Three findings from MoneyPeel's own sample. Each one prints the query behind it, so nothing here has to be taken on trust.
The library, by freshness
738 cards
Those pages still open and their reports still read, but the amount never enters a total and never takes a rank. Most libraries sort stale rows next to fresh ones; this one removes them from the arithmetic and says so.
SELECT count(*) n, sum(coalesce(revenue_stale,0)) FROM startup s LEFT JOIN report_spec p ON p.slug=s.slug WHERE s.status='published'Not an error: MRR is a run-rate at a moment, the 30-day amount is a window of what arrived. A product that just raised prices, or refunded heavily in one month, shows the gap. Dividing the two is a coverage indicator, not an accounting split.
SELECT count(*) FROM startup WHERE status='published' AND mrr > revenue_last30dMostly marketplace fees, managed services and one-off sales, where a monthly run-rate is not a meaningful measure. The page prints "MRR not applicable" and the API returns null — never 0, because a zero would drag every median and percentile with it.
SELECT sum(case when mrr>0 then 1 else 0 end), sum(case when revenue_last30d is not null then 1 else 0 end) FROM startup WHERE status='published'What this brief cannot say
It is not a market census, not profit, not retention, and not a benchmark for the SaaS industry. It describes this self-selected sample only.
How to cite
MoneyPeel, "Research brief No. 0", data as of 2026-11-11, accessed YYYY-MM-DD, https://moneypeel.com/research/2026-11
