How do you split renewal-team comp between CSM and AE in 2027?

The Tuesday morning that exposes the problem

Picture a $180K ACV account sixty days from renewal. The CSM has run the account for two years, drove adoption from 40% to 78% of seats, and has a champion who returns her Slack messages inside an hour. Three weeks ago that champion mentioned the security team wants the compliance module — call it $70K incremental. The CSM logs the opportunity, builds the business case, and gets verbal agreement.

Then the AE who originally landed the logo sees the opportunity in the pipeline report, notices it is now a $250K total contract value event, and asks the sales manager whether he should "take point on the commercials." The sales manager says yes. The CSM finds out when the AE forwards a redlined order form to her champion, cc'ing her.

What happens next is the thing every renewal-team comp design is actually built to prevent. The champion gets two points of contact with two slightly different pricing narratives. The order form sits for eleven days while the CSM and AE argue in a private channel about who is booking the expansion. The customer's procurement lead, who was ready to sign, now has time to ask whether they should run a competitive look. The deal closes — but forty-one days later than it should have, at a 9% deeper discount than the original proposal, because the AE conceded on price to end the internal fight faster.

Nobody behaved badly. The CSM was protecting her number. The AE was protecting his. The sales manager made a defensible call. The failure was structural: there was no written rule saying who owns a $70K expansion inside a renewal, so ownership got negotiated per deal, in the middle of the deal, with the customer watching.

That is the exact failure a banded split eliminates. Before the renewal opens, both people already know the answer. A $70K expansion on a healthy account is CSM-led with AE support at roughly 70/30 on the expansion component. The CSM runs the customer conversation. The AE joins the pricing and paper conversation. Renewal credit — the full $180K — never leaves the CSM regardless. There is nothing to argue about, so there is no eleven-day gap, no dual-narrative confusion, and no discount conceded to end an internal dispute.

The cost of the unwritten rule is not theoretical. In practice, ad-hoc credit allocation adds somewhere between two and four weeks to a renewal cycle that involves expansion, and it does so on the deals that matter most — the biggest ones, where expansion is present and the CSM-AE overlap is unavoidable. If you run 400 renewals a year and 30% of them carry expansion, that is 120 deals bleeding two to four weeks each. It also produces a slower, quieter cost: CSMs stop surfacing expansion signals, because surfacing one is how they lose the account conversation to an AE.

Two clarifying details from that Tuesday scenario are worth holding onto, because they drive the whole design. First, the CSM found the opportunity. Any comp rule that punishes her for that is a rule that guarantees fewer opportunities get found. Second, the AE genuinely added something — enterprise procurement negotiation is a real skill, and a CSM who has never run a redline cycle with a Fortune 500 legal team will give away terms she does not know she is giving away. A good split respects both facts instead of picking a winner.

How the banded split actually works

The mechanism is a decision tree evaluated once, early, and then locked. The evaluation happens at the renewal-planning checkpoint — typically 120 days out for enterprise, 90 for mid-market, 60 for SMB — and the band assignment gets written to the opportunity record as a field, not left in someone's head.

The four bands. Band one is the flat renewal: no incremental expansion, no competitive threat. The CSM owns it end to end and takes 100% of renewal credit. No AE is staffed, no AE credit is paid. This is typically 55–65% of a renewal book, and it should be the cheapest motion you run.

Band two is renewal plus expansion under $25K. Still CSM solo — 100% of renewal credit and 100% of expansion credit. The reasoning is throughput: a $15K seat add does not need enterprise negotiation, and routing it through an AE adds a handoff that costs more in cycle time than the AE's skill adds in price realization. Band two is usually 20–25% of the book.

Band three is renewal plus expansion between $25K and $100K. Renewal credit stays 100% CSM. Expansion credit splits roughly 70% CSM / 30% AE. The CSM owns the customer relationship and the narrative; the AE owns pricing structure, discount approval strategy, and contract mechanics. This is the band where most comp designs break, because it is the band where both parties genuinely contribute. Band three is roughly 10–15% of the book.

