Multi-Year SaaS Contracts Fell From 48% to 26% in One Survey Year. The Same Survey Says That's the Riskiest Direction to Move.
KeyBanc/Sapphire's SaaS survey shows multi-year contracts collapsing, right where its own churn-by-term data says they cut churn the most.
Ask a SaaS pricing team why they've stopped pushing multi-year contracts and you'll usually get two honest answers: buyers won't sign them anymore, and the discount doesn't cover the risk of locking up revenue that far out. Both show up clearly in the KeyBanc Capital Markets/Sapphire Ventures Private SaaS Company Survey, the industry's longest-running private benchmark of its kind. The share of SaaS companies closing multi-year deals fell from 48% to 26% in a single survey year — the sharpest one-year move the series has recorded. The same survey family also happens to be the clearest evidence anyone's published that multi-year terms are exactly where churn goes to die.
The collapse, in one table
The move is big enough that it's worth seeing both editions side by side rather than as a single percentage. The 2025 edition's full breakdown also shows where the lost multi-year share actually went: mostly into straight one-year terms, not into month-to-month, which stayed a small slice of the mix.
| Contract term | KBCM-2025 edition | KBCM-2024 edition |
|---|---|---|
| One year or less (combined) | 74% | 52% |
| — of which: one-year term | 67% | not broken out |
| — of which: month-to-month | 7% | not broken out |
| Multi-year (2+ years) | 26% | 48% |
A nearly 50% relative drop in multi-year adoption in one survey cycle isn't a gradual drift — something changed in how deals got negotiated and closed across an entire data year. And it happened at the same time as the trend we covered in our piece on co-terming, where a separate survey, Zylo's 2026 SaaS Management Index, found multi-year agreements growing as a share of the broader enterprise software stack. Those two numbers aren't actually contradictory once you separate what each one measures: Zylo is tracking what enterprise buyers hold across every vendor in their stack, while KBCM/Sapphire is tracking what SaaS vendors themselves are closing in new and renewing deals. It's entirely possible for buyers to be consolidating existing contracts onto longer, co-termed dates while vendors are simultaneously losing the fight to get new multi-year commitments signed in the first place.
Why the direction is backwards, according to the same survey family
The KBCM-2024 edition also published a churn-by-contract-term breakdown, and it's one of the more useful tables in any SaaS benchmark report precisely because it isolates one variable — term length — against annual gross dollar churn.
Source: KeyBanc Capital Markets/Sapphire Ventures Private SaaS Company Survey, KBCM-2024 edition, p.12.
Run the gap between the ends of that chart and multi-year terms cut churn by roughly 79% versus month-to-month — a bigger single-variable swing than almost any onboarding fix, pricing change, or customer-success hire is going to deliver in a year. That's the backdrop against which 22 points of multi-year share just evaporated from the market in twelve months. Vendors aren't choosing to give up the biggest churn lever they have access to. They're losing the ability to require it.
Why buyers won't pay for the protection anymore
The economics explain most of it. The same Zylo dataset that shows multi-year agreements growing across enterprise stacks also shows the discount for committing to one shrinking: 16.4% off for a 12-month term versus 13% off for 36 months, a gap of about 3 points for tripling the length of the commitment. A few years ago that math looked different — vendors would trade a meaningfully steeper discount for the longer lock-in, which gave a CFO a real reason to trade budget flexibility for savings. At a 3-point spread, the discount barely covers the cost of capital on paying further in advance, let alone the risk of being stuck in a contract if the product stops earning its renewal.
Layer vendor consolidation on top of that math and multi-year gets harder to justify from the buyer's side for a second reason, not just a pricing one. We've covered how consolidation reviews are running at scale across procurement teams right now — a three-year signature is a three-year commitment to a line item that might get flagged for cutting in year two of the term, with no annual off-ramp to act on that decision. A one-year or month-to-month default gives a buyer the same right-sizing option every renewal cycle instead of banking it against a hoped-for discount that's gotten too thin to be worth the inflexibility.
The benchmark itself isn't as stable as it looks
There's a second warning buried in the same survey series, and it matters for anyone benchmarking their own churn against it: the numbers get revised, and not in the vendor's favor. The KBCM-2024 edition reported 2022 and 2023 gross dollar retention at 89% for both years — churn of roughly 11%. The KBCM-2025 edition, fielded a year later with more time for slow-resolving cancellations, disputed invoices, and non-renewals to actually finish playing out, restated those same two years down to 86% and 85%. The churn rate for 2023 effectively grew from 11% to 15% in hindsight, a 4-point miss on a number that had already been published as final.
