GRR and NRR get reported together in almost every SaaS board deck. They sit side by side in the retention slide, defined in the appendix, tracked in the same model. The reporting is not the problem.
The problem is that NRR gets all the attention. It is the number in the pitch deck, the benchmark in the investor memo, the metric that defines category leaders. 120% NRR is a headline. 94% GRR is a footnote.
That emphasis is backwards — or at least, incomplete. NRR and GRR are not two ways of measuring the same thing. They are answering two different questions. NRR tells you how your existing customer base is growing. GRR tells you how stable it is. You need both answers to understand what your revenue is actually doing, and treating one as primary and the other as supporting detail means you are half-reading your own business.
What each number actually answers
Gross Revenue Retention answers one question: of the ARR you had at the start of the period, how much did you keep? Expansion is excluded. If you started with $10M and retained $9.2M before any upsell, your GRR is 92%. That number tells you how stable your base is — how much of what you built last year is still standing.
Net Revenue Retention answers a different question: after churn, contraction, and expansion, what is the net effect on that same starting base? If you retained $9.2M and added $1.4M in expansion, your NRR is 106%. That number tells you how efficiently your existing customer base is growing.
Both are real. Both matter. They are not interchangeable.
The number that gets the headline
NRR earns its prominence. A number above 100% means your existing customer base is growing without a single new logo — it is a powerful signal of product stickiness and expansion potential, and investors are right to care about it.
But NRR is a net number. It folds stability and growth into a single figure, which means a strong NRR can coexist with a deteriorating base — and the headline will not tell you that.
Consider two companies, both reporting 108% NRR:
Company A: GRR of 94%, expansion of 14%. Retaining most of its base, growing on top of it. The floor is solid.
Company B: GRR of 78%, expansion of 30%. Losing nearly a quarter of its base every year, offset by aggressive expansion from a shrinking pool of surviving accounts.
The net number is identical. The businesses are not. Company B has a retention problem behind an expansion headline — and NRR, on its own, will not show you that. GRR will.
That is not a criticism of NRR. It is an argument for reading both. The famous number tells you what your team did with the base. The quieter one tells you whether the base is worth building on.
Why you can't plan off NRR alone
NRR is a trailing summary. It tells you what happened. It does not tell you what is coming.
When you sit down to build next year's renewal forecast, NRR is not the number you reach for. You reach for GRR — specifically, the GRR of the cohort that is up for renewal in the coming period. That is the number that tells you how much of your renewal base you can reasonably expect to keep. NRR, which includes expansion, does not answer that question cleanly.
The same is true for headcount planning, quota setting, and the ARR bridge. If you are modeling next year's net new requirement, you need to know how much of the base will churn — that is a GRR question. NRR tells you the outcome after CSMs have worked the accounts, after expansions have landed, after the quarter has closed. It is a result, not an input.
Planning off NRR instead of GRR is like navigating off your destination instead of your current position. The destination is useful. But you need to know where you are first.
They belong together
This is not an argument against NRR. It is an argument for always reading both with equal seriousness.
GRR sets the floor. It tells you how stable the business is, how much of last year's revenue is still in the building, and what the renewal base actually looks like going into next year. NRR tells you what your team did with that floor — how much expansion landed on top of the retention, and whether the net effect is growth or erosion.
A healthy SaaS business needs both to be true at once: a GRR high enough that the base is stable, and an NRR high enough that the base is growing. A company with strong NRR and weak GRR is running a retention problem behind an expansion headline. A company with strong GRR and modest NRR has a stable floor and an expansion opportunity. Those are not the same situation, and they do not call for the same response.
The two numbers together tell you something neither can tell you alone: whether you are building on solid ground.
One source, two views
The reason GRR and NRR get reported inconsistently — or quietly disagree when someone cross-checks them — is that they are usually computed separately. GRR comes out of one report, NRR out of another, both referencing the same customers but not necessarily the same contracts, the same period boundaries, or the same definition of churn.
When they disagree, finance spends time reconciling two metrics that should be mathematically linked. They are both views of the same starting ARR — the only difference is whether you include expansion in the numerator. If they are computed from the same contract base, against the same snapshot, they cannot disagree. The arithmetic guarantees it.
That is how HarborOS computes them: GRR and NRR come from the same query against the same locked snapshot, so the two metrics are always consistent with each other and with the ARR bridge. The denominator is identical. The only difference is what goes in the numerator. There is no reconciliation because there was never a second source to reconcile against. If you want to understand how the underlying terms are defined, the glossary covers GRR, NRR, and the retention mechanics in one place.