4% → 13%: The Contract Length Collapse

Recently, Jason Lemkin sat in a board meeting at a B2B AI startup that just crossed $100M ARR. The new VP of Sales was exasperated. "The good news? Everyone wants to buy. The bad news? They aren't sure what they want in AI next year. So no one wants to sign a long-term deal."

Lemkin published the moment in SaaStr on April 20. His framing was not "buyer pushback." It was "structural shift." The data behind it is what every CFO and CEO in B2B software needs to read closely.

ICONIQ's State of Go-to-Market 2026, based on a January 2026 survey of 150+ B2B software executives:

Buyers are deciding faster and committing for less time. The contract length your salesforce is fighting to defend is the variable buyers have decided is theirs to control.

This issue covers the contract length collapse — why it is happening, what it costs companies that fight it, and the model that wins instead. The Signal is the data. The Build is Snowflake's quarterly earnings. The Move is what to do before next Thursday. Reading time: about 12 minutes.

Buyers are not pulling back. They are pricing their own optionality, and your three-year contract is now a risk premium they refuse to pay.

The mechanism is straightforward. AI capability is repricing every B2B software category every six months. A buyer signing a three-year deal today is making a bet on which vendor wins their category in 2027. They are not compensated for taking that bet.

Chris Degnan, former CRO of Snowflake, said it directly in the ICONIQ report: companies want to give their employees choice in the tools they use, but they will not sign long-term contracts in an AI world where the best solution can change in a matter of months. Buyers want to keep their options open.

The pricing infrastructure has shifted to accommodate them. 48% of B2B software companies now use hybrid pricing as their primary model — a mix of subscription and consumption — with usage-based and outcome-based models rising. When the price the buyer pays is variable, the buyer's finance team has even less appetite for a multi-year commitment at fixed terms. AE compensation is moving with it: comp tied to Net New Recurring Revenue jumped from 25% of companies in 2025 to 33% in 2026, eight points in one year. The market is repricing the salesperson the same way it is repricing the contract.

The trap most boards fall into is fighting the shift with multi-year discounts. SaaS Capital's 2023 survey of 1,500 SaaS companies, as summarized by SaaStr, found that multi-year contracts have only a modest impact on NRR — they appear to defer churn more than prevent it. A discounted three-year deal is a customer who resents the commitment by month 18 and churns at renewal anyway.

The non-obvious read for a contract-signer: this is not a slowdown. Top-quartile ARR growth reaccelerated to 111% for sub-$50M ARR companies in H2 2025. NDR sits at 110% to 123% across revenue bands. Buyers are buying. They are expanding. They are doing it on shorter contracts because that is what they trust. The CEO question is not how to defend the three-year deal. It is whether the post-sale motion is strong enough that the renewal stops being a negotiation.

How Snowflake built a $4B+ business on the opposite of multi-year lock-in.

Snowflake's most recent SEC filing is the cleanest data point in the category. Q4 fiscal 2026, ending January 31, 2026: product revenue of $1.23 billion, 30% year-over-year growth, Net Revenue Retention of 125%. The same NRR they reported in Q1, Q2, and Q3 of the same fiscal year. Five consecutive quarters in the 124% to 126% range, per Snowflake's quarterly earnings releases.

Snowflake's model is not a traditional seat-based subscription. It is consumption-based, with capacity contracts where customers commit to a spending level over a term and consume against it. Their own filing language draws the distinction: "Our consumption-based business model distinguishes us from subscription-based software companies that generally recognize revenue ratably over the contract term." Customers consume above contracted capacity during the term and roll unused capacity into future periods at renewal.

Snowflake's customers are not locked into a seat-based subscription. They pay for what they use. They can scale down. They can leave. The cohort still grows spend by 25%, year over year.

Chris Degnan ran the sales motion that scaled this model from early stage through public-company maturity. Three operating principles map directly onto the contract length problem.

Land for the use case, not the seat count. Initial commitments are sized to a specific workload the customer is solving today, not to a long-term seat forecast. The first deal is small enough that procurement does not price it as risk.

Make consumption visible. Customers see what they spend in near real time. The renewal conversation is not a negotiation over contract length. It is a discussion of what the customer has used and what workload comes next.

Customer success as the engine, not the safety net. Expansion comes from new workloads onboarding into existing accounts, not from price escalators in renewal language. The post-sale team is built to find the next use case, not to defend the last one.

Snowflake's 125% NRR is structural — it is what happens when customers can leave and choose not to. Depth of consumption wins, not length of paper.

Three actions before next Thursday.

PwC's 2026 AI Performance Study, released April 13 and surveying 1,217 senior executives across 25 sectors, found 74% of AI's economic value is captured by just 20% of companies. The differentiator is not budget, tool count, or which vendors you picked. It is what your AI is pointed at, and whether the foundation under it can support what you are building. The contract length problem sits downstream of that question.

The 20/80 AI Growth Benchmark tells you which side of the split you are on. Fourteen questions across five dimensions drawn from PwC's 2026 AI Performance Study and IBM's Transformation Edge framework — Growth Orientation, Decision Clarity, Data Readiness, Autonomy Level, Governance Maturity. Twenty minutes. I score every submission personally and walk you through the result on a 15-minute call. No credit card. No email list. No sales pitch. One matched recommendation. If we are not the right fit, I tell you so.

The companies that figure this out first are the 20%. The ones that don't, won't.

Take the 20/80 Benchmark →

Next Thursday: the renewal cliff. What happens to the 2022-2023 multi-year deals coming up for renewal in a market that has been repriced underneath them.

Reply if your board has already started asking the contract length question.

Elizabeta Kuzevska Co-Founder, Revenue Experts AI https://revenueexperts.ai