Benchmark Update — October 01, 2026: The 2.7x Multiple, 12-Month Cohorts, and the Death of the Three-Year Lock-In

Benchmarks are not laws of physics. They are snapshots of what buyers, boards, and capital allocators are currently willing to tolerate. And as of this morning, several of those snapshots have been retaken — with worse lighting.

a16z Growth’s September State of Markets deck landed with a number that is going to get screenshotted into a hundred board decks this week: horizontal B2B SaaS now trades at 2.7x revenue. Not 2.7x forward ARR with a growth-adjusted premium. Just 2.7x. That is the floor becoming visible, and it reprices everything downstream — your burn multiple tolerance, your CAC payback ceiling, your LTV assumptions, and most urgently, how long you can afford to be unprofitable.

Here is what actually moved, and what to do about it before your Q4 planning cycle locks.

Signal 1: The Runway Benchmark Is Now the Scariest Number on the Board

The same a16z deck reports that 55% of unicorns have under 24 months of runway. Read that again in the context of the 2.7x multiple. A company with 18 months of cash, growing 40% year-over-year on a horizontal product, is no longer a “strong Series C candidate.” It is a company that needs to either hit profitability or accept a flat-to-down round inside four quarters.

The counterweight in the same dataset: a cohort of new startups growing 500%+. This is the bifurcation thesis made numerical. The market isn’t dead. It’s sorted. Top-decile growth gets frontier pricing — see “Instinct” hitting a $10B valuation in 33 days, which is velocity pricing we last saw in 2021. Everyone else gets the discount bin.

Benchmark revision: the old “default alive” standard of 18 months of runway is obsolete. In a 2.7x-multiple market where down rounds are repricing events rather than rescue events, model 30 months minimum if you’re in the horizontal middle. That’s not conservatism. That’s the cost of not having a forced-sale conversation at the bottom of the cycle.

Signal 2: Contract Length Collapsed — Rebuild Your Churn Model on 12-Month Cohorts

This is the most actionable benchmark shift of the quarter, and it came from SaaStr: stop pushing multi-year deals. Shorter terms are now the rational B2B default in the AI era. Buyers will not lock into a model architecture that may be obsolete two quarters from now — and after today’s GPT-6 “Astra” launch from OpenAI, that fear is validated, not paranoid.

The KPI consequences cascade fast:

• Gross revenue churn will look worse. Not because your product got worse, but because renewal events now happen annually instead of triennially. A cohort with 8% annual gross churn and three-year contracts reported something closer to 3% in any given year. That cosmetic cushion is gone.
• NRR leans harder on expansion. If best-in-class NRR was 120% with multi-year lock-ins doing half the work, you now need expansion revenue — seats, usage, modules — to carry 25–30 points on its own. Benchmark reset: NRR above 110% is the new “good” for mid-market B2B; 120%+ is now genuinely elite rather than merely competent.
• LTV compresses mechanically. LTV is ARPA divided by churn rate. Double your effective churn by halving contract length and you halve LTV. Every CAC payback target built on the old number is now wrong.

The CAC Payback Recalibration

The standard target has been 12 months for SMB, 18 for mid-market, 24 for enterprise. With shorter contracts and compressed LTV, those numbers need to tighten — particularly in enterprise, where 24-month payback assumed a customer who was contractually immovable for 36. If your buyer can walk at month 12, a 24-month payback means you can lose money on a full customer lifecycle.

Working target for Q4 2026: SMB 9 months, mid-market 15, enterprise 18. If you can’t get there on blended CAC, the answer is almost never “spend more on paid.” It’s pricing, packaging, or a narrower ICP.

Signal 3: Vertical Still Clears the Bar — The Fresha Proof Point

Against all of that gloom, Fresha raised $80 million from KKR at a $1 billion valuation today. A beauty and wellness booking marketplace. Not a frontier AI lab, not an agentic anything — a workflow-embedded vertical platform that owns the transaction.

This is the cleanest available evidence that the 2.7x multiple is a horizontal number, not a SaaS number. Vertical platforms that sit inside a transaction — payments, scheduling, inventory — carry structurally better retention because ripping them out means ripping out operations. Their NRR benchmarks hold up precisely because the expansion motion is usage-linked rather than seat-linked.

Same pattern in Lio’s $30M a16z-led Series A for AI procurement: the agentic wedge that works is the one embedded in a workflow with a measurable savings number attached, not a chat interface over a general model. OpenAI’s “Dots” always-on agent launch and Microsoft’s MAI-Transcribe-2 — which undercuts OpenAI, Google, and ElevenLabs on both price and speed — make the point from the other direction. API-layer margin compression is accelerating. If your gross margin depends on a model-provider price that nobody else can get, it isn’t a margin, it’s a lease.

The India Angle: A Working-Capital Squeeze Nobody Has Priced

Today’s wire carried no India-specific funding news — flagging that honestly rather than inventing it. But the contract-length shift is going to hit Indian SaaS harder than anyone is currently modelling.

Indian B2B SaaS has run on annual-prepaid as a capital strategy. Collecting twelve months upfront is how Bengaluru, Chennai, and Delhi NCR vendors funded growth without burning equity — it’s a structural advantage over US peers who discount heavily for annual terms. If global buyers normalise monthly or quarterly billing