Growth Stage Assessment: Reading the Evidence an ARR Number Can't Show

A growth stage assessment is a structured diagnosis of which growth stage a company is actually in, built from observable evidence, retention behavior, whether the acquisition motion repeats without the founder, CAC payback, growth rate against a dated benchmark, and org shape, rather than from a revenue milestone alone. Two companies can sit at the same $4 million in ARR and be in genuinely different stages: one still searching for a segment that keeps using the product, the other with a repeatable motion and a real scaling problem. The ARR number looks identical. The evidence underneath it doesn't.

Most teams skip the diagnosis and read their stage off a slide. They hit a funding round, a headcount milestone, or a strong bookings quarter, and declare themselves past product-market fit or ready to scale. That usually says more about what the team wants to be true than what's happening in the business, and it drives real decisions: hiring a VP of Sales before there's a motion to run, or building a pipeline sized for volume the company doesn't have. Getting the stage wrong doesn't just waste a quarter, it points hiring, budget, and process at a problem the company doesn't have.

This piece stays narrowly diagnostic. It isn't a catalog of what a growth model is made of, that's growth model components, and it isn't a guide to which tools run the motion, that's growth tech stack design. What follows is the question underneath both: which stage is a company actually in, right now, and what has to change because of it. For the structural picture this sits inside, start with what growth frameworks are.

Key Facts: Growth Stage Assessment Reality Check

• Sean Ellis's product-market-fit survey, first published in 2009, set the 40% "very disappointed" threshold that's now routinely misattributed to later growth writers who popularized it. (Sean Ellis, Startup Marketing Blog, May 2009)
• Reforge founder Brian Balfour argued a flattening cohort retention curve, not a survey response, is the harder behavioral evidence of fit: "If it flattens off at some point, you have probably found product market fit..." (Brian Balfour, 2013)
• "T2D3" (triple, triple, double, double, double) was coined by Neeraj Agrawal of Battery Ventures in 2015 to describe the annualized revenue growth path his firm's fastest-scaling portfolio companies followed, not a target every company should hit. (Neeraj Agrawal, Battery Ventures, 2015)
• SaaS Capital's 2026 survey of more than 1,000 private B2B SaaS companies found a median growth rate of 22%, down from 25% the prior year, a reminder that any single-year median is a snapshot, not a permanent rule. (SaaS Capital, 2026 Growth Rate Benchmarks)
• SaaStr's Jason Lemkin puts it at about 70% of first VP of Sales hires not making it past 12 months, an estimate from SaaStr's own observation of the market rather than a formal study, and it tracks with companies hiring a sales leader before the acquisition motion was repeatable. (Jason Lemkin, SaaStr)

Why an ARR Band Is a Bad Diagnostic

ARR is easy to report and compare, which is why it became the default way teams describe their own stage. But it's mostly a function of pricing and deal size, not evidence of what's true underneath. A company selling $2,000 contracts and one selling $80,000 contracts can both hit $3 million in ARR at very different points in real maturity: one with hundreds of data points about what works, the other with forty logos and an untested motion.

The SaaS growth stages model is a useful companion for how stages map to company milestones. This piece stays narrower: the specific evidence that should override a revenue milestone whenever the two disagree.

Every row describes something real happening in the business. None of them, alone, says which stage the company is actually in. The five evidence categories below do that job instead.

Signal One: What the Retention Curve Is Actually Doing

Retention behavior is the hardest evidence to fake, because it's what users do over time, not what they say they'll do. Plot the percentage of a cohort still active each week or month since signup. A curve that keeps declining toward zero means no segment has found a reason to stay. A curve that flattens at some non-zero level, even a low one, is the strongest behavioral signal of real value.

Self-reported evidence is a useful second data point, not a substitute. Sean Ellis's 40%-would-be-very-disappointed survey is useful precisely because it's cheap to run, but it measures stated intent, not behavior, and a segment can score well while its usage curve keeps declining. A strong score paired with a flattening curve is a real signal; a strong score with a still-declining curve usually means people haven't built the habit yet.

Companies still searching for that first flattening segment get more from the early-stage growth model than from a scorecard that assumes a stability that doesn't exist yet. This borrows the same read used in more depth in the conversion optimization framework and the growth metrics hierarchy. For how B2B curves differ from the consumer cohorts most writing on this topic was built around, see product-market fit for SaaS.

Signal Two: Does the Motion Repeat Without the Founder

A motion counts as repeatable when someone other than the founder can run it and get a similar result, not when the founder closes another deal. This gets missed constantly, because a founder closing five deals in a row feels like proof of a working motion. It's proof the founder can sell, using relationships and improvisation a new hire won't have on day one.

The test differs by type: for self-serve, the signup-to-activation funnel converts consistently without a founder onboarding each account by hand; for sales-led, a rep who isn't the founder closes at a comparable rate using the same playbook, pricing, and messaging.

The go-to-market framework covers how to choose and formalize a motion once this test passes; high-velocity sales covers what repeatability looks like for a high-volume, short-cycle motion.

Signal Three: CAC Payback and Unit Economics

Unit economics answer a question retention and repeatability can't: even if the motion repeats, is it worth running at this price. CAC payback, the months it takes gross margin from a new customer to cover its own acquisition cost, is the cleanest single number here, since it folds cost, price, and margin into one figure a team can track over time.