Band four is renewal plus expansion over $100K, or any renewal — regardless of expansion size — where a competitor is actively bidding. Renewal credit stays 100% CSM. Expansion credit inverts to roughly 70% AE / 30% CSM. The AE leads the commercial motion; the CSM stays in the room for continuity, product context, and champion management. Band four is 5–10% of the book but often 30–40% of the incremental revenue.

The one non-negotiable rule. Renewal credit never splits. Ever. Under any band. The CSM who carried the account gets full renewal dollars whether the expansion is $0 or $400K. The moment you split renewal credit, you have told the CSM that her retention work is contestable, and she will respond rationally by disengaging from the renewal motion — which is the one motion nobody else in the company is running. Organizations that split renewal credit tend to see gross retention degrade by several points within two to three quarters, and the degradation is hard to reverse because it is a trust problem, not a mechanics problem.

The competitive-risk override. Band four triggers on competitive risk even when the expansion is small or zero. A $200K flat renewal where the incumbent's contract is being shopped is not a CSM-solo motion — it is a defensive sales cycle, and it needs someone who negotiates against alternatives for a living. In that specific case, and only that case, the renewal itself carries an AE-led split. Keep this exception narrow and require a named competitor plus evidence (an RFP, a pricing request, a stated evaluation) before it can be invoked, or every CSM with a hard renewal will claim competitive risk to pull in AE support.

Locking the band. Once written, the band should be difficult to change. Allow exactly one re-evaluation trigger: expansion scope materially changes — meaning it crosses a band threshold by more than 25% — before the proposal goes out. After the proposal is with the customer, the band is frozen. Otherwise you recreate the exact mid-deal ownership fight the model exists to prevent, just with more paperwork.

Who arbitrates. A RevOps deal desk, staffed at director level, owns band disputes with a published SLA — five to ten business days for a standard call, same-week for anything blocking a signature. The deal desk does not negotiate the split; it interprets the written rule. If it finds itself inventing a new rule, that is a signal the playbook needs an amendment, which goes into the quarterly review rather than getting decided ad hoc.

The numbers that make a split defensible

Comp design arguments get won with arithmetic, not philosophy. Here is the arithmetic you need in front of the CFO.

Comp cost per expansion dollar. This is the single number finance cares about. Compute it per band. In band one, expansion comp cost is zero — there is no expansion. In band two, if your CSM expansion commission rate is around 8–10%, your comp cost is 8–10% of expansion revenue, all of it CSM. In band three at a 70/30 split, with CSM at roughly 10% and AE at roughly 15%, you pay 7% (CSM's 70% share at her rate) plus about 4.5% (AE's 30% share at his rate) — call it 11–12% blended. In band four at 70/30 inverted, you pay roughly 10.5% AE plus 3% CSM, or about 13–14% blended. The pattern is intentional: comp cost rises with deal complexity, because the deals that need AE involvement are the ones where a point of price realization is worth more than the incremental commission.

Run that math before you set rates, not after. A common mistake is picking split ratios first and discovering afterward that band four costs 19% of expansion revenue because both rates were set independently by two different functions.

Cycle time by band. Set explicit targets so you can detect comp-induced drag. Reasonable targets: flat renewals close in 14–21 days from initiation; band two in 21–30 days; band three in 30–45 days; band four in 45–60 days, longer if procurement and legal are heavy. Track actuals monthly. If a band runs more than 20% over target, the first hypothesis should be a comp problem — usually one party has no credit stake and is deprioritizing the work. The classic tell is band-one renewals slipping because an AE was informally pulled in and had zero reason to move fast.

Book composition. Audit twelve trailing months and classify every renewal into a band retrospectively. You need this before you set thresholds, because thresholds that land badly against your actual distribution produce absurd outcomes. If 40% of your expansions cluster at $22K–$28K, a $25K threshold means half your book will be arguing about which side of the line a deal landed on. Move the threshold to where the distribution is thin — maybe $35K — rather than where the round number is.

