In the world of software investing, few metrics command as much reverence as Net Dollar Retention (NDR), also known as Net Revenue Retention (NRR). This metric, which measures a company's ability to grow revenue from its existing customer base after accounting for upgrades, downgrades, and churn, has become the North Star for evaluating both private and public software businesses. An NDR exceeding 120% is the gold standard, signaling a product so valuable that customers naturally expand their usage over time.
But a troubling trend is emerging, starting in my small bubble of venture-backed startups and now stealthily creeping into public company filings: this once-sacred metric is being distorted with increasing frequency.
Figma's recent S-1 filing provides a textbook example. The company's celebrated 132% NDR was a cornerstone of its IPO narrative, easily placing it among the very elite tier of software companies. But there’s a catch. See if you can spot it:
We calculate Net Dollar Retention Rate as of the applicable period of measurement by starting with the ARR as of the date of measurement from all Paid Customers with more than $10,000 in ARR that were also Paid Customers with more than $10,000 in ARR as of twelve months prior to such date of measurement (“Current Period ARR”). We then calculate the ARR for those same customers as of twelve months prior to the date of measurement (“Previous Period ARR”). We then divide Current Period ARR by Previous Period ARR to calculate our Net Dollar Retention Rate for the applicable period of measurement.
Translation: to be included in their calculation, a customer must maintain over $10K in ARR for the entire measurement period. Any customer who churns completely or drops below the threshold simply vanishes from the calculation.
Of course, this is a ridiculous distortion of NDR. At best, this metric is something like a net expansion rate. But let’s be frank, this formula is designed to measure expansion only among the company’s best, most stable, and continuously growing customers, while explicitly ignoring any revenue contraction or loss from that same top-tier segment.
This practice of creative calculation is not new and not even malicious. Even after normalizing for the distortion, Figma is an exceptional business and one of the most incredible PLG success stories. The real question is not how to standardize NDR, but rather: why do companies with legitimate success feel compelled to window-dress their metrics?
Einstein believed that “everything should be made as simple as possible, but not simpler.” We’ve somehow managed to violate both parts of that wisdom, making NDR both too complex and too simple. Businesses are complicated, yes, and metrics will need to evolve as business models change. My hope is that we land on some sane middle ground: honest complexity, clearly stated.