How should I structure SDR commission to discourage gaming MQL counts in 2027?

The two structures you are actually choosing between

Every SDR commission conversation eventually collapses into a binary: pay on an ungated unit the rep can produce unilaterally, or pay on a gated unit that a second party with opposing incentives must approve. Everything else — scoring thresholds, activity minimums, meeting-booked bonuses, "quality coaching" — is decoration on top of that one choice. Understanding both options honestly, including what the ungated version genuinely gets right, is the only way to make the switch stick.

Option A: the MQL-volume plan. The rep earns a per-MQL bounty, or a per-meeting-booked bounty against an MQL quota. Marketing scores the lead, the SDR works it, and the count is the payable unit. This design is fast to implement, needs almost no CRM engineering, produces a daily feedback loop the rep can feel, and gives leadership a clean scoreboard that updates in real time. Those are real virtues, and they explain why the design keeps getting rebuilt at company after company despite a twenty-year track record of failure. It is the path of least operational resistance.

Its fatal property is that no party in the transaction has an incentive to say no. Marketing benefits when MQL counts rise. The SDR benefits when MQL counts rise. The cost of a bad MQL lands on a third party — the AE — who has no vote at the moment of creation. When a measure becomes a target it ceases to be a good measure, and here the measure is not merely targeted but directly monetized, with no counterweight anywhere in the loop.

Option B: the SAO-gated plan. The rep earns a bounty per Sales-Accepted Opportunity, where an AE must affirmatively accept against a written checklist, and the bounty reverses if the AE disqualifies the opportunity inside a defined window for a quality reason attributable to the SDR's qualification work. A separate quarterly kicker pays on downstream conversion. This costs more to build: you need a two-signature event in the CRM, reason codes, an SLA on the AE side, an audit cadence, and an adjudication path for disputes. In exchange you get a unit that cannot be inflated unilaterally.

The comparison in practice:

There is a third position worth naming so you can rule it in or out deliberately: paying SDRs primarily on closed-won revenue. It sounds like the ultimate gate — the customer's signature is the least gameable approval in the entire funnel — and for AEs it is exactly right. For SDRs it is a known anti-pattern. It makes an SDR's income hostage to AE skill and a 60-120 day sales cycle, destroys the fast motivational loop the role depends on, and is brutally unfair to a strong prospector paired with a weak AE. A small closed-won overlay is fine as a revenue-connection gesture. As the primary unit it produces a different pathology: reps stop prospecting hard because the payoff is too distant and too far outside their control.

How to decide, and how to tell which failure you already have

Before you redesign anything, diagnose which of three distinct behaviors people mean when they say reps are gaming the numbers. Each has a different countermeasure, and a plan that fixes one while ignoring the others just relocates the problem.

Volume gaming — sandbagging the threshold. The rep pushes marginal contacts over whatever line defines the payable unit: logging touches, firing one templated sequence to trip an engagement score, marking a half-hearted call as qualified. Nothing here requires fraud. It requires only optimism, and optimism is free. The countermeasure is the second signature.

Definition gaming — exploiting ambiguity. Where criteria read "shows interest" or "good fit," reps interpret generously and consistently in their own favor. Every ambiguous field is a free option the rep will exercise. The countermeasure is an objective, checklist-based definition with as few free-text vibes fields as you can tolerate.

Timing gaming — stuffing the window. Reps batch weak hand-offs at quarter end to reach accelerators, or hold strong leads back to smooth a future quarter. The countermeasures are the clawback window, which makes quarter-end stuffing dangerous because the reversal lands in the next period, and rolling quota credit that reduces the payoff from cliff behavior.

Notice what is absent from all three countermeasures: better lead scoring. The most common instinct when SDRs game MQLs is to fix the MQL — add behavioral signals, raise the threshold, retrain the model. This does not work, and the reason matters. Tightening the score moves the line the rep games against; it does not change the fact that the rep is paid on an ungated unit. A more elaborate scoring model is a more elaborate proxy, and the same law applies to elaborate proxies exactly as it applies to simple ones. Worse, sophistication manufactures false confidence: leadership believes the smart model solved quality while reps quietly learn which signals to trip. Keep lead scoring for routing and triage, where it genuinely helps an SDR decide what to work first. Never let it touch a commission line.

The decision tree above is also a diagnostic order. Run it top down. Most orgs stop at the first branch and never get to the subtler failures, which is why so many second-generation plans — gated but with vague criteria, or gated but with no consequence for AE stonewalling — decay back into dysfunction within two quarters.

One more decision input that gets skipped: is this a discipline problem or a design problem? Leaders instinctively read gaming as a character issue and reach for better hiring, integrity coaching, tighter management. That framing is comfortable because it locates the fault in the rep rather than the plan. Test it directly. Take your single most honest, most diligent SDR and put them on a pure pay-per-volume plan. Within two quarters, watching less scrupulous peers out-earn them on padded numbers, even that rep starts rounding up — not because they became a worse person, but because the plan made honesty financially irrational. If your best rep games under a given plan, the plan is the variable to change. You cannot hire or coach your way out of a structurally broken incentive.

The concrete numbers behind each structure

Percentages below are a defensible market-standard default, not physics. Tune them to your motion, deal size, and ramp.

Pay mix. For a quota-carrying SDR, a defensible split sits at roughly 60-65% base, 35-40% variable. SDR is a higher-base-weighted role than AE on purpose: so much of an SDR's output depends on lead flow, territory quality, and product-market fit — inputs the rep does not control. A 50/50 split, standard for AEs, is too aggressive here and manufactures exactly the desperation gaming you are trying to kill. Variable then splits roughly 70% per-SAO bounty / 30% quarterly quality kicker.

Worked example — mid-market SDR, US metro:

The arithmetic that makes gaming lose. Walk it from the rep's chair, because the rep certainly will. A junk SAO earns ~$113 today. If the AE disqualifies it inside 12 business days — highly likely for genuinely junk work — the rep loses the $113. Net comp impact: zero, plus spent calendar time and spent AE goodwill they will need on the next hand-off. Worse, that junk SAO sits in the denominator of the conversion kicker. A rep who books 16 real SAOs at 50% conversion clears the 45% gate and banks $2,325. The same rep padding to 22 SAOs at 33% conversion misses the kicker entirely, forfeiting $2,325 to chase ~$113 bounties that mostly reverse. Expected value of gaming is negative. That is the whole design. You are not relying on honesty, monitoring, or culture — you have made the dishonest path the lower-paying path, so honesty becomes the dominant strategy financially.

Use this as a design test: if you cannot draw on a napkin the arithmetic showing the gaming strategy earns the rep *less* than the honest strategy, the plan is not finished. Keep tightening the clawback window and the conversion gate until the napkin math works.

Quota, from capacity rather than ambition. A clean structure sitting on an unreachable number still produces gaming, because a rep who literally cannot hit an honest quota will manufacture a dishonest one. Build bottom up: ~21 working days a month × productive prospecting hours per day, realistic touches per hour, then the cascade of connect rate → conversation rate → meeting booked → meeting held → AE accepted. For a typical B2B SaaS motion this lands at 12-22 SAOs per SDR per month. Outbound-heavy motions with larger deals sit at the low end (10-15); inbound-assisted motions with real marketing air cover sit higher (18-25). Set quota near the median achievable so roughly half the team can exceed it. A quota only the top decile can reach is itself a gaming incentive.

Attainment bands:

Generous accelerators are safe *because* the unit is gated and clawback-protected. Overperformance now means more real opportunities, which is precisely what you want to pay a premium for. Bolt those same accelerators onto an ungated MQL count and you have simply turbocharged the inflation.

Ramp. New SDRs sit on a ramped quota — commonly 25% / 50% / 75% / 100% across four months — with partial guaranteed commission during the ramp. Holding a new rep to full quota on day one is the most predictable single source of gaming in the entire system: they are not yet capable of the honest number, and a rep who cannot earn a living wage honestly will find a dishonest way to earn one. That is rent and groceries, not bad character. The ramp guarantee is not generosity; it is a gaming control.

Territory equity, the quota input nobody audits. Two SDRs on identical quotas but unequal territories are not on the same plan regardless of what the spreadsheet says. Hand one rep a rich vertical with strong air cover and the other a thin patch, and the second faces a higher *effective* quota and a correspondingly stronger gaming incentive. Size each territory's realistic SAO capacity before launch and either equalize the patches or adjust the numbers. Reps detect territory inequity faster than almost anything else, and the rep who believes the game is rigged against them is the rep most likely to rig it back.

Multi-year drift. Quotas creep upward because last year's top performer silently becomes this year's baseline. When quota inflation outpaces genuine capacity growth from better tooling and enablement, you are slowly rebuilding the desperation problem you just solved. Re-run the bottom-up capacity model annually; never apply a percentage bump to last year's number and call it planning.

Building the gate, the clawback, and the kicker

The whole structure rests on a crisp definition of the accepted unit. If the definition is mushy, reps game the definition instead of the count and AEs reject good work to dodge effort. Spend real time here.

Write acceptance criteria as an objective checklist. Trimmed from the BANT/MEDDICC families to what an SDR can realistically establish:

• ICP fit — company size, industry, geography match a documented segment. Near-binary.
• Authority or a named path to it — the contact either owns the budget or has explicitly named who does and agreed to an introduction.
• Identified, articulated pain — a specific stated problem in the prospect's own words, logged verbatim. Not "might be interested."
• Plausible timeline — some indicated evaluation window, even loose: this half, before the Q3 renewal.
• Meeting held, not booked — a discovery meeting actually occurred. This single criterion kills no-show stuffing outright.

Require structured CRM fields for the verbatim pain quote, the named budget owner, and the stated timeline, and require the AE to confirm them at acceptance. Fewer free-text fields, less room to game the definition.

Two signatures, timestamped, immutable. The SAO is created only when the SDR submits and the AE accepts, both logged with timestamps. Make it a discrete event record, not an inference from a stage picklist, so it cannot be silently backdated or flipped. Minimum fields: SAO ID, source SDR, accepting AE, submitted-at and accepted-at timestamps, a criteria snapshot captured at acceptance so later edits cannot rewrite history, clawback status with reason code, a pointer to the downstream opportunity so per-rep conversion is computable, and a rejection record for declined submissions. Build it as an append-only log. If a junior admin can overwrite an acceptance timestamp, the entire clawback mechanism is compromised — treat this event with the integrity discipline you would give a financial transaction, because through the bounty it is one.

Close the loophole in the other direction. A gate only the AE controls creates a fresh gaming surface: the AE can reject good work to keep their calendar light or to make SDR-sourced pipeline look weak in a turf war. Three controls. First, an acceptance SLA — accept or reject within 2 business days, with silence defaulting to accepted, so no AE can kill an SAO through inaction. Second, structured reason codes on every rejection (out of ICP, no pain identified, duplicate, no authority and no path); free-text-only rejections are not allowed. Third, a frivolous-rejection audit where RevOps samples rejected opportunities monthly, and an AE who rejects work that another rep then accepts and converts gets scrutinized. AEs should also carry a pipeline-coverage expectation so starving themselves of SAOs damages their own forecast. This is symmetric accountability: every gate that can punish one role needs a control punishing the counterparty for abusing it. Asymmetric accountability — SDR clawed back, AE untouched — is the second most common reason these plans fail.

Clawback mechanics. Ten to fifteen business days is the working range: long enough for a genuine first discovery call, short enough that pay is not hostage for months. Tie the clock to a triggering event — twelve business days after acceptance, or after the first discovery call, whichever comes first — so a slow AE calendar does not punish the rep. Then be ruthless about scope:

The governing principle: the SDR is accountable for qualification, not for the outcome. A well-qualified opportunity that still lost is a good SAO and the rep keeps the bounty. Punish reps for losses they could not influence and you push them straight back into volume gaming, because if you get clawed back either way you may as well book more. Pair this with a defined dispute path (SDR manager and AE manager jointly, or RevOps, ruling on logged criteria), a statute of limitations so a cleared bounty is permanently earned with no retroactive reversals, and a small monthly pool of manager-discretion adjustments as a pressure-release valve for edge cases the rules did not anticipate. Healthy clawback rates run in the low single digits. Above roughly 10% means the definition is too loose or AEs are over-rejecting — investigate which. A near-zero rate is its own warning: the gate may be rubber-stamping.

The kicker. Tie ~30% of variable to a quarterly quality metric, in rough order of preference: SAO-to-Stage-2 conversion rate (what share of accepted opportunities advanced past discovery), SAO-sourced pipeline dollars (better when deal sizes vary widely), or SAO-to-closed-won as a smaller lagging overlay. Measure quarterly, never monthly — a single SDR generates too few SAOs a month for a stable conversion rate, and one or two coin-flip deals swing it wildly. A rolling quarterly window gives enough sample to reflect skill rather than luck, and it spans the period boundary, which further dampens quarter-end behavior. Set the launch threshold near the historical median so roughly half the team clears it on day one.

With both components live the rep faces a clean, legible optimization: book as many *real* opportunities as possible because volume drives the bounty, and never book a fake one because fakes reverse *and* drag down the conversion rate that pays the kicker. The plan is finally telling the truth about what the business values.

Sequencing the migration, and the governance that keeps it honest

Moving an existing org off volume-based pay is a change-management project, not a spreadsheet edit.

Phase 1 — Instrument (weeks 1-3). Before changing a dollar of pay, make sure you can measure the new unit. Build the two-signature event, the reason codes, the clawback clock, and per-rep conversion reporting. Then run the new model in shadow mode alongside the live plan for at least a full month so quotas and bounty rates get calibrated against real data rather than guesses.

Phase 2 — Calibrate (weeks 3-6). Use shadow data to set quota at realistic capacity, back the per-SAO bounty out of target variable, and set the kicker threshold at the historical median. Pressure-test with the napkin math: confirm the gaming strategy nets less than the honest one.

Phase 3 — Communicate (weeks 5-7). Over-communicate, and lead with the *why* — the proxy problem, the fairness argument, the AE partnership — before the mechanics. Walk every rep through a personalized worked example of their likely earnings. State explicitly that quotas and rates were set so a good rep earns at least as much as before. If reps read the change as a stealth pay cut, they resist and game harder. Many orgs run a one-quarter hold-harmless floor guaranteeing no rep earns below their trailing average during the transition.

Phase 4 — Launch and monitor (quarter 1). Go live at a quarter boundary, never mid-period. Stand up the audit cadence immediately and watch disputes closely for six weeks — early disputes almost always reveal genuine ambiguity in the criteria, so treat them as free QA and tighten the wording.

Phase 5 — Tune (quarter 2+). Review attainability, conversion stability, and clawback rates, then adjust at the next quarter boundary through governed change control.

Governance keeps it from decaying. The criteria, the clawback rules, and the dispute process belong to a neutral function — RevOps — not to Marketing and not to Sales. Marketing left alone loosens the definition to make lead numbers look good; Sales left alone tightens it to dodge work. Version the criteria, announce changes, and make them effective only at quarter boundaries. Redefining a paid unit mid-quarter is a comp-fairness violation and should be treated as one.

Run a standing monthly audit whose explicit job is detecting proxy inflation. Watch four things: conversion-trend divergence (SAO volume rising while SAO-to-Stage-2 conversion falls is the canary — treat a sustained divergence as a fire, not a curiosity); rejection patterns by rep and by AE, flagging outliers in both directions; per-rep cohort curves where consistently high volume paired with bottom-decile conversion is a coaching trigger; and definition drift, confirming nobody quietly loosened the criteria. Report cadence: per-rep volume versus quota weekly, acceptance and rejection rates weekly, conversion by rep monthly, clawback rate and reasons monthly, full plan-health review quarterly. The single most important chart is volume plotted against conversion over time. When the lines cross, the unit is being inflated and you intervene that week.

Plan stability is itself an anti-gaming control. If management re-cuts the plan every time someone earns "too much," reps learn the real game is predicting management rather than serving customers, and they hedge, sandbag, and game accordingly. Publish the plan, commit to it for the fiscal year barring genuine emergencies, and change it only through the governed process.

Hiring and coaching carry the last stretch. Screen prospectors for pride in clean hand-offs — ask a candidate how they decided a lead was *not* ready, and listen for real disqualification judgment. A candidate who cannot describe walking away from a marginal opportunity will pad your numbers under pressure. Make hand-off quality a standing 1:1 topic, have AEs give structured feedback on sourced SAOs, and celebrate the best conversion rate as loudly as the highest volume. The fastest cultural signal available to leadership is publicly thanking an SDR for a good disqualification — the lead they correctly chose not to push through.

Adapting the pattern to adjacent roles and motions

The underlying rule — pay on the first unit a counterparty with opposing incentives must approve — generalizes well beyond sales development, though the specifics shift.

Inbound reps (BDR/MDR) work marketing-generated demand, so connect and acceptance rates run higher and quotas sit at the top of the range. The gaming risk is *sharper*, not milder, because a fat pile of scored leads sits right there waiting to be pushed through. The gate and the clawback matter more for inbound, not less. Speed-to-lead is fine as a small modifier; never let it substitute for the quality gate.

Outbound reps source their own opportunities, so volume is lower, each SAO is worth more, and quotas run 10-15. They control the entire top of funnel, which means more room to game — but hired well, they also carry more pride in pipeline quality. The clawback and conversion kicker do the heavy lifting.

Account executives face the analogous risk in a different shape: sandbagging and pipeline inflation, stuffing the CRM with low-probability deals or pulling business into a quarter to reach an accelerator. The analogous fix attaches the meaningful money to closed-won revenue — the customer's signature being the ultimate opposing party — and governs forecast accuracy as a tracked competency rather than a vibe.

Customer success and renewals hit it again with a twist: pay on gross retention or expansion that survives a post-renewal window, not on a "saved account" flag the CSM sets unilaterally. Any self-reported save is an ungated unit.

Channel and partner incentives are the same problem outsourced. A partner-sourced registration nobody validates is an MQL by another name; require a joint acceptance step before the referral fee attaches.

Marketing itself deserves the mirror treatment. If SDRs stop being paid on MQL volume but the demand-gen team is still bonused on it, you have moved the inflation upstream rather than removed it. Move marketing's number to SAO-sourced pipeline or accepted-opportunity contribution and both functions finally point at the same unit — which, incidentally, ends the perennial lead-quality argument that no dashboard has ever settled, because there is now one agreed unit under neutral governance.

The narrow honest exception. There is one scenario where paying on a leading metric is temporarily defensible: a brand-new motion with no AE bench, no historical conversion data, and a first SDR cohort that needs a functioning plan now. In that bootstrapping window a meetings-held bounty — never raw MQL count; at minimum require a held meeting — can work for one or two quarters, provided you pay only on held meetings, instrument the full downstream funnel from day one even though you are not paying on it, publish a hard migration date so reps know the plan moves, and keep variable modest so the bounty cannot fund a serious gaming spree before you migrate. The exception is scaffolding you remove on schedule. Two years into a "temporary" meetings plan and you have quietly rebuilt the problem you set out to solve. A lesser caveat: on teams of one or two SDRs, formal clawback and dispute machinery is heavier than the problem warrants — run the principle on lightweight manager judgment and add the apparatus past roughly five SDRs, when consistency starts to matter more than speed.

What it costs, honestly. The gated model is not free: CRM engineering for the two-signature event, RevOps time for the monthly audit, manager time for adjudication, and a genuine change-management effort. But for any team past a handful of SDRs, that cost is trivially smaller than what it eliminates — inflated CAC, AE hours burned on garbage, a poisoned marketing-sales relationship, attrition among your good reps, and a pipeline number leadership cannot trust. The volume plan only *feels* cheaper because its costs are diffuse and hidden. The gated plan concentrates a modest visible cost and removes a large invisible one.

Related questions

What if SDRs complain their pay now depends on the AE?

Partly the point, and it is fair because the dependence runs both ways. The AE must respond inside an SLA, must supply a reason code, and is audited for frivolous rejection. The SDR depends on the AE; the AE is held to a standard. Communicate it as partnership.

What if marketing's lead volume genuinely drops?

The gated plan exposes that problem instead of hiding it — under volume-based pay, thin lead flow was disguised by inflation. If reps cannot source enough accepted opportunities, you have a real demand-generation problem to fix at the source. The comp plan is now telling you the truth.

