Ins and outs of Revenue Retention Rate

Revenue Retention rate
You guys may have noticed that various Saas companies will report some form of a revenue retention rate. Generally the formula is, “(Starting MRR + expansion – downsell – churn)/Starting MRR” This isn’t a GAAP standard so companies are free to define and calculate whatever works best for their company. I do think it is important to understand what they are calculating and how what the company chooses to report can either misinform or inform us as investors. A while back cloudera was reporting their Net expansion rate (NER) was 136% but somehow they only grew their revenue about 20%. It turned out that they were cherry picking only their largest best customers to calculate their NER. As an opposite example paycom (PYC) reports “revenue retention rate” of around 90%, which seems bad until you read their quarterly report, “Dollar-based retention excluding the benefit of upsells, based on GAAP subscription revenue” So basically they have excluded upsells which every other company I know includes.

Some key takeaways:
The various rates that companies are reporting aren’t very comparable between companies.
Yes you can paint with some broad strokes but a AYX 130% DRR is not the same as an OKTA 121% or a PYC 90%. AYX’s 130% isn’t necessarily better than OKTA’s 121% or PYC’s 90%.

A company’s Revenue retention rate or whatever they call it can be used to track a company’s performance for how “sticky “ their product is. In general companies can boost this number by price raises, adding new products that they then sell to existing customers, and decreasing loss or downsells of customers (i.e. moving from netflix’s 11.99 plan to their 8.99 plan).

Sometimes companies will start messing with how they calculate the number so that they can report a prettier number to investors. I.E. cloudera cherry picking only their best customers.

As a company matures they will switch to a cohort analysis. Generally they will calculate their version of a revenue retention rate(RRR) on customers they acquire by year. This is useful because otherwise their RRR will be artificially low and not be a useful number as they have a bunch of mature customers that are spending as much as they are going to spend on the platform. Those customers are targeted differently than new customers so it is helpful to see if what they are doing is affecting the new customers vs the old customers. See New Relic or Salesforce.