Breakdown of Key SaaS Metrics

Why retention — not headline growth — reveals the real financial profile of a SaaS business

By Paul Inouye, Founding Partner, Western Hills Partners

For a founder, growth is the number everyone celebrates. It's also the number that hides the most. A SaaS business can post impressive MRR or ARR gains quarter after quarter while quietly failing at the one thing that determines its long-term value — keeping the customers it already has. That's why, when I evaluate a software business, I look past top-line growth to retention, and specifically to net revenue retention (NRR). Retention is where the truth lives. If customers aren't staying and spending more over time, it usually means the product isn't solving a real problem — there's no truly compelling value proposition underneath the growth. And that gap matters more with every quarter, especially once a company is past product-market fit and into its scaling phase.

Let's start with the basics.

The three retention metrics that matter

Three numbers tell you whether a business is actually holding onto its revenue:

• Customer (logo) retention — the share of customers you keep from the start of a period to the end. It counts noses, not dollars.
• Gross revenue retention (GRR) — the share of recurring revenue you keep from existing customers, excluding any expansion. It's the floor.
• Net revenue retention (NRR) — the same measure, but including expansion (upsell, cross-sell, added seats). Because expansion counts, NRR can exceed 100%; GRR and logo retention never can.

Hold onto that last point. It's the key to reading these numbers correctly.

MRR vs. ARR: which number should you trust?

A lot of founders treat MRR and ARR as interchangeable, and in a sense they are — both measure the recurring revenue base, just at different time resolutions. In most models, ARR is simply MRR multiplied by twelve. But as an operating signal, MRR is usually the one I'd trust more, for two reasons. First, granularity: MRR is measured every month, so it captures new bookings, expansion, contraction, and churn as they happen. ARR is a snapshot that annualizes a single month — and because it multiplies by twelve, one unusually strong month (a large annual prepay, a lumpy enterprise close) gets magnified into a full year of implied run-rate that may never repeat. Second, discipline: computing MRR forces you to isolate what is genuinely recurring, where loosely calculated ARR can quietly absorb one-time items like setup fees or services and overstate the base.

The nuance worth stating plainly is that this isn't absolute. For a business built on annual or multi-year contracts, ARR is the more meaningful and stable figure, because customers have actually committed for the term; the monthly number is a derived allocation. For month-to-month or usage-based businesses, MRR is the truer signal. The right way to hold both: MRR gives you more granular, timelier visibility into momentum and is harder to distort with a single strong month, while ARR is the cleaner headline for annual-contract businesses.

Think of it like driving. MRR is your speedometer — what the business is doing right now. ARR is the projection on the trip computer — where you'll end up if this exact speed holds for a year. Read only the projection, and one fast mile fools you into thinking you'll travel twelve times as far as you actually will.

Why NRR is the metric I care about most

If I could keep only one retention metric, it would be NRR, because it captures expansion — whether you're growing revenue inside your existing base, not just adding new logos. NRR above 100% means your existing customers are spending more over time before you win a single new account; that's the signature of strong product-market fit and real pricing power. Below 100%, your ARR decays — you're refilling a bucket that leaks, and every new sale is partly just replacing what you lost.

But NRR has a catch: it's a lagging indicator, and it's a net figure. A single number that nets several opposing forces together can hide as much as it reveals — which is why you can't stop at the headline.

Decomposing NRR: the headline can lie

NRR is built from four moving parts, measured on a fixed cohort of customers over a set period: the base recurring revenue you started with, plus expansion (upsell, cross-sell, added seats, price increases), minus contraction (downgrades, reduced usage), minus gross churn (customers who leave entirely).

Here's why the breakdown matters. Two companies can both report 105% NRR and be nothing alike underneath. Company A holds 95% gross retention and adds ten points of broad-based expansion — a sticky product with healthy, distributed upsell. Company B loses twenty points to churn and contraction but masks it with twenty-five points of expansion concentrated in a handful of accounts. Same 105%, radically different businesses. Company B is a leaky bucket papered over by aggressive upselling to a few customers — and if that expansion slows even slightly, or one big account walks, the whole number collapses.

That's why NRR should never travel alone. Pair it with GRR, which strips out expansion and shows the floor — how much you keep before any upsell. The gap between GRR and NRR is your expansion engine. Read together, they tell you whether a strong retention number rests on a durable base or on a few accounts you can't afford to lose.

The analogy I use: NRR is your weight on the scale — one number. It won't tell you whether you gained muscle or fat. GRR, churn, and expansion are the body-composition breakdown that tells you whether that number is actually healthy.

What good looks like

Benchmarks move year to year, but recent private-SaaS research from SaaS Capital puts median net revenue retention right around 100%, with top-quartile companies near 110% or better and best-in-class businesses north of 120%. Gross retention tends to cluster near 90%, running higher for enterprise-grade products with larger contracts and dedicated support. Retention also strengthens with deal size — higher-ARPA businesses generally see lower churn and more expansion — and the very best retention profiles hold up remarkably well regardless of company stage.

The metrics that sit around retention

A handful of metrics live close to retention and are worth watching alongside it. ARPA (average revenue per account) shapes almost everything downstream — sales-cycle length, contract tenure, onboarding cost — and higher ARPA tends to pull NRR up with it. On the acquisition side, CAC matters less as a raw number than as a payback period: how many months of recurring revenue it takes to recover what you spent to land the customer. The SaaS quick ratio — new plus expansion revenue divided by churned plus contracted revenue — gives you a single read on the quality of your growth: how much you're adding for every dollar you're losing. And to track momentum over time, compound monthly growth rate — CMGR = (ending value / beginning value) ^ (1 / number of months) − 1 — smooths the lumpiness that any single month can introduce.

Why this matters if you might ever sell

Retention isn't only an operating metric — it's one of the strongest predictors of the multiple a software business commands at exit. Acquirers and growth investors scrutinize GRR and NRR precisely because they can't be faked by a single good quarter; durable retention is often worth more than an extra point of growth. A founder who can show a high, well-decomposed retention profile — a solid gross floor with a genuine expansion engine on top — negotiates from a position of strength. That's the difference between growth that looks good on a chart and growth that's actually worth paying for.

If you're building toward that outcome, the time to understand your retention story is well before you need it.

Benchmark figures: SaaS Capital private-SaaS retention research (2025–2026). Figures are approximate and shift year to year.