Expansion is now 40% of net new ARR at the median B2B SaaS company, and a dollar of it costs half what a dollar of new-logo ARR costs. That is the whole argument for building the motion deliberately, and most teams still have not.
The pattern in the audits I run is consistent enough to describe as a pattern. A board deck says "land and expand" three times. The revenue mix says north of 90% of new ARR comes from new logos, expansion is in single digits, and net revenue retention sits just under 100%. That is not land and expand. That is land, ship the contract, and forget the customer until renewal panic three months before the date.
Most B2B SaaS teams talk about land and expand the way undergrads talk about networking. Everyone knows the words. Almost nobody runs the motion. The result is a growth model that needs to refill the bucket from scratch every year, because the bucket has holes and nobody is pouring more water in from the customers already inside it.
I've spent the last ten years building RevOps and GTM systems for B2B SaaS teams, currently as Founding GTM Engineer at Peec AI. Every land and expand audit I run finds the same five gaps. This post is the playbook, with the 2026 numbers attached to each part of it.
Share of net new ARR that comes from expansion at the median B2B SaaS company, rising to 44% in the low-growth cohort and past half above $50M ARR. Expansion also carries a $1.00 CAC ratio against $2.00 for new logos, so it is the cheaper half of the growth model as well as the larger one (2026 Aleph x Benchmarkit SaaS and AI Performance Benchmarks, 342 companies, full-year 2025 data).
What land and expand actually means in 2026
Land and expand is a GTM motion where the first contract is intentionally smaller than the account's potential. You win on one team, one product, or one use case. Then you grow inside the account through three levers: more seats, more products, and more usage.
The "intentionally smaller" part is what most teams get wrong. They sell whatever the prospect will buy, then call expansion the responsibility of customer success, who get a budget of zero and a comp plan that says nothing about it.
A real land and expand motion needs four things working together:
- Pricing that lets a small first contract grow without a re-negotiation.
- A first contract scoped to a single team, product, or use case so adoption is fast.
- An expansion owner with a quota and a comp plan, not just a CS manager with a feeling.
- An expansion signal system that tells the owner when an account is ready to grow.
Skip any one and the motion breaks. I've seen teams nail pricing and signal and still get 102% NRR because no human owned expansion. I've seen teams hire account managers and still flatline because pricing made every expansion a six-week procurement fight.
The benchmarks that actually matter in 2026
Before you build the motion, anchor on the numbers. NRR is the headline. The two numbers underneath it, gross retention and expansion share, tell you whether the motion is actually working or just covering for something.
The median NRR across 342 SaaS and AI-native companies on full-year 2025 data is 102%, and expansion accounts for 40% of net new ARR at that median, 44% in the low-growth cohort (2026 Aleph x Benchmarkit SaaS and AI Performance Benchmarks, June 2026). SaaS Capital's 2026 survey of more than a thousand private B2B SaaS companies lands in the same place, a 103% median for the $3M to $20M ARR band, with the 90th percentile at 117.9% (SaaS Capital, 2026).
The 118% number is not the elite bar, and this page used to say it was
Almost every land and expand article, including an earlier version of this one, quotes 118% as top-quartile NRR. It is not. 118% is the median for enterprise books, meaning ACV above $100K. Mid-market sits at 108% and SMB at 97%, so the median SMB SaaS shrinks every year before new logos are counted (Optifai, citing SaaS Capital segment data). Once you split the benchmark by contract size the single median stops meaning much, and an SMB product at 97% is at benchmark rather than failing.
Check which number you are being held to before you accept the target. Enterprise ACV and an SMB motion are twenty-one points apart at the median, and no comp plan closes that gap.
The number that changed this year is gross retention, not NRR
Median GRR fell to 84% in 2025, down four points from 88%, and the 75th percentile fell from 95% to 91%. The drop shows up in every quartile, not in a tail of bad companies (The SaaS CFO on the 2026 benchmarks, Signal).
That pairing is the thing to sit with. A 102% NRR sitting on an 84% GRR means expansion is buying back eighteen points of churn before it adds a single point of growth. If your own numbers look like that, the land and expand motion is not your growth lever yet. It is your patch. Build it anyway, since the alternative is a shrinking base, but fix the leak in parallel and do not report the NRR number to your board without the GRR number next to it. The net revenue retention guide covers how the two are calculated and where the definitions usually slip, and the gross revenue retention guide takes the 84% number apart: the quartile spread behind it, the publishers who still report 88% to 91%, and why contraction counts against you even when the customer stays.
