How to design SDR compensation that retains top performers in 2027

The outcome you should expect

A compensation plan built this way does not eliminate SDR turnover — nothing does, because the role is by design a two-to-three-year launchpad into AE, customer success, or RevOps seats. What it does is change *who* leaves and *when*. The failure state most teams live in today is that the top quartile leaves first, at month 14 to 20, right as they become genuinely productive. The success state is inverted: your bottom quartile self-selects out in the first two quarters because the variable component is real and unforgiving, while your top quartile stays through month 30 because the accelerator ceiling is high enough that leaving costs them money.

Concretely, here is what a well-designed plan produces on a 20-rep team over 12 months. Voluntary attrition in the top quartile drops from something like 40–50% annually to the 15–20% range. Overall team attrition may barely move — it might even rise slightly in the first two quarters as low performers exit faster — but the *composition* of that attrition shifts. Average productive tenure rises, which means the fixed cost of ramp is amortized over more months of output. If your ramp is roughly five months and median tenure is 26 months, you get 21 productive months per hire. Push median tenure to 32 months and you get 27 productive months — a 28% increase in output per hire with zero change in per-rep quota.

The second-order outcome is that your AE bench improves. SDR-to-AE promotion is the single highest-ROI hiring channel most SaaS orgs have, and it only works if reps survive long enough to be promotable. A rep who leaves at month 15 was never a promotion candidate; a rep who reaches month 28 with two years of accepted-opportunity history is a known quantity with a proven work ethic and full product fluency. Companies that keep SDRs longer end up spending materially less on external AE recruiting, which is a cost line that rarely gets attributed back to SDR comp design but should be.

The third outcome is forecast quality. When variable pay is gated on AE-accepted opportunity rather than raw meetings, the pipeline that enters your forecast is cleaner by construction. Bad-fit meetings do not get booked, because they do not pay. This has a knock-on effect on AE morale — AEs stop treating SDR meetings as a tax on their calendar — and on marketing attribution, because sourced-pipeline numbers stop being inflated by meetings that never should have counted. RevOps leaders consistently report that fixing the SDR payable event is the cheapest pipeline-hygiene intervention available, because it changes rep behavior in a single pay period rather than requiring a data-governance project.

What you should not expect: this plan will not fix a broken ICP, a product with weak market pull, or a manager who does not coach. Compensation is a steering mechanism, not an engine. If your reps cannot book quality meetings because the list is bad and the message is generic, no accelerator structure will retain them — they will hit 60% of quota, earn 60% of variable, and leave for a company where the phone gets answered. Comp design assumes the underlying motion works. Diagnose that first.

What drives that outcome

Five levers do nearly all the work, and they interact rather than operate independently. Understanding the interaction is what separates a plan that retains from a plan that merely pays.

Lever one: pay mix set by motion, not job title. Enterprise outbound SDRs working a small named-account list — six to ten accounts, multi-threaded, long research cycles — should sit closer to 75/25. Their feedback loop is measured in months, so a heavy variable component reads as instability rather than opportunity, and the highest-variance outcome in their quarter is often outside their control. High-velocity SMB SDRs running large daily touch volumes tolerate 60/40 comfortably because their loop closes in days and their statistical sample size per month is large enough that effort reliably converts. Mid-market anchors at 65/35. Get this wrong in either direction and you generate attrition: too much variable in a long-cycle role produces anxiety-driven exits; too little variable in a fast-cycle role produces boredom-driven exits among top performers who want the upside.

Lever two: the payable event. This is the highest-leverage decision in the whole plan. Paying on meetings booked pays for calendar entries. Paying on meetings held pays for show rates but still tolerates bad-fit prospects. Paying on AE-accepted opportunity — where the AE affirmatively marks the opportunity as qualified into a defined stage within a fixed window, typically five business days — pays for pipeline. The mechanic needs three supporting pieces to be fair: a written acceptance definition that both SDRs and AEs sign annually, a dispute path that resolves inside one pay period, and an SLA on the AE that forces disposition within the window so reps are not left waiting on someone else's CRM hygiene. Without the SLA, the gate becomes a source of resentment rather than alignment.

