What's the right way to handle territory disputes between AEs without killing morale in 2027?

The Tuesday morning that costs you two reps

Picture a 40-rep SaaS org, second week of Q3. An inbound demo request lands from a 900-employee logistics company headquartered in Columbus with a distribution center outside Reno. Marcus, the Midwest AE, has three logged calls with the VP of Ops from a conference last quarter. Priya, who covers the West and owns the named-account list for logistics, has a signed NDA with the same company's procurement lead from a deal that stalled in February. Both are looking at the same opportunity record. Both are right.

What happens next determines whether you keep both of them. In the average org, what happens is a Slack thread, then a DM to the VP of Sales, then a hallway conversation, then a decision that arrives four days later with no stated reasoning. Priya gets the account because she's at 112% and Marcus is at 71%. Nobody says that out loud, but everyone on the floor infers it within a week.

The damage is not that Marcus lost. Reps lose accounts constantly and shrug. The damage is that thirty-eight other AEs just learned the actual decision rule — attainment buys accounts — and that rule is nowhere in any document. From that Tuesday forward, every rep's behavior changes: they log speculative touches on accounts they have no intention of working, they escalate earlier and louder, and the ones who don't have the VP's ear start quietly updating their LinkedIn.

Now run the same Tuesday through a system. The routing layer reads the account against the named-account list, sees logistics is a designated vertical with Priya as owner, and assigns it to her automatically at 9:04am — before either rep has time to form a grievance. Marcus gets a notification explaining which rule fired. If he thinks the rule was applied wrong, he files a dispute in a queue everyone can see, and a decision with written reasoning lands within two business days. He may still lose. But he lost to a published rule, applied identically to the rep at 112% and the rep at 71%, and he can read the reasoning and predict the next one.

That difference — predictable loss versus arbitrary loss — is the entire subject. Everything else in this answer is mechanics for producing it.

The scenario generalizes well beyond net-new account claims, which is worth naming early. The same failure pattern shows up in renewal-versus-expansion ownership fights between AEs and Customer Success, in SDR meeting-credit disputes when two reps worked the same buying committee, in partner-sourced deals where a channel manager and a direct AE both registered the opportunity, and in professional-services attach where a solutions consultant expects credit for a deal they technically unblocked. If your Rules of Engagement only cover AE-to-AE geography, you have solved maybe half the surface. The teams that get this right write one document that adjudicates all the boundaries where revenue credit can be claimed by more than one person.

How the mechanism actually works

The operating model has six moving parts, and they only work as a set. Skip one and the failure mode is predictable enough that you can name which one is missing from the symptoms alone.

The scoring formula. Every account carries a numeric score built from five to seven weighted inputs — a workable set is revenue potential, ICP fit, growth signal, geography, and existing rep relationship. The weights are published. A rep can pull up any account and reconstruct its score in a few minutes. Keep it under seven variables: past that, the model becomes both opaque and gameable, and a formula nobody can verify is functionally identical to no formula at all. Practitioners at Pavilion and similar RevOps communities converge on the same test — if a rep can't recalculate their own book, the model is broken regardless of how defensible the math is.

The Rules of Engagement document. This is the constitution. It answers, in writing and in advance: does named-account assignment beat geography (almost always yes), who owns expansion versus net-new, how long a deal registration holds, what counts as a qualifying touch, and how credit splits when more than one rep legitimately contributed. Every AE signs it at hire and re-acknowledges it each quarter. Version history stays public. There are no silent edits.

The routing layer. The RoE becomes code here — in native CRM territory management or a dedicated routing tool sitting on top of it. This is the step most orgs skip, and skipping it is why so many well-written RoE documents rot in a shared drive. A rule that a human has to remember and apply is a rule that gets applied inconsistently. A rule encoded in the routing engine fires the same way at 9am on a Tuesday and 6pm on the last day of the quarter.

The escalation SLA. Some fraction of cases will genuinely be edge cases the rules don't cover. That's fine and expected. What matters is the clock and the owner: dispute filed, both reps submit CRM evidence within 24 hours, RevOps makes a preliminary call from the data by hour 48, and if it's still contested the CRO decides — personally, in writing, published. Delegation reads as avoidance to the floor. So does silence.

The freeze. No mid-quarter territory changes except for termination, leave of absence, or a performance-plan outcome. This is the cheapest rule in the entire system and the one most often broken, usually by a leader who is certain this particular exception is obviously fair. It never reads that way from the floor. Every rep watching an exception is now computing whether their own book is next.

The comp plan. If the plan pays disproportionately for contested accounts, reps will fight for them no matter how elegant the territory model is. Flatten payouts on cross-boundary deals, kill per-logo bonuses, put accelerators above quota so the marginal hour is better spent closing than arguing, and add a clawback window so nobody profits from claiming a bad-fit account and dumping the support burden.

The loop at the bottom is the part most implementations miss. Every escalated dispute is evidence that a rule was missing or ambiguous. If you resolve the case and don't feed the gap back into the RoE, you'll adjudicate the same shape of dispute again next quarter, and the floor will notice that escalation is a permanent feature rather than a shrinking exception.

Worth noting what the mechanism does to the manager's job. In the broken version, a frontline sales manager spends real hours per week on adjudication — hearing sides, brokering splits, managing the emotional aftermath. Those hours come directly out of coaching and deal inspection, which are the only two activities that reliably move a team's number. Managers who get pulled into dispute mediation are managers who stop reviewing pipeline. The system doesn't just protect rep morale; it gives the manager layer its week back.

What the numbers actually say

Territory work gets dismissed as soft process because the costs are diffuse. They aren't. They're just billed to accounts nobody watches closely.

Replacing a regretted AE. Industry analyst benchmarks generally put the fully-loaded cost of losing a productive rep at somewhere in the range of six to nine months of that rep's total compensation. For an enterprise AE with a base in the mid-six-figures on-target, that's a low-to-mid six-figure event per exit — recruiting spend, ramp time, manager hours, and the pipeline that goes cold while the seat is empty. Add the orphaned book: accounts under a departed rep's name typically see a meaningful productivity gap during the handoff and replacement ramp, often in the range of a third of prior output, and the successor rarely recovers the historical run rate inside two quarters even with clean documentation.

How much of that is territory-attributable. Analyst research on sales force effectiveness consistently finds that perceived territory unfairness is a top-tier driver of unwanted attrition — commonly cited in the neighborhood of a fifth to a quarter of regretted exits. The word doing the work in that sentence is *perceived*. Reps don't leave because they lost a dispute; they leave because they concluded the adjudication was rigged. Two orgs can produce identical dispute outcomes and get opposite attrition results based purely on whether the reasoning was visible.

Mid-quarter disruption. The cleanest finding in the whole space is that territory stability correlates with attainment. Reps whose books shift mid-quarter show markedly lower quota attainment than the stable cohort — the commonly cited SaaS benchmark data puts the median drop in the range of fifteen percentage points. Conversely, reps holding a stable territory across three or more consecutive quarters close at a measurably higher rate. The mechanism is unglamorous: relationship depth, account knowledge, and the compounding value of a second and third conversation with the same buyer.

Capacity ceilings. Two numbers govern whether a territory is workable at all. First, current-state ARR per AE — most SaaS orgs land somewhere in the range of eight to ten million in existing revenue per enterprise rep before coverage quality degrades, though this varies enormously by segment and motion. Second, active account count. The commonly cited saturation threshold sits around 150 accounts; past that, reps triage rather than work, and the accounts that get triaged out are a leading indicator of both logo churn and the rep's own frustration. A rep with 300 accounts isn't covering a large territory; they're covering forty accounts and ignoring 260.

Dispute frequency as a health metric. Track disputes as a percentage of opportunities created per quarter. Under three percent is generally healthy — a small edge-case tail is a sign the rules are specific enough to cover the normal cases. Above five percent, patching individual cases is wasted effort: the RoE itself has a structural gap and needs a rewrite, not another adjudication. This single ratio is the most useful instrument you have, because it moves before attrition does. Attrition is a lagging indicator by six to twelve months. Dispute rate moves within a quarter.

Split-credit conventions. There's no universal formula, but the patterns that show up repeatedly in practitioner guidance are worth knowing as starting points rather than laws: single-touch deals pay 100% to the closer; multi-touch deals commonly split along the lines of a majority to the closer, a meaningful minority to the prospector, and a small slice to the account of record; strategic accounts with genuine co-ownership often run an even split. Deal-registration windows typically run thirty to ninety days depending on sales-cycle length — shorter for transactional motions, longer for enterprise. Clawback windows on early churn are commonly set at twelve months.

The comp-variance test. A useful and under-used diagnostic: what share of your total variable-comp variance traces to disputed or split accounts? If it's more than a couple of percent, your comp plan is manufacturing the disputes your territory model is trying to prevent, and no amount of RoE tightening will fix it. Fix the plan first.