If you are below $20M ARR and expansion is under 20% of new ARR, you are running a hunter motion that calls itself land and expand. The fix is not a CS hire. The fix is upstream.
Why most land and expand motions stall
Five reasons, in order of how often I see them.
1. The first contract is too big
Sales reps are paid on first-year ACV. So they sell as much as they can on day one. The prospect signs a 5-team, 200-seat enterprise deal. Six months later, 40 seats are activated and the buyer is dodging renewal calls because they cannot defend the spend.
The "land" in land and expand has to land. That means a contract small enough that the buyer can prove value in 90 days without a politicking. If the first contract requires a board meeting, your land motion is broken.
2. Pricing is built for procurement, not for growth
Annual contracts with per-seat pricing, fixed at signing, paid up front. That is a procurement-friendly model and a growth-hostile one. Every seat add is a contract change request. Every usage spike is a phone call.
Pricing architecture is the strongest structural predictor of NRR in the 2025 data. Usage-based companies post a 108% median NRR against 98% for seat-based, a ten-point gap, and the 75th percentile for usage-based models reaches 155% (2026 Aleph x Benchmarkit benchmarks). Seat-based at 98% is below the break-even line, which is a new and uncomfortable place for the default SaaS model to sit.
Adoption is moving the same way but slower than the pricing commentary suggests. Around 38% of SaaS companies run usage-based pricing in 2026, up from 27% in 2023, and hybrid models with a seat base plus a metered layer sit near 43% today with forecasts around 61% by the end of the year (Culta, ValueAdd VC). Read those adoption numbers loosely: each report defines usage-based differently, some counting only pure consumption models and others counting any product with a usage element.
The mechanism matters more than the adoption curve. Usage-based revenue grows without anybody signing anything, and seat-based revenue does not, which is also why seat models are exposed right now: a customer consolidating roles or handing work to an AI agent shrinks their own seat count without ever churning. If your contract has no expansion path that avoids legal review, expect a flat NRR.
3. CS owns expansion but has no quota
Pick one. Either CS owns expansion with a quota, a target, and a comp plan that pays on it. Or CS owns adoption and renewals, and a separate account manager or expansion AE owns growth.
The hybrid "CS owns relationship and feeds expansion to sales" model works at companies above $100M ARR with mature ops. Below that, it produces a queue of warm leads that nobody picks up, because CS does not get paid to qualify them and AEs do not get paid to chase them.
4. No signal system
Most expansion happens because a customer asks. That is not a motion. That is order taking. A real expansion system tracks product usage, account changes, and adjacent team adoption, and routes a signal to the owner when conditions are met. No system, no proactive expansion. In the books I have looked at, reactive expansion lands somewhere just above the median and stays there, which is a judgment from engagements rather than a published benchmark. The published version of the same point is that the median company sits at 102% NRR, and the median company does not run a signal system.
5. The product does not have a second act
Sometimes land and expand fails because there is nothing to expand into. Single product, single user persona, no usage variable. In that case, no comp plan or signal system fixes the problem. You need a second product, a second persona, or a usage meter inside the existing one. Pricing and packaging is the real lever, not sales motion.
If your CS team owns expansion with no quota, nobody owns expansion. Pick a model and pay for it.
Hybrid ownership without comp alignment is the single biggest cause of stalled land and expand motions I see at Series A and B B2B SaaS.
The land and expand motion that works
Here is the four-stage motion I build with B2B SaaS teams that want expansion pulling its median weight, which is 40% of net new ARR, rather than whatever arrives unasked.
Stage one: land small
The first contract should be the smallest commercial unit that delivers measurable value in 90 days. In practice that usually means a single team license, no platform fee, a short minimum term, and a stated adoption gate that has to be met before the 12-month renewal conversation happens.
The trade the founder has to accept is a smaller number on the first invoice in exchange for a larger one at month 24. It works because the win rate on a right-sized first deal is higher and the account renews on evidence rather than on hope. I do not have a published, consent-cleared client number to put behind that claim, so treat it as the pattern I see rather than as a benchmark. The published part is the CAC gap above: expansion ARR costs about half what new-logo ARR costs, which is the arithmetic that makes landing small rational. If you want that argument in unit-economics terms, the LTV:CAC guide and the CAC payback page work the same numbers from the acquisition side.
