Inside the SaaS efficiency gap
The widening gap between early-stage startups and everyone else
👋 Hi, it’s Kyle Poyar and welcome to Growth Unhinged, my weekly newsletter exploring the hidden playbooks behind the fastest-growing startups.
In the last eight years I've surveyed 4,000+ private SaaS companies to figure out what's going on in the industry. Never have I seen benchmarks go out of date so quickly.
Let's separate hype from reality and find out what's really going on today. Take the 9th annual SaaS benchmarks survey here.
This report is super special to me. I'm so excited to continue the journey with my friends at High Alpha. The full report will drop in the Fall and all participants will get early access to the data (I’ll also share it here).
THANK YOU for your support. Now, on with the newsletter.
Let’s rewind the clock to 2022. The Oscars slap happened. Elon bought Twitter. Wordle became a thing.
Back then the typical <$1M ARR software startup had 13 people. “T2D3” was still the name of the game. Hiring a big team was a flex; in fact, the highest quartile of <$1M ARR SaaS companies had 38 (!) people.
Well, software startups have changed. As of last year, the typical <$1M ARR software startup had just 7 people. Expectations around ARR per employee went through the roof. AI agent workflows have become the thirst trap of LinkedIn (and perhaps this newsletter?). The new flex is reaching the top of the Lean AI leaderboard.
I thought now is a good time to shed preconceived notions and investigate the new SaaS org chart, by the numbers. The data comes from the most recent SaaS benchmarks report, which covered 800+ software companies including both VC-backed and bootstrapped companies.
The widening gap between early-stage startups and everyone else
Zooming out, there’s an increasing divergence between early-stage startups and everyone else.
Software companies used to get way more efficient as they scaled with a $50M+ ARR company typically making 2-3x the ARR per FTE of a $1-5M ARR startup. This was part of the magic of SaaS! But it’s no longer consistent with what’s happening.
More mature SaaS companies are facing dual crises of team bloat paired with decelerating growth.
Median ARR per employee was down slightly (-9% year-on-year) among businesses with $5-20M ARR.
It dipped even further (-17% year-on-year) for those with $20-50M ARR.
It fell materially for companies with >$50M ARR (-34% year-on-year).
Looking at public software companies, there’s a worrying level of inefficiency, too. Public SaaS companies ended Q1 2025 with an average gross margin-adjusted CAC payback period of 57 months based on estimated data from Clouded Judgement1. Said differently, it takes the best companies nearly five years to recover their sales and marketing spending.
This isn’t some extreme outlier quarter. The average CAC payback period over the past twelve quarters is now 41 months. CAC payback periods haven’t been healthy since Britney Spears’s conservatorship ended in 2021!
Public companies can afford for their GTM to be inefficient in the short-term. They have plenty of profitable legacy customers who can subsidize sales and marketing. But the subsidies can’t go on forever. And a reckoning could be on the horizon.
Early-stage startups have gone in the opposite direction. Early-stage teams are much smaller than in previous years and seeing far higher ARR per employee.
Among software companies with <$1M ARR, median ARR per full-time employee (FTE) increased 75% year-on-year to $71k. It increased by 33% year-on-year for those with $1-5M ARR and is now at $120k. Interestingly, growth rates have actually increased year-on-year despite (or because of?) the small team sizes.
The rise of the two-pizza startup
You might’ve heard of Amazon’s famous two-pizza teams. Today we’re seeing the rise of the two-pizza startup.
The median <$1M ARR SaaS startup now has 7 full-time employees — small enough to be fed by two extra-large pizzas. This is down from an average of 12 in 2023 and 13 in 2022. The range (bottom quartile — top quartile) is fairly tight, between 5-15. Here’s a snapshot of a typical org chart.
Engineering is consistently the largest team, representing 43% of headcount on average. At this point companies usually have one team member in product and design, one in sales, and one in customer success and support. They might have a marketer on the team, especially if they have a PLG strategy, but often rely on founder-led marketing along with support from fractional or agency resources.
The median $1-5M ARR SaaS startup has 25 full-time employees. This is also down materially from 34 in 2023 and from 46 (!) in 2022. The range (bottom quartile — top quartile) is between 16-40 people. Here’s a snapshot of a typical org chart.
Product and engineering remains the largest function, representing about half a typical $1-5M ARR startup. A major focus is building out the go-to-market side, moving from founder-led sales to a repeatable GTM. GTM roles quadruple from an average of 2.5 people to roughly 10 including: 3-4 in sales, 2 in marketing, 3 in customer success and support, and 1 in professional services.
Part of this shift toward lean teams could be attributed to what Elena Verna calls the rise of the AI-native employee. Another likely factor is the Series A crunch highlighted by Peter Walker from Carta.
The Carta fundraising data shows that the median ARR at Series A is now $3M, up by 100% compared to 2021 when it was $1.4M ARR. Seed to Series A graduation rates have fallen; only 13% of SaaS companies graduate from Seed to Series A in two years or less. Early-stage funding needs to stretch for much longer than before — which necessitates going further with a small team.
Where startup hiring is heading
It’s clear that the old spreadsheet-based approach to startup hiring is in the rearview mirror. The new goals are to minimize burn and maximize speed.
What I’m seeing in the next batch of breakout startups:
Fewer management layers of people. Even senior hires are player-coaches who do their share of IC work.
More management layers of AI agents. Each team member will have their own "org chart" consisting of a mix of AI agents, contractors, advisors and (maybe) full-time hires.
Hiring fewer, but higher agency people. There’s a premium on those with side hustles along with former founders — people who are used to tackling unfamiliar challenges. These folks all need to be what Eric Simons of bolt.new calls ‘high context’ so they can make decisions independently.
Generalists over specialists. Leaner teams means fewer specialists. A marketer might own content and paid programs. Or community and partnerships. A GTM hire might look more like a GTM engineer doing a mix of pipeline generation, RevOps, and growth.
A premium on AI fluency. This is becoming a requirement for more and more roles. Not everyone needs to be an AI automation engineer, but they’ll increasingly be measured based on their ability to apply AI to what they’re doing.
We might have reached this point partly out of necessity, but small teams are proving to be surprisingly effective. They can ship releases at a blistering pace, ride the wave of improving LLMs, automate processes from day one, and be a first-mover as new opportunities arise.
The real question: will legacy software companies will be able to maintain the pace?
Public software companies typically do not report their CAC payback period. The metrics shown here have been derived with the following formula: [(Previous Q S&M) / ((Current Q Subscription Rev x 4) - (Previous Q Subscription Rev x 4)) x Gross Margin] x 12. Read more in Clouded Judgement.
Love it!
Excellent article. What I'm seeing (my main company does staff aug for tech startups) is that companies under $5mm or $3mm ARR are building leaner AI-augmented team. I think long gone are the days of building 10 person teams who build MVPs in 12 months.