95/102 beats 85/110

Two retention numbers describe every MSP. Owners quote the flattering one. Buyers price the other — and the gap between the two is where a surprising amount of enterprise value quietly leaks out.

Two MSPs. The first has 95% gross revenue retention and 102% net revenue retention. The second has 85% gross retention and 110% net. Which is the better business?

On a slide, the second one wins. Net retention of 110% is the number SaaS founders brag about — it says the base is expanding, upsells are landing, the machine is compounding. Most owners, asked for one retention number, quote NRR, because it is the flattering one.

A buyer prices the first business higher. Sometimes much higher.

Why the honest number is the gross one

Net retention blends two things a buyer badly wants kept separate: how much revenue you kept, and how much you expanded. An 85/110 business lost fifteen percent of its base and papered over it with expansion — often expansion concentrated in a handful of accounts. A 95/102 business kept nearly everything and grew modestly on top.

The entire premise of an MSP's multiple is that the recurring revenue recurs. That is what a buyer is paying six, eight, ten times earnings for — the belief that the revenue will still be there after the wire clears. Fifteen percent gross churn says some meaningful part of it will not be. No amount of upsell changes what that says about the underlying asset; it just changes who the risk is concentrated in.

This is not exotic buyer knowledge. It is standard diligence. What is remarkable is how rarely owners hold the distinction themselves — partly because most tools report a blended number, and partly because two erosions that gross retention catches are nearly invisible in a normal P&L:

Seat-level contraction inside retained clients. Nobody churns. No logo is lost. A 40-seat client quietly becomes a 31-seat client across three renewal cycles, and the revenue erodes without a single event that would trigger anyone's attention.

The difference between logo churn and revenue churn. Losing ten five-seat clients is noise. Losing one 200-seat client is a diligence event. A blended churn percentage treats them identically; a buyer does not.

What to do with this

Compute your GRR honestly — revenue retained from the base, no expansion counted, trailing twelve months. If you cannot compute it from your PSA without an afternoon of spreadsheet work, that itself is worth knowing, because a buyer's analyst will compute it in the first week of diligence, from your own data, whether you have ever looked at it or not.

Then hold the two numbers side by side. If the gap between your NRR and your GRR is wide, you are growing on top of a leak — and buyers price the leak, not the growth.