Quota construction. CSM quota should be built from two components: a retention target expressed in gross retention dollars against the assigned book, and an expansion target that reflects the realistic expansion capacity of that book. If a CSM's book is 70% flat renewals, giving her the same expansion quota as a CSM with an expansion-rich book is a design error that reads as favoritism. Rebuild books and quotas together, quarterly.

Time investment. Plan capacity around observed effort: a CSM typically spends 8–15 hours per renewal when the split is clear, rising sharply — 25 hours and up — when ownership is contested. An AE typically spends 3–8 hours on a band-three expansion and 15–30 on band four. Use those figures to size the team. A CSM carrying 40 accounts with a band-three-heavy book is carrying meaningfully more work than headcount models usually assume.

Dispute rate as a health metric. Track the percentage of renewals that generate a band dispute. In the first month after rollout, expect something in the mid-teens as edge cases surface. By month three, a well-specified model should be in the low single digits. If it plateaus above 8%, your band definitions have a hole in them — find it by categorizing the disputes rather than resolving them one at a time.

Payment timing. Pay renewal and expansion credit within 30 days of contract signature. Longer cycles measurably erode engagement, and they do it asymmetrically: AEs, who are used to fast commission cycles from new business, disengage from renewal work first. Do not run renewal comp on a quarterly true-up if new-business comp pays monthly.

Comp ROI. Roll it up: total renewal-and-expansion comp spend divided into retained-plus-expanded revenue. A functioning banded model generally lands in the 3.5x–5x range. Below 3x, either your rates are too rich for the motion or your band thresholds are pushing too much volume into expensive AE-led territory.

What you give up, and what else you could do

Banded splits are not free, and the honest case for them requires naming the costs.

The administrative cost is real. Someone has to classify every renewal, write the band to a record, keep the classification logic current in the commission system, and arbitrate the edge cases. That is a meaningful fraction of a RevOps analyst's time — plan for 10–20% of an FTE at a few hundred renewals a year, more if your product mix is complex. Organizations under roughly 150 renewals a year often find the overhead exceeds the friction it removes.

Threshold gaming is real. Any bright line invites behavior at the line. A CSM sitting on a $105K expansion has a rational incentive to structure it as $98K now plus a $7K add-on next quarter, keeping it in band three where she takes 70% instead of 30%. You cannot eliminate this; you can only make it visible. Track the distribution of expansion sizes and look for an unnatural cluster just under each threshold. If you see one, the fix is usually a taper — a transition zone from, say, $90K to $110K where the split moves gradually from 70/30 to 30/70 rather than flipping at a single dollar amount. Tapers cost you simplicity and buy you honesty; whether that trade is worth it depends on how much clustering you actually observe.

The alternatives are legitimate options, not strawmen.

*Double credit* — pay both the CSM and the AE 100% of the expansion — eliminates every ownership dispute instantly. Nobody argues about a split that does not exist. The cost is comp expense roughly doubling on expansion-bearing deals, and finance will push back hard. It is genuinely the right answer in one situation: a company under about 40 sellers total, where the friction cost of a split exceeds the dollar cost of paying twice, and where the expansion volume is small enough that the absolute dollars are modest. It stops being viable somewhere around 100 quota-carriers.

*Pure CSM ownership* — the CSM owns renewal and all expansion, at every size, with no AE involvement — is clean and cheap. It works when your expansions are structurally simple: seat adds, tier upgrades, usage overages that price themselves. It fails when expansion requires multi-stakeholder enterprise selling, because CSMs generally have not been trained to negotiate against procurement and will leave margin on the table without knowing it. If your average expansion involves a security review and a legal redline, this model will quietly cost you more in discount than it saves in commission.

*Pure AE ownership* of all expansion is the mirror image. It maximizes negotiation quality and completely destroys CSM expansion motivation. The CSM who spots the signal has zero reason to log it. Net revenue retention degrades not because expansions fail but because they never get discovered.

