Most startups skip the only step that makes growth data useful

You have dashboards. Maybe a lot of them. You check MRR every morning and MAU every Monday. You can rattle off your growth rate to two decimal places in an investor update.

And you still don't know whether your business is healthy.

That gap — between having numbers and understanding what they mean — is where most early-stage companies get stuck. Not because they lack data, but because they skipped a step: decomposing their growth into the components that actually drive it. The step is called growth accounting, and it's the foundation that makes everything else in your data stack worth building.

The growth accounting framework in two minutes

Growth accounting breaks your topline metric into the forces pushing it up and the forces dragging it down. For users, that means four buckets:

New — people who showed up for the first time this period.

Retained — people who were active last period and came back.

Resurrected — people who had churned but returned this period.

Churned — people who were active last period and disappeared.

The math: change in active users = new + resurrected − churned.

For revenue — which is what matters for B2B SaaS — the buckets expand: new MRR, expansion MRR (existing customers spending more), contraction MRR (existing customers spending less), churned MRR (customers who left), and reactivation MRR (customers who came back).

Jonathan Hsu at Social Capital built this into a single diagnostic number when his team was evaluating hundreds of startups for investment: the quick ratio. Quick ratio = (new + resurrected) / churned. Above 4 is strong for enterprise SaaS. Below 2, you have a churn problem that no amount of acquisition will outrun. Below 1, you're shrinking — full stop — regardless of what your topline chart says.

A spreadsheet handles all of this. The question is why so few founders actually do it.

Why this gets skipped

Modern BI tools make it dangerously easy to confuse visibility with understanding.

You sign up for Hex or Metabase. You connect your warehouse. By 4 PM you have fifteen charts — MRR over time, user growth, a cohort heatmap that looks impressive in a board deck. The experience feels productive. You can see the lines going up.

But none of those charts decompose anything. They show the topline and the trend. They don't show you what's inside the number.

A real scenario: a founder sees 10% month-over-month net user growth and calls it strong momentum. The board agrees. The investor update is optimistic. But underneath that 10%, new user growth is 30% — and churn is 20%. The company is acquiring users fast and losing them almost as fast.

That's not growth. That's a leaking bucket with a bigger hose. You can run on that treadmill for a year before the economics catch up. And by then, your CAC has compounded while your retention never improved.

The only way to see this is to take the number apart.

What you actually find when you decompose

The first time a founder runs growth accounting, they almost always find something hiding in plain sight.

Groove, a SaaS helpdesk startup, was growing but noticed 4.5% monthly churn made their unit economics unsustainable. When they finally decomposed user behavior — rather than watching the topline — they found that users whose first session lasted under 35 seconds were almost guaranteed to leave. Users who stayed past 3 minutes and 18 seconds almost always stuck around. Two radically different populations hiding inside one aggregate "new user" number. The fix wasn't more acquisition spend. It was rebuilding the first-run experience to get users past that 3-minute mark.

Or take a pattern I've seen repeatedly at early-stage B2B companies: $80K MRR with 12% month-over-month growth. The pitch deck looks great. Decompose it:

• $18K new MRR from first-time customers

• $5K expansion from upgrades

• $9K churned MRR from customers leaving

• $2K contraction from downgrades

Net new is $12K — that's the 12% growth. But gross loss is $11K. Almost everything coming in is simultaneously walking out the door. A few happy customers expanding their usage are papering over a retention crisis that will compound as the base grows.

Without decomposition, this company raises a Series B on "12% MoM growth." With it, they fix onboarding, reduce churn from 11% to 6%, and net growth jumps to 18% — without acquiring a single additional customer.

The number that tells you if your growth compounds

Net revenue retention (NRR) is the clearest output of honest growth accounting. It measures whether your existing customers are growing or shrinking — independent of new sales.

NRR above 100% means existing customers expand faster than they churn. The B2B SaaS companies that compound — the ones investors fight over — run NRR above 120%. At that rate, your installed base generates meaningful revenue growth on its own. Acquisition accelerates it. You're not dependent on it.

Below 100%, every new dollar of acquisition needs to replace lost dollars before it contributes to growth. The lower your NRR, the harder and more expensive growth becomes. It's the difference between pushing a boulder uphill and riding one downhill.

Growth accounting gives you the component parts to calculate NRR with precision. Without the decomposition, NRR is just another number someone reports in a board meeting — calculated differently by every team, benchmarked without context.

How to set this up before next Monday

You need three things. None require a data team.

A specific definition of "active." For B2B SaaS, this usually means logged in and performed a core action within the last 30 days. Not "has an account." Not "received an email." The definition needs to be precise enough that two people in your company would independently classify the same user the same way. If your product is a reporting tool, "active" might mean "ran at least one query." Write it down. That's your definition.

A stable user ID. This trips up more early-stage companies than you'd expect. If users access your product on web and mobile, you need one identifier across both. Without it, the same person shows up as "new" on mobile and "retained" on web, and your decomposition is fiction. Teams at Meta and Tencent made stable ID resolution a prerequisite for growth tracking — features didn't count toward performance reviews unless the logging was properly instrumented.

A consistent time window. Monthly is standard for SaaS. Weekly for high-frequency products. Pick one and don't switch — mixing windows makes cohort comparisons meaningless.

With those three pieces, build a growth accounting table. For each month: count new, retained, resurrected, and churned. Calculate the quick ratio. Do the same for revenue.

This is how we think about it at Lopus — the FDE engagement builds this framework and wires it into discovery chat so you can ask "break down my net growth this month" in plain English instead of maintaining a spreadsheet. But the framework itself is just math and definitions. You can start it this weekend.

The uncomfortable patterns that always surface

Two things I see almost every time an early-stage company decomposes for the first time:

Resurrection is doing more heavy lifting than acquisition. The user base isn't growing because of new users — it's growing because old users keep reactivating. That feels fine until you realize your current acquisition channels aren't pulling their weight. You're coasting on past investment, not building forward momentum.

Or: one segment is carrying the entire business. Growth looks distributed on an aggregate chart, but slice by plan tier, geography, or acquisition channel and 60% of retained revenue comes from a single pocket. That's a concentration risk no topline chart will reveal.

Growth accounting doesn't fix these problems. It names them. And you can't fix what you haven't named.

Start here, not with dashboards

The instinct is always to add more charts, more tools, more metrics. Every analytics product on the market is built to encourage that behavior — it's their business model.

Growth accounting goes the other direction. One framework. A handful of numbers. A clear picture of what's pushing your growth up and what's dragging it down.

Decompose first. Dashboard second. The founders who do this early don't just track better data — they make different decisions. Because they can see what's actually driving the number, and they know exactly where to push when it stalls.