One caution on all of these numbers. Published benchmarks in sales operations come from self-selected survey panels, skew toward companies willing to report, and vary wildly by segment, motion, and sales-cycle length. Use them to size the problem and to sanity-check your own figures, not as targets to hit. Your own dispute rate measured over three quarters is worth more than any external benchmark, and it's the only one you can actually move.

The trade-offs nobody puts in the deck

Every design choice here buys something and costs something. Pretending otherwise is how you end up with a model that looks rigorous on the slide and gets ignored on the floor.

Named accounts versus geography. Named-account models produce far fewer disputes because ownership is explicit — the list either has the logo or it doesn't. The cost is that the list becomes the fight. Reps stop arguing about individual deals and start arguing about list composition, which is a slower, more political, and considerably more corrosive argument because it happens once a year in a room they're not in. Geographic models are self-evidently fair and cheap to maintain but generate constant boundary disputes at the edges — multi-site companies, remote buyers, and headquarters that don't match where the actual decision-maker sits. Most orgs above meaningful scale end up hybrid: named accounts for the top tier, geography for everything else, with the named list published and version-controlled.

First-touch-wins versus territory-map-wins. First-touch rewards hustle and is intuitively fair to reps. It also gets gamed within about six weeks of implementation — reps spray single LinkedIn connection requests across hundreds of accounts to establish claims. The fix is an activity floor: a claim requires several substantive logged touches inside a defined window, with content, not just a connection request. Territory-map-wins is un-gameable and produces near-zero disputes, but it demoralizes the rep who genuinely sourced an account outside their lines and watched someone else close it. Reasonable compromise: territory map governs, but sourcing outside your territory earns a documented finder's share.

Transparency versus flexibility. A fully published model with a public decision ledger is the strongest morale protection available. It also removes your ability to quietly make an exception when an exception is genuinely warranted — a strategic logo that needs your most senior rep regardless of geography, a rep coming off leave who needs a soft landing. You cannot have both. Choose transparency and accept that exceptions must be published with reasoning, or choose flexibility and accept that the floor will eventually discover the exceptions and price in the discovery. The second path is almost always worse, because undiscovered exceptions get discovered at the worst possible time, usually by the rep who lost a dispute the week before.

Speed versus thoroughness in adjudication. A 48-hour SLA is aggressive and occasionally produces a decision that a week of investigation would have reversed. That's an acceptable trade. An unresolved dispute is corrosive in a way a slightly-wrong resolved dispute is not — the rep in limbo can't work the account, can't forecast, and spends the interval building a case rather than building pipeline. Decide fast, publish the reasoning, and treat a reversal as a normal outcome rather than an embarrassment.

Overlays and specialists. Vertical specialists, product overlays, and named-account teams all multiply the surface where two people can legitimately claim the same deal. If you run overlays, the RoE has to state precedence explicitly and the routing layer has to encode it — manually-routed overlay models collapse into dispute floods within a quarter or two. The precedence order most orgs land on: named-account ownership is the orchestrator, vertical specialist enters when the deal triggers vertical-specific compliance or configuration requirements, product specialist enters when the deal includes something the core rep can't sell competently, and credit follows a documented formula rather than a negotiation.

Where these programs actually break

Six failure modes account for most of the wreckage. Each has a recognizable signature and a specific recovery.

Override theater. The CRO publicly rules for the top rep in a visible dispute. The RoE becomes decorative within one quarter, and the recovery cost is steep because every prior decision retroactively looks suspect. Recovery: the CRO recuses from disputes involving the top handful of producers for two quarters, delegates with the recusal itself published, and lets the floor watch the rules apply to someone with leverage.

Claim spam. Reps discover that a single logged touch establishes ownership, and the queue fills with bad-faith claims on accounts nobody intends to work. Recovery: add a substantive activity floor — several logged touches with actual content over a defined window — and auto-expire stale claims. Both changes go in the RoE with worked examples, because a rule that requires interpretation gets interpreted generously by the person it benefits.

The obviously-fair exception. Somebody breaks the freeze mid-quarter for a reason that is genuinely defensible. It doesn't matter that it was defensible. The floor now knows the freeze holds only when convenient, and every rep begins hedging. Recovery: a longer freeze with an explicit multi-signature exception process, and every exception published with its reasoning. The point is not to make exceptions impossible; it's to make them expensive and visible.

Hidden carve-outs. Strategic accounts assigned through back channels, discovered later by a rep who was told the list was complete. This is the most damaging failure on the list because it converts a process problem into a trust problem, and trust recovers on a multi-quarter timescale. Recovery: publish the complete carve-out list with a rationale per account, commit in writing to no undocumented carve-outs, and submit to a quarterly audit by someone other than the CRO's own team.