The sales comp plan has to reflect this. If reps get paid on first-year ACV only, they will not sell small. Build a 12 to 24 month vesting structure where expansion revenue inside year one accrues to the original closer, then transitions to the expansion owner. That keeps the rep honest about right-sizing the first deal.
Stage two: activate
CS has 60 to 90 days to get the customer to a defined activation milestone. Not "they logged in." A real adoption metric. For a CRM data tool: enrichment runs scheduled on at least 5,000 records. For an analytics product: three custom dashboards in active use. For an AI product: 70% of seats used the product in the last 14 days.
If activation fails, the expansion motion fails. There is no point in signaling expansion to an owner if the underlying account is not adopted. I see teams skip this gate constantly. They push expansion plays to accounts that are not even using the base product, and the owner burns the relationship.
Stage three: signal
This is where most teams fall apart, because it is the part that needs RevOps work, not just sales work.
The signals you want to track:
- Usage threshold crossed. Account hit 80% of seat limit, 90% of API quota, or 70% of the included usage tier.
- New team activity. A new team or department started using the product. Even one user from a new domain or department code.
- Champion role change. Your buyer got promoted or moved to a parent company. Both are expansion triggers.
- Adjacent product event. A relevant integration was set up. A feature was used that maps to your second product.
- Account growth. Headcount grew 20%, a new round was raised, a competitor was named in a recent earnings call.
You build this in HubSpot (or whatever CRM is the source of truth) with custom properties on the company object, fed from your product database via reverse ETL and from enrichment sources like Clay. If you want the depth on reverse ETL, see the reverse ETL playbook. For the enrichment side, the Clay waterfall guide covers the data layer.
The signal then routes to the expansion owner via HubSpot workflows or, for more complex orchestration, an n8n flow. The output is a task with the account, the signal, the suggested play, and a 5-day SLA. Not a Slack notification anyone can ignore. A task with a name on it.
Stage four: expand
The expansion play depends on the signal. Five plays cover 90% of motions:
- Seat expansion triggered by seat limit. Owner reaches out, references usage data, proposes the next tier. Close cycle: 2 to 3 weeks.
- Usage tier upgrade triggered by API or consumption threshold. Often automated with an in-product prompt, then human follow-up for negotiation.
- Team expansion triggered by adjacent team activity. Owner reaches out to the new team's leader, references the existing internal champion.
- Product expansion triggered by feature usage signaling fit for a second product. Owner schedules a discovery on the second product use case.
- Renewal uplift triggered by approaching renewal with healthy adoption. Owner pre-empts the renewal with a multi-year deal at a higher ARR.
Each play has a defined trigger, a script, and a target close cycle. Build them in HubSpot sequences or your sales engagement tool, but tie the play directly to the signal that fired it. Generic outreach to expansion accounts is the same as generic outbound: low reply, low close, burned relationships.
Comp plan that actually drives expansion
I will keep this section short because it is the most contentious part of the motion. The teams that get expansion right pay for it directly.
The single biggest shift I have seen is paying expansion AEs (or account managers, name is not the point) a full quota on net new expansion ARR. Same OTE as a new business AE, sometimes higher base because they own the customer relationship.
The math works because expansion deals convert on incumbent advantage. Reported win rates for expansion and renewal business run around 40 to 60%, against 15 to 30% for enterprise new business, and expansion deals close faster than new logos by most published cuts of the data (Development Corporate, Zenit Data 2026 win-rate benchmarks). Put that next to the $1.00 expansion CAC ratio against $2.00 for new ARR and the case is settled on arithmetic: the expansion rep closes more of what they touch, sooner, at half the acquisition cost. Quota attainment is easier, and the ARR behind it is cheaper.
For a deeper look at how to structure these comp plans without breaking the budget, the sales compensation playbook walks through the math.
The RevOps backbone
Land and expand is a RevOps problem dressed up as a sales motion. The data plumbing matters more than the script.
Here is the minimum stack:
- CRM with custom objects for account-level expansion tracking. HubSpot or Salesforce both work. If you are evaluating, the HubSpot vs Salesforce 2026 comparison covers the trade-offs.
- Product database synced into CRM via reverse ETL (Hightouch, Census, or a custom n8n flow). Usage data on the account record updates in real time.
- Enrichment layer for account context like headcount, funding, and tech changes. Clay is the standard.
- Signal routing in workflows. HubSpot workflows for simple cases, n8n for multi-step orchestration with external triggers.
- Reporting on expansion as its own pipeline. Separate from new business. Different stages, different velocity, different conversion model.
