Expansion revenue needs an owner. Who should carry it?

Three people claim the same upsell in the comp review. The CSM spotted the need, the account manager wrote the order form, and the AE who closed the logo two years ago says the customer still calls him first. Finance pays two of them, and net retention does not move.

The money in that fight grows with the company. Benchmarkit’s 2026 report puts expansion at a median 40% of new ARR, and at 62% for companies above $100 million in ARR. High Alpha’s 2025 benchmarks found companies above $50 million in ARR generate roughly 60% of new ARR from existing customers. ChurnZero’s 2026 study of 580 customer success and sales leaders put the median share of total revenue from existing customers at 70%.

Outside the AI leaders, net retention sits close to 100%. Benchmarkit’s 2026 median is 102%, while ICONIQ’s Pacesetters, the top-quartile growers that are AI-native or AI-driven, post medians of 120% under $100 million in ARR and 130% above it. Public SaaS companies sit in between: the ones Jamin Ball tracks in Clouded Judgement showed a median of 110% in his Oct. 2, 2026 update. At 100%, a company books exactly enough expansion to replace what its churned and downgraded customers took with them.

Expansion is also the easier sale, so an upsell with no owner is an expensive miss. In Ebsta’s 2025 benchmarks, produced with Pavilion before Fullcast bought Ebsta, expansion deals won 45% of the time against 19% for new business, and closed in 52 days against 91. Clari’s March 2025 study of 10 million opportunity records from more than 100 enterprises found expansion deals made up 5% of pipeline but drove 13% of won revenue per quarter.

Ebsta also found accounts with ongoing engagement were 1.9 times as likely to expand. The person who keeps showing up finds the next deal, and four ownership models can put that person in place.

CSM-owned

The customer success manager carries renewal and expansion targets for a book of accounts. It fits mid-market books where expansion means more seats or a higher tier on standard paper, and where the CSM already runs the relationship.

Pay a variable tied to gross retention on the book, with expansion commission that starts only above a retention floor. The risk is trust: a CSM with a quota walks into each health check as a seller, and a customer with a problem may stop raising it.

Account manager-owned

A dedicated account manager carries renewal and expansion quota, and the CSM stays off quota to run adoption. It fits enterprise and upper mid-market accounts that negotiate their renewals and buy more than one product, as long as the contract can fund a second person on the account.

The AM’s plan pays on net expansion bookings against a quota built from the book’s whitespace, gated by renewal rate, while the CSM earns on adoption and gross retention. The cost is the handoff chain: the customer meets an AE, then an AM, then a CSM, and explains its business again to each one.

AE-owned

The AE who closed the logo keeps the account and sells the expansion. It fits enterprise land-and-expand deals where each expansion looks like a new sale into a new business unit, and early companies without an AM layer. Ebsta’s 2025 benchmarks found 46% of companies adopting full-cycle selling, where one seller prospects, closes and nurtures the account through its first 12 months.

Count expansion inside the AE’s quota at a lower rate than new logos, and make renewal attainment a gate on the payout. The Bridge Group’s 2026 study puts median AE quota at $960,000 and found 48% of reps hit their annual number. An AE who is behind in the last week of the quarter will work the new logo and let the renewal slide unless the gate is there.

Pooled or hybrid

The CSM owns gross retention and flags expansion for an expansion desk or AM to close, and the whole pod shares a net retention target. It fits product-led and usage-priced businesses, where expansion arrives through product use more than through a sales cycle.

Benchmarkit’s 2026 report puts median net retention at 108% for usage-priced companies and below 100% for seat-priced ones, and the usage a customer grows on its own is the expansion least in need of a commission. The same report has product-led companies at 79% median gross retention against 88% for sales-led ones, which is why the CSM in this model owns retention and hands expansion to someone else. Pay the pod a shared bonus on net retention, and pay the desk on assisted upgrades only.

Pay for each dollar once

Ownership fights end in the comp plan, so write the answer there. Four rules apply whichever model you pick.

• One quota-carrier per dollar. Each expansion dollar pays one owner at full rate. Overlays such as the CSM or sales engineer get a smaller share, written into the plan before the quarter starts.
• Gate expansion pay on retention. Accelerators pay only when the owner’s book clears a gross retention floor. Gainsight advises putting gross retention first when it sits below roughly 88% or has trended down for two straight quarters. Benchmarkit’s 2026 median of 84% is already below that line.
• Separate the renewal from the upsell. Pay renewals at a low rate or treat them as a gate, and pay expansion at a rate that rewards new scope.
• Leave the weather out. Escalators and true-ups that fire on their own, and usage growth nobody sold, stay out of commissionable expansion unless a person changed the contract.

How it breaks

• Double pay. The CSM and the AM both take full credit for one upsell, and cost of sales climbs with no new revenue behind it.
• Orphaned renewals. The first person to open the renewal does it 30 days out, with a price increase the customer has not heard about.
• Masked churn. A few large expansions lift net retention while small logos leak away. In Benchmarkit’s 2026 report, median net retention ticked up to 102% in 2025 while median gross retention fell four points, from 88% to 84%. Gainsight warns that a pure NRR focus pulls energy toward your largest accounts.
• The handoff cliff. The AE disappears at signature and the customer goes quiet, which throws away the ongoing engagement Ebsta found makes expansion 1.9 times as likely.

Choosing a model

You can settle most of the choice with two questions. Does a seller create the expansion, or does the product? And is the average contract big enough to fund a commercial owner on top of a CSM?

• Product creates it, contracts too small for a second owner: pool it. The CSM owns gross retention and an expansion desk handles assisted upgrades.
• Product creates it, contracts big enough for a second owner: an AM owns the renewal commit and the consumption plan, and the CSM drives adoption.
• Sellers create it, contracts too small for a second owner: the CSM owns the book with commercial training and a retention gate, and AMs cover only the accounts big enough to fund one.
• Sellers create it, contracts big enough for a second owner: the AE keeps the account for the first 12 to 18 months, then hands it to a strategic AM.

Before choosing, answer the first question with your own numbers. Match last year’s expansion in finance’s ARR bridge against the closed-won expansion opportunities in your CRM. Set aside price escalators and upsells a rep closed but never logged, and the dollars left without an opportunity are expansion the product created. If those dollars make up most of your expansion, you belong in one of the two product rows.

Pull the next 50 renewals on the calendar and write one name beside each. Where you need two names, you are about to pay twice for one dollar. Pick the name before the comp plans go out, because after that, finance pays both.