*A dedicated renewal-manager role* sidesteps the split by creating a third function that owns all renewals and expansions, with CSMs on pure adoption metrics and AEs on pure new logo. It is genuinely clean and it is what many companies past a few hundred million in ARR converge on. The cost is headcount and an additional customer handoff — and handoffs are where relationships go to die. It also takes two to three quarters to stand up.

The trade-off nobody names. A banded split optimizes for predictability over optimality. On any individual deal, a smart manager assigning ownership case-by-case would sometimes beat the band. The band wins anyway, because it decides before the deal opens instead of during it, and because a rule people trust produces behavior a rule people negotiate never will. You are trading a few points of per-deal optimization for the elimination of an entire category of internal conflict. That is usually the right trade — but you should make it knowingly.

Where these models break, and how to prevent it

Five failure modes account for most of the damage.

Splitting renewal credit. This is the fatal one. A CFO looking at comp spend will eventually propose splitting renewal dollars on AE-led deals — it looks like an obvious saving. It is not. The CSM reads it as her retention work being contestable, and the observable consequence is disengagement from the renewal motion within a quarter or two, followed by gross retention slipping several points. Prevention: write "renewal credit is 100% CSM in all bands" into the plan document as a stated principle with the rationale attached, so the next person who proposes changing it has to argue against a reasoned position rather than an arbitrary one.

Ad-hoc allocation. Deciding ownership per deal, during the deal, is the default state most teams drift back into when the playbook gets stale. It reliably adds two to four weeks per expansion-bearing renewal and it teaches CSMs not to surface opportunities. Prevention: make band assignment a required field that blocks the opportunity from advancing past the planning stage. Systems enforce what documents only suggest.

No arbitration path. Without a named arbiter and an SLA, disputes escalate to VPs, consume leadership bandwidth, and get decided by whoever has more political capital — which teaches everyone that the real rule is influence. Prevention: name the RevOps deal desk as the sole arbiter, publish the SLA, and require that VPs redirect escalations back to it rather than ruling personally.

Undocumented edge cases. Four recur constantly and should be written down before rollout rather than litigated during it:

*Expansion discovered by the CSM.* If a CSM independently sources and qualifies an expansion before any AE involvement, she keeps a materially larger share than the band would otherwise give — often 100% up to the next threshold. Without this rule, the model punishes discovery.

*Mid-cycle expansion growth.* A renewal opens flat and an expansion appears at week six. Rule: re-band once if it crosses a threshold before the proposal goes out; freeze after. Write the freeze point explicitly.

*Multi-product renewals.* A customer renews product A flat and expands product B. Treat each product line as its own renewal for banding purposes. Trying to band the blended deal produces nonsense — a $500K flat renewal with a $30K expansion should not become an AE-led motion because the total is large.

*Marketing- or product-sourced expansion.* An in-app upgrade prompt or a campaign generates the opportunity. Neither the CSM nor the AE sourced it, and paying full rate to both is expensive. Reduce the total credit pool and split what remains per band; document the reduced rate up front so it does not feel like a clawback.

Rolling out without a parallel period. Going live on a new split without running the old and new calculations side by side for a month guarantees a payroll error in month one, and a comp error destroys more trust than the friction you were fixing. Prevention: month one audits the book and sets thresholds; month two builds the plans and runs parallel calculations; month three goes live with a 60-day grace period where a three-person committee — RevOps, CS, and Sales leadership — meets weekly on disputes. Publish the rules to the whole team in a single session, from the CRO, with the reasoning stated, before any of it takes effect.

Then review quarterly: dispute volume, cycle time by band, comp cost per expansion dollar, and the distribution of expansion sizes against thresholds. Amend the playbook once a quarter, never mid-quarter, so nobody's plan changes underneath them.

Related questions

Should the AE who originally landed the account keep any permanent claim on it?

No. Permanent landing claims create a tax on every future expansion and disincentivize the CSM entirely. Give the original AE credit at land, then release the account into the renewal-team model. If retaining relationship continuity matters, staff that AE into band-four deals by preference — but as a role, not an entitlement.

How do you handle a renewal where the CSM changed mid-cycle?