If you want the broader picture of how to assemble this without buying every tool on the market, see the minimal RevOps tech stack and our CRM and RevOps work.
The four changes, in the order they usually have to happen
A note on what this section is not. We publish client outcomes only when the client has cleared the numbers for publication, and no expansion engagement is cleared today, so there is no NRR-went-from-here-to-there story on this page. What follows is the sequence itself, which is the part that transfers anyway.
- Reprice the first contract. Lower the entry point and put a metered layer on top of the seat base, so the account can grow without a contract change request. This is the change that takes longest to agree internally and gates everything after it.
- Name one expansion owner. One person, quota on net new expansion ARR, OTE matched to new business. Not a shared responsibility, not a CS side-project.
- Build the signal system. Product usage piped into the CRM on the company object, five or six defined signals, each routed to the owner as a task with an SLA. Reverse ETL if you have it, a weekly export if you do not.
- Align comp last. New business AEs vest on expansion inside the first year so they stop over-selling the land. CS bonuses attach to the activation milestone, not to a satisfaction score.
Do them in that order. Hiring the expansion owner before the pricing can carry expansion gives that person a quota they cannot hit, and a signal system built before anyone owns the output produces a dashboard nobody opens. The build cost of stage three is usually measured in weeks of RevOps time rather than in software, which is the part founders consistently get backwards.
Running a flat NRR despite a growing customer base?
Book a free 30-minute audit and we will map the three changes most likely to move your expansion number this quarter.
Book an audit →FAQ
What is a healthy NRR for B2B SaaS in 2026?
The 2025 median is 102% across 342 companies, and SaaS Capital's survey of private B2B SaaS puts the $3M to $20M band at 103% with the 90th percentile at 117.9%. Judge yourself by contract size rather than by that single median: enterprise books (ACV above $100K) sit at 118%, mid-market at 108%, SMB at 97%. Below 100% means the existing base is shrinking before new logos are counted. Whatever your NRR, read it next to gross retention, where the median fell to 84% in 2025.
Should the same AE who closed the deal own expansion?
It depends on company size. Below $10M ARR, yes, the same AE often owns expansion because there is not enough volume to justify a dedicated role. Between $10M and $50M ARR, separate the roles and pay both. Above $50M ARR, build a full account management function with named accounts and territory rules.
How small should the first contract be?
Small enough that the buyer can prove value in 90 days without a board meeting. As a rule, the first deal should be below the buyer's personal signing authority where possible. For most mid-market buyers that is $25K to $50K. For SMB, $5K to $15K. Land small, grow inside.
Is usage-based pricing required for land and expand?
No, but the data says it is the single biggest structural advantage available. Usage-based companies post a 108% median NRR against 98% for seat-based on 2025 actuals. Usage-based or hybrid models grow contracts without re-negotiation, which is the biggest friction point in seat-only expansion. Pure seat-based pricing can still work if you have a clear team-to-team expansion story and a fast contract amendment process, but you are starting ten points down.
How do I build the signal system if I am not technical?
Start in the CRM. Define five signals on the company object as custom properties. Feed them manually from a weekly product usage export. Once the plays prove out, invest in reverse ETL to automate the data flow. Skip the big platform purchase until you have proven the signals actually drive closed expansion ARR. We cover the build pattern in our AI automation work.
What to do this week
Pull three numbers: NRR for the last four quarters, gross retention over the same period, and the percentage of net new ARR coming from expansion. The benchmarks to hold them against are 102% NRR, 84% GRR, and 40% expansion share at the median, adjusted for your contract size. If expansion is below 25% of new ARR, the motion is broken. If NRR looks fine but GRR is in the low eighties, you have a churn problem wearing an expansion problem's clothes, and stage two is where to start.
Pick one of the four stages above where the gap is biggest and fix that first. Do not try to rebuild all four at once. The teams that succeed pick the highest-impact stage and run it for 90 days before touching anything else.
A note on the numbers. The benchmarks on this page were taken from search result extracts of the reports linked above rather than from a direct read of each publisher's page, so treat them as reported rather than verified first hand. The Aleph and Benchmarkit figures were corroborated across independent passes, and where sources disagree, as they do on usage-based pricing adoption, the page says so. Confirm anything headed for a board deck at the source.
If you want a second set of eyes on where your motion is leaking, reach out. We will walk through your numbers and tell you the one change that would move expansion ARR most in the next quarter.