The 47-variable model. A scoring formula so elaborate that no rep can verify it. Functionally identical to no model, except it also cost you a quarter of RevOps time and produced the false confidence that the problem was solved. Recovery: cut to five to seven inputs, republish with worked examples on real accounts, and adopt the rule that adding a variable requires removing one.

Comp working against the map. The territory model is clean, the RoE is signed, the routing fires correctly — and reps still fight, because the plan pays enough on contested accounts to justify the effort. This is the failure that survives all the others being fixed, and it's the one leaders diagnose last. Recovery: audit variable comp by territory tier, flatten cross-boundary payouts to base rate with no accelerator, and eliminate per-logo bonuses entirely.

There's a seventh worth naming because it's cultural rather than mechanical: treating the whole program as a RevOps project rather than a leadership commitment. A territory operating system built by RevOps and tolerated by sales leadership will be abandoned the first time it produces an inconvenient answer. The CRO has to own the SLA personally, has to be the one who publishes the hard decision, and has to be visibly bound by rules they wrote. Reps are extremely good at detecting the difference between a leader who follows a system and a leader who has a system available for use when convenient.

Running the reset without killing morale in the process

The transition is more dangerous than the steady state. Announcing a new territory model to a floor that has been burned before reads as "your book is about to change," which triggers exactly the behavior you're trying to eliminate. Sequence matters enormously.

Diagnose before you touch anything, roughly the first two weeks. Count active and recent disputes by rep, by account size, by type. Pull attainment by rep and look for the pattern where low attainment correlates with territory disruption rather than skill. Run an anonymous single-question pulse survey — "do you believe the territory model is mechanically fair?" — administered by someone outside the sales chain of command. Map the routing logic that actually exists today, which is frequently not the logic anyone believes exists. Read the comp plan specifically looking for clauses that reward fighting. Change nothing during this phase and say so explicitly, because the diagnostic phase is when rumors do the most damage.

Draft with a rep in the room, weeks three and four. Build the scoring formula and the RoE with at least one respected AE participating, ideally someone mid-pack rather than a top producer — top producers optimize for rules that favor top producers, and the floor knows it. Draft the SLA. Get explicit CRO commitment to personally own the 48-hour decision, in writing, because verbal commitment evaporates the first time it's inconvenient.

Build and encode, weeks five through seven. Encode the formula in the CRM or routing layer. Move the capacity math somewhere auditable. Rewrite the comp plan. Stand up the public ledger — a shared document is fine; the surface matters far less than the fact that it's readable by everyone without asking permission.

Socialize before you launch, weeks eight and nine. Open a genuine comment period on the RoE where every AE reads it and can object. Take some of the objections. Visibly changing the draft in response to rep feedback is worth more for adoption than any amount of leadership roadshow, because it demonstrates the document is a real constitution rather than a decree.

Launch at a quarter boundary, and only at a quarter boundary. Publish the model, the RoE, the freeze, the SLA, and the comp changes simultaneously. Staggering them means reps experience each piece as a separate threat. Then resolve the first dispute under the new system publicly, within the SLA, with full written reasoning — and hope the first one goes against a strong rep, because that's the case that proves the system to everyone watching.

Reinforce for a quarter. Weekly SLA compliance check. Monthly RoE review that actually incorporates the gaps escalated disputes revealed. Re-run the pulse survey at quarter end and publish the score to the floor, including if it's bad. A leadership team that publishes a mediocre fairness score and states what they're changing buys more credibility than one that publishes a good score.

Expect dispute volume to spike in the first quarter, not drop. New rules surface edge cases that the old ambiguity hid. That spike is the system working — every escalation is a rule gap being found and closed. If the volume is still elevated in quarter three, the RoE has a structural problem rather than a maturity problem, and it needs a rewrite rather than more patience.

One last thing on the morale question specifically, since that's what the question is actually about. The instinct when a rep loses a disputed account is to soften the blow — a sympathetic 1:1, a vague promise about making it up to them. That instinct is wrong in an important way. What actually restores morale is a concrete, published offset: priority on the next two or three high-potential accounts that route through the queue, documented in the same ledger as the decision that went against them. Sympathy without a mechanism reads as management theater. A mechanism without sympathy still works. The rep doesn't need you to feel bad; they need to know the system will make them whole in a way they can verify.