That's not a one-off data error — it's a structural property of any survey that asks respondents to report current-year churn before the year's slowest cases have fully resolved. If you're using an industry benchmark as a ceiling to measure your own number against, the newest edition of that benchmark is usually the most flattering version it will ever look, and the real comparison point shows up twelve months later once the restatement lands.
What this means for your own renewal calendar
If your own book is drifting toward shorter default terms — whether by buyer pressure, a sales team optimizing for faster close rates, or just following the market — the practical consequence isn't abstract. Shorter terms mean more renewal events per account per year, and the churn-by-term data says each of those events carries a meaningfully higher failure rate than the multi-year renewal it's replacing.
- Model the churn delta against your own discount schedule, not the industry average. Run your actual term mix through our dollar retention rate calculator to see what shifting even 10 points of your book from one-year to two-year terms would be worth in retained ARR, before you decide whether a deeper multi-year discount is worth proposing again.
- If you're defending a multi-year default against buyer pushback, lead with the churn gap, not a stability pitch. Procurement teams respond to a specific number — "three-year terms churn 79% less than month-to-month in the industry's own data" lands differently than "we think longer relationships are healthier."
- Treat every shorter-term account as a higher-frequency renewal risk, not just a smaller one. The same CRM discipline we described in our piece on auto-renewal notice windows — tracking the real notice deadline, not just the renewal date — matters more when a book is skewing toward one-year and month-to-month terms, because there are simply more of those deadlines to miss across the same number of accounts.
- Don't quote last year's published churn benchmark as this year's ceiling. Given how often these numbers get restated upward, build in a buffer when setting internal churn targets against an external survey, rather than treating the freshest published figure as the bar to clear.
The mechanism underneath all of this is the same one that shows up in a self-serve cancellation flow, just compressed into a much shorter cycle: fewer renewal decision points generally mean fewer chances to lose the account, and more of them mean more. A SaaS company whose book is quietly shifting toward shorter terms is accumulating exactly the kind of renewal exposure that a monthly subscriber already lives with every single billing cycle — which is also why the reason captured at each one of those decision points matters more, not less, as they get more frequent. CancelFlow exists for that compressed version of the same problem: catching the specific reason behind a cancellation at the moment it happens, instead of losing that signal in an annual churn number that won't even be final until next year's survey restates it.
Frequently asked questions
Are SaaS contracts getting shorter in 2026?+
Yes, sharply. The KeyBanc Capital Markets/Sapphire Ventures Private SaaS Company Survey — the industry's longest-running private benchmark — shows multi-year contract adoption falling from 48% of deals in its 2024 edition to 26% in its 2025 edition. The share of companies running on terms of one year or less rose from 52% to 74% over the same single survey year, the sharpest move in the series.
Does contract length actually change a SaaS company's churn rate?+
The same survey family's contract-term breakdown says it's one of the largest single levers available: roughly 14% annual gross churn on month-to-month terms, 10% on one-year terms, 6% on two-year terms, and 3% on terms of three years or longer. Moving a cohort from month-to-month to a three-year term cuts churn by about 79% on that data, a bigger swing than most product or customer-success initiatives will ever move in a single year.
Why don't SaaS companies just require multi-year contracts if they churn less?+
Because the discount buyers get for committing longer has shrunk to the point where it barely compensates for the extra budget risk. Zylo's 2026 SaaS Management Index puts the average discount at 16.4% for a 12-month term versus 13% for 36 months — a roughly 3-point gap for locking up three times the commitment. Add a procurement environment actively hunting for vendors to cut, and a three-year signature is a harder ask than the churn math alone would suggest it should be.
Can I trust a SaaS churn benchmark the first time it's published?+
Treat it as provisional. The KBCM-2024 survey reported 2022 and 2023 gross retention at 89% for both years. The KBCM-2025 edition, published a year later with more resolved data, restated those same two years down to 86% and 85% — churn that looked like 11% at the time turned out to be 14-15% once the slow-resolving cancellations and non-renewals had finished working through the numbers. The newest edition of any survey is usually the most optimistic one it will ever look.
Stop losing subscribers today
One script tag. One function call. A live cancellation flow in under 10 minutes.
Start free trial →