A company can have a repeatable motion and still be earlier than it thinks if payback is long enough to make growth cash-negative for years. SaaS Capital's 2026 spending survey of the same 1,000-plus company panel found companies with $3 million to $5 million in ARR spending roughly 12% of revenue on selling costs and 8% on marketing, a combined 20% of ARR on acquisition, a useful anchor for whether a company's spend sits in a normal range. Full mechanics live in CAC payback optimization; this piece uses payback as one input among several.

Signal Four: Growth Rate Against a Dated Benchmark

Growth rate matters only when compared against a benchmark that's named, dated, and sized, not a remembered rule of thumb. "T2D3," the triple-triple-double-double-double path Neeraj Agrawal described from Battery Ventures' own portfolio in 2015, gets quoted constantly as a target every SaaS company should hit. It described what the fastest-scaling companies in one venture portfolio did, not a median, and treating it as a floor sets most companies up to feel like they're failing a standard that only ever described outliers.

The more useful discipline is naming a real, dated benchmark, checking your rate against companies genuinely similar in size and stage, then re-checking every year, since the number moves. SaaS Capital's 2026 survey of over 1,000 private B2B SaaS companies put the median growth rate at 22%, down from 25% the year before, with equity-backed companies at 25% and bootstrapped companies at 20%. Those figures will differ again next year, which is why quoting last year's median as a permanent rule is a mistake.

The B2B SaaS growth framework covers how growth rate expectations shift across a full SaaS lifecycle in more depth than a single benchmark can.

Signal Five: Org Shape

Org shape, who the company has hired and in what order, is a lagging signal, but a real one, because hiring decisions reveal what leadership believed about the stage, whether or not that belief was correct. A company that hires a VP of Sales before its motion repeats without the founder is usually mistaking a founder's close rate for a scalable motion, and the org chart shows the scar.

That mismatch has a measurable cost. Lemkin's estimate that about 70% of first VP of Sales hires don't survive their first year tracks closely with hiring a sales leader before there was a documented motion to hand them. The org shape that fits a stage isn't about headcount, it's about which roles exist because the evidence justified them, versus which exist because a board deck implied they should.

When to hire RevOps and the enterprise pipeline model both cover later-stage org-shape decisions in more depth than fits here. Stage and operating maturity are also not the same thing, and they routinely diverge: revenue operations maturity covers what the operating system underneath a company can actually support, which is the constraint that decides whether a stage-appropriate motion is safe to add yet.

The Stage-by-Stage Diagnostic Table

Putting all five signals side by side is the actual diagnostic. A company should sit at roughly the same stage across most rows; when rows sharply disagree, that's itself useful information, since it usually means one part of the business has outrun the evidence supporting it.

What Breaks at Each Transition

Every transition breaks something that worked fine at the smaller size, usually invisible until it's already caused damage. Naming the failure in advance is cheaper than diagnosing it after the fact.

None of this is a reason to avoid moving forward, just a reason to name the specific failure before it happens, so the fix is a plan instead of a scramble. Growth model components and growth tech stack design are the next two reads once the stage above is settled.

A Scoring Rubric You Can Run in an Afternoon

This doesn't require new instrumentation, just an afternoon, the five signals above, and an honest scorer, ideally not the founder, since founders systematically overrate their own stage on the signals hardest to fake from the outside.

Score each signal 0, 1, or 2 using the anchors below, then read the total and the spread across categories, not just the sum.

A score of 7 or higher, with no single category at zero, is a reasonable case for treating the company as further along than an ARR number alone suggests. A low total, or a wide spread with one or two categories at zero, is the more common and more useful finding: it names exactly which piece of evidence hasn't caught up to what the team wants to believe. Re-run the same five rows every quarter, since a stage scored six months ago can drift out of date quietly.

Common Misdiagnoses

Three patterns account for most of the wrong calls teams make about their own stage: a real, positive event gets mistaken for a broader signal it never was.

A funded ramp gets read as product-market fit when a large marketing or hiring push produces real growth on a chart; retention and payback tell a different story once the spend slows, since growth built on subsidy was rented, not earned. One channel gets read as a repeatable motion when a single paid channel converts well and the team declares "we found our motion," when the real test is whether a second channel, run by someone else, gets the same result. A hiring spike gets read as scale when a larger sales team is treated as evidence of scaling, though org shape without retention and repeatability behind it is just spend.

Conclusion

A growth stage assessment earns its keep the moment it overrides a comfortable ARR-based story with an uncomfortable but accurate one: retention, motion repeatability, unit economics, a dated benchmark, and org shape read together instead of one milestone standing in for all five.

None of this is a one-time exercise. A stage accurate six months ago can have shifted, a channel that once repeated can stop, and a benchmark that looked generous last year can look tight this year. Companies that keep the assessment current treat it as a recurring discipline, not a label chosen once and defended long after the evidence moved on. Past the last rung on most stage ladders the bar changes again, because the audience reading the evidence becomes a public market rather than a board, which is the subject of the IPO-ready growth model.

Related Topics

• What Are Growth Frameworks
• Early-Stage Growth Model
• Growth Model Components
• CAC Payback Optimization
• Growth Metrics Hierarchy
• B2B SaaS Growth Framework
• Product-Market Fit for SaaS
• When to Hire RevOps