That is, in the end, the whole thesis restated. RevOps can eliminate almost every dispute through encoded rules, and the residual few can be decided fast and in public. What you cannot do is eliminate the losing — someone loses every contested account. Handling disputes without killing morale means making the losing legible: this rule, applied to you the same way it's applied to the rep at 130%, with the reasoning written down and the offset already scheduled. Reps can absorb losing. They cannot absorb not knowing why.

Related questions

How often should territories be rebalanced?

Once or twice a year, aligned to the fiscal calendar, using the published scoring formula rather than leadership intuition. Announce at least two weeks ahead. More frequent rebalancing destroys the account-knowledge compounding that makes stable territories outperform in the first place.

Who should own the final dispute decision — CRO or VP Sales?

The CRO, personally, except in genuine recusal cases. Delegation reads as avoidance to the floor and is the most common way a resolution SLA quietly collapses. If the CRO has a close relationship with one disputant, recuse explicitly and publish the recusal alongside the decision.

Should Customer Success be included in the Rules of Engagement?

Yes. Expansion-versus-net-new ownership between AEs and CS is one of the highest-frequency dispute categories and is usually left undefined. Write it into the same document, with explicit language on who owns upsell, who owns cross-sell into new business units, and how renewal-adjacent expansion credits.

What's a reasonable deal-registration window?

Thirty to ninety days, scaled to sales-cycle length — shorter for transactional motions, longer for enterprise. Pair it with a substantive activity floor so registration requires real work, and auto-expire stale registrations so accounts don't sit locked under a rep who stopped working them.

How do you handle disputes on partner-sourced or channel deals?

Same document, separate clause. Define whether partner registration beats direct-rep territory claim, how long partner registration holds, and the split formula when both worked the deal. Channel-versus-direct is structurally identical to AE-versus-AE and fails for identical reasons when left unwritten.

FAQ

What if two AEs both claim they touched an account first?

Resolve it on logged CRM evidence against a defined activity floor, not on recollection. A qualifying claim should require several substantive touches with actual content inside a defined window — a single connection request doesn't count. If both clear the floor legitimately, apply the published split formula rather than picking a winner, and log the reasoning either way.

How do you stop reps from hoarding accounts they never work?

Enforce capacity ceilings in the model — a cap on both current-state ARR and active account count per rep — and add an activity-based release rule that returns accounts with no meaningful engagement over a defined period to the routing pool. Publish the release rule in advance so nobody is surprised, and give the rep a warning window before release fires.

Does a public decision ledger create more conflict by making losses visible?

The opposite, consistently. Private decisions get reconstructed through rumor, and the rumored version is always less flattering than the real one. Publishing the rule that fired and the reasoning behind it converts a personal defeat into an impersonal one, and lets every other rep calibrate their own behavior against a real example rather than a guess.

What do you do when a rep is genuinely right that their territory is unfair?

Fix it at the next quarter boundary through the model, not immediately through an exception. In the interim, offer a bounded, published offset — priority routing on the next high-potential accounts, or a time-limited incentive on hard-to-penetrate segments. The offset addresses the individual without teaching the floor that complaining loudly works.

How do you know whether the model is working?

Watch three numbers: dispute rate as a share of opportunities created per quarter, SLA compliance percentage, and an anonymous quarterly fairness score. Dispute rate and SLA compliance move within a quarter; attrition lags by two to four. If the fairness score is poor while dispute rate looks healthy, believe the fairness score — perception is what drives the exit decision.

Is this worth building for a team of eight reps?

A lightweight version, yes. You don't need a routing platform or a planning tool at that size, but you do need a written RoE, a named decision-owner, and a visible record of decisions. The document takes a day to write and prevents the specific failure where an eight-rep team's informal norms stop scaling somewhere around fifteen and nobody notices until two people quit.

Sources

1. Gartner — Sales practice research and sales force effectiveness
2. Forrester — B2B sales and revenue operations research
3. The Bridge Group — SaaS AE and inside sales metrics reports
4. OpenView Partners — SaaS Benchmarks
5. Bessemer Venture Partners — State of the Cloud / Atlas
6. SaaStr — sales compensation and go-to-market guidance
7. Salesforce — Sales Cloud Territory Management
8. LeanData — lead routing and territory orchestration
9. Anaplan — connected planning for sales
10. Xactly — sales performance and incentive compensation management

Related on PULSE

• How should RevOps design a sales compensation plan that doesn't create perverse incentives?
• What's the right account-to-rep ratio before coverage quality starts degrading?
• How do you split credit on multi-touch deals without starting a war?
• When should a sales org move from geographic territories to named accounts?
• What causes regretted AE attrition and which signals show up first?