Credit follows the work, not the seat. Prorate renewal credit between the departing and incoming CSM by time-in-account during the renewal cycle, with a floor for the person who closes it — typically no less than 50% to the closer, since the last mile carries the risk.

Does this model work for multi-year renewals?

Yes, with one adjustment: band on annual contract value, not total contract value. A three-year $90K/year deal is a $90K renewal, not a $270K one. Banding on TCV pushes routine renewals into AE-led territory and inflates comp cost with no corresponding negotiation complexity.

What if the CSM has no sales training at all?

Then band two and three will underperform on price realization, and you should narrow them temporarily — drop the CSM-solo ceiling to $10K and route more volume to CSM-led-with-AE-support — while you train. Widen the bands back as capability builds. Comp design cannot substitute for missing skills.

Who should own the expansion forecast?

Whoever owns the deal in that band. Forecast ownership and comp ownership must match, or you get a forecast submitted by someone with no stake in its accuracy. Roll both up through RevOps into a single renewal-and-expansion forecast rather than maintaining separate CS and Sales views.

FAQ

What is a banded ownership model in plain terms?

It is a rule set that changes who owns a renewal based on the deal's size and risk, decided in advance rather than per deal. Flat renewals stay entirely with the CSM. Larger expansions progressively shift commercial leadership toward the AE. The bands are written down, applied mechanically, and arbitrated by RevOps when a deal sits ambiguously between two of them.

Why does the CSM keep 100% of renewal credit even when the AE leads the deal?

Because retention and expansion are different jobs. The CSM's renewal credit pays for two years of adoption work, escalation management, and relationship maintenance that made the renewal possible at all. The AE's expansion credit pays for the negotiation that happened in the final six weeks. Splitting renewal credit charges the CSM for accepting help, which is exactly the behavior you do not want to price.

Where should the thresholds actually sit?

Wherever your expansion size distribution is thin. The $25K and $100K figures are common starting points, not universal truths. Pull twelve months of expansion data, plot the sizes, and place thresholds in the gaps. A threshold sitting in the middle of a dense cluster will generate constant disputes and constant deal-structuring games at the line.

How long does a rollout take?

Roughly 90 days done properly: one month auditing the book and setting thresholds, one month building plans in the commission system and running parallel calculations against live deals, one month live with a weekly dispute committee. Compressing it below 60 days usually means skipping the parallel calculation, which is where payroll errors get caught.

What is the earliest sign the split is not working?

Rising cycle time on a single band while others hold steady. That pattern almost always means one party in that band has insufficient credit stake and is deprioritizing the work. Dispute volume is the second signal, but it lags — people complain after they have already slowed down.

Can a company run this without a formal deal desk?

Below roughly 150 renewals a year, a named individual in RevOps with a published response SLA is sufficient — a formal desk is overhead you do not need. Above that, and especially with a multi-product catalog, the volume of edge cases justifies a standing function. What matters is that the arbiter is named, reachable, and not one of the two disputing parties' managers.

Sources

• https://www.gainsight.com/blog/
• https://www.forrester.com/research/
• https://www.gartner.com/en/sales
• https://www.worldatwork.org/resources
• https://openviewpartners.com/blog/
• https://www.bain.com/insights/topics/customer-strategy-and-marketing/
• https://hbr.org/topic/subject/sales-compensation
• https://www.saastr.com/category/customer-success/
• https://www.mckinsey.com/capabilities/growth-marketing-and-sales/our-insights

Related on PULSE

• [How do you comp a hybrid AE/CSM who handles expansion in their book?](/knowledge/q13)
• [How do we transition comp plans when we move from transactional to land-and-expand?](/knowledge/q273)
• [What is the 2027 typical CSM comp plan with NRR component?](/knowledge/q12103)
• [How do you design a CSM comp plan — and should CSMs have quota?](/knowledge/q10889)
• [How do we comp reps on expansion/upsell deals alongside a CSM or account manager?](/knowledge/q271)
• [How do you design an AE sales pod (2 BDRs + 1 SE + 1 CSM) in 2027?](/knowledge/q12681)