome
We identified and fixed a $2.1M discrepancy in ARR by standardizing all teams on one certified metric layer. This cut board reporting prep from three days to four hours. At the next board meeting, the CFO presented a single ARR figure backed by full data lineage. The metric layer also served as the core data source for the Series C data room.
Nailing Down Our Metrics Before We Built Anything
The problem wasn’t a tech bug. Finance, sales, and product were all measuring ARR differently, each with their own set of rules. So, we pulled everyone into a workshop. We mapped out how each team was doing it, called out the conflicts, and hammered out one clear ARR definition that worked for everyone. After that, we built the solution around that shared understanding.
We rolled up our sleeves with dbt and Looker to clean up the data. Then, we built a certified metric layer that got all our reports speaking the same language.
Here’s how we tackled it: first, we got super clear on what each metric actually meant and who was responsible for it. Next, we followed the data trail from start to finish. Then, we built the metric layer in dbt and put automated tests in place that run every time the data updates. Last step was connecting that layer to Looker’s semantic model. The payoff? No more guessing games with the numbers when the board shows up.
Results
- $2.1M ARR discrepancy resolved
- 3 days to 4 hours Board reporting preparation time
- 5 to 1 ARR definitions consolidated to single certified source
- Series C Metric layer used as data room foundation
We used to track five different ARR numbers, and the board had basically stopped trusting any of them. Thinklytics worked with our three teams to agree on one definition and built the metric layer in eight weeks. That $2.1 million discrepancy is gone, and now the board is asking more useful questions.