Should SDR earnings be capped?

No. With a gated, clawback-protected unit, a rep earning well above OTE is generating a lot of real pipeline — exactly what you want more of. Caps signal distrust and push your strongest reps to sandbag or leave. Use accelerators, not ceilings.

Monthly or quarterly bounty payout?

Pay the per-SAO bounty monthly for a fast feedback loop, but only on opportunities that have already cleared the clawback window — so a late-month SAO pays in the following cycle. Pay the conversion kicker quarterly for sample-size stability.

Can tighter MQL scoring fix this without changing comp?

No. Raising the threshold only moves the line reps game against; the unit is still ungated. Sophisticated scoring adds false confidence while reps learn which signals to trip. Keep scoring for routing and prioritization, never for payment.

FAQ

Why do SDRs game MQL counts in the first place?

Because the plan pays for it. When commission attaches to a unit the rep can produce unilaterally, maximizing that unit is doing the job correctly as written. Nobody upstream of the AE has a reason to say no — marketing benefits from higher counts, the SDR benefits from higher counts, and the cost lands on a third party who had no vote. The defect is in the definition of the payable unit, not in the rep's character.

How exactly does a clawback window discourage padding?

It extends the rep's accountability past the moment of payment. A junk opportunity earns a small bounty today and reverses ten to fifteen business days later when the AE disqualifies it, netting zero while consuming calendar time and AE goodwill. It simultaneously dilutes the conversion rate that gates the quarterly kicker, so the padder loses twice. Gaming stops being a strategy because it stops being profitable.

What base-to-variable split should an SDR plan use?

Roughly 60-65% base and 35-40% variable, with the variable split about 70% per-accepted-opportunity bounty and 30% quarterly quality kicker. SDRs carry a higher base weighting than AEs because so much of their output depends on lead flow, territory, and product-market fit — factors they do not control. Pushing toward a 50/50 AE-style mix manufactures the desperation that drives gaming.

What clawback rate indicates a healthy system?

Low single digits as a share of accepted opportunities. Persistently above roughly 10% means either the acceptance criteria are too loose so reps submit marginal work in good faith, or AEs are over-rejecting — investigate which before adjusting rates. A near-zero rate deserves scrutiny too, since it often means the gate is being rubber-stamped rather than genuinely applied.

Who should own the definition of the accepted unit?

A neutral function, in practice RevOps. Marketing left to itself loosens the definition to flatter lead numbers; Sales left to itself tightens it to reduce workload. Neutral ownership with versioned, quarter-boundary change control keeps the unit stable, and stability is what lets reps optimize within the plan instead of gaming against it.

Is there ever a defensible reason to pay on a leading metric?

Yes, narrowly. A brand-new motion with no AE bench and no conversion history can run a meetings-held bounty — never raw MQL count — for one or two quarters, provided the downstream funnel is instrumented from day one, variable comp stays modest, and a hard migration date is published up front. It is scaffolding with a removal date, not a permanent structure.

Sources

• https://hbr.org/2015/04/motivating-salespeople-what-really-works
• https://www.saastr.com/the-sdr-comp-plan-that-works/
• https://openviewpartners.com/blog/sales-development-metrics/
• https://www.gartner.com/en/sales/topics/sales-compensation
• https://blog.hubspot.com/sales/sales-compensation-plans
• https://www.salesforce.com/resources/articles/sales-compensation-plans/
• https://www.mckinsey.com/capabilities/growth-marketing-and-sales/our-insights/sales-compensation-that-drives-growth
• https://hbr.org/2018/01/how-to-set-sales-quotas-that-motivate-your-team
• https://www.bridgegroupinc.com/sdr-metrics
• https://en.wikipedia.org/wiki/Goodhart%27s_law

Related on PULSE

• What's a fair OTE for an enterprise AE selling $100k+ ACV deals?
• What accelerator multiples are typical past 100% of quota for SaaS AEs?
• What's the right SDR-to-AE ratio for a Series C SaaS company?
• How do you set sales quotas from capacity math instead of top-down targets?
• How do you separate NRR, GRR, and logo retention when auditors ask which is real?
• How should marketing be compensated when SDR pay moves off MQL volume?