Lever three: accelerators with no cap. Flat per-opportunity commission is the quiet retention killer, because it teaches your best reps to cruise. A rep who can hit quota by the 22nd of the month has no financial reason to work the last week. A tiered structure — standard rate to quota, a meaningful step-up in the first band above quota, a larger step-up in the second, and something aggressive at the top band — turns the last week of the month into the most profitable week of the rep's year. The design principle: a top-quartile rep should be able to earn 130–150% of OTE in a strong quarter. That number is what a competing recruiter has to beat, and it is much harder to beat than base salary.

Lever four: clawbacks sized to protect quality, not to punish. A 180-day window on disqualified or bad-fit opportunities reverses the commission at the next pay period. The purpose is behavioral, not financial recovery — the dollar amounts are small relative to payroll. What matters is that "book-and-burn" stops being profitable. This lever matters more in 2027 than it did five years ago, because AI-assisted prospecting has made it trivially cheap to generate volume. When the marginal cost of a bad meeting approaches zero, the comp plan is the only remaining brake.

Lever five: tenure bonuses at the cliff. The attrition spike sits between roughly month 18 and month 26 — precisely when a rep has ramped, learned the playbook, built a network, and become maximally poachable. A one-time cash bonus at month 13 and a larger one at month 24, both conditional on hitting a threshold like 80% of trailing-twelve quota, cost far less than a single replacement cycle. Present this to finance as turnover-cost offset, not headcount inflation, and it clears without friction.

The interaction effect is what most plans miss. Accelerators without a quality gate produce a flood of junk pipeline from reps chasing the top band. A quality gate without accelerators produces reps who hit exactly 100% and stop. Tenure bonuses without either produce well-paid mediocrity. The five levers are a system; implementing two of them and declaring victory usually makes things worse than the flat plan you replaced.

Benchmarks and realistic ranges

Compensation benchmarks move fast and vary enormously by geography, funding stage, and sales motion, so treat any single number as a starting hypothesis to validate against live market data rather than a rule. What follows are the structural ranges that hold up across most mid-market and enterprise SaaS environments in 2026–2027.

OTE and mix. Mid-market SaaS SDR OTE commonly lands in the low-six-figure range in major US metros, with base representing roughly two-thirds. Enterprise SDR roles skew higher on base and lower on variable; SMB and high-velocity roles skew the other way. Regional spread between a top-tier tech metro and a lower-cost secondary market is frequently 20–30% on base — meaningful enough that a single national band will simultaneously overpay one location and lose reps in another. Remote-first companies increasingly run two or three geographic bands rather than a per-city matrix, which is administratively survivable and defensible in a board comp discussion.

Quota and payable-event pricing. Work backward, always. Pick the annual variable number, divide by twelve for a monthly variable target, then divide by the monthly accepted-opportunity quota to get per-opportunity value. If the resulting per-opportunity number feels absurdly high or low, the quota is wrong, not the mix. A monthly accepted-opportunity quota in the low double digits is typical for mid-market; enterprise runs materially lower with a higher per-opportunity value; SMB runs higher with a lower value. The sanity check that matters: at 100% attainment, a rep should be able to reach quota working a normal week. If quota requires heroics to hit at plan, you have designed a plan that pays out at 70% and reads to reps as a pay cut.

Attainment distribution. A healthy plan produces roughly 60–70% of reps at or above quota, with a long right tail. If 90% of your team is over quota, the quota is too low and you are overpaying for baseline effort. If 30% are over, the quota is too high, your median rep is earning well below OTE, and your attrition is about to spike regardless of what the plan document says. Model the distribution before you launch, using last year's actual per-rep opportunity counts run through the new payable event definition. That backtest is the single most valuable thirty minutes in the whole design process, and it is the step teams skip most often.

Ramp. Budget roughly four to six months to full productivity, and pay a ramp guarantee — typically a declining percentage of target variable over the first three to four months — so new hires are not financially punished for a learning curve you designed. Ramp guarantees are cheap and dramatically reduce first-year attrition, which is where a large share of total SDR turnover concentrates.

Cost of turnover. The full-loaded replacement cost of an SDR — recruiting, onboarding, management time, lost pipeline during vacancy and ramp — typically runs a substantial multiple of the tenure bonus you are considering. This is the number that unlocks CFO approval. Build the model explicitly: vacancy weeks times weekly sourced-pipeline value, plus ramp months at partial productivity, plus recruiting cost. Compare it to the total tenure-bonus outlay across the team. The comparison is usually not close, and it converts the conversation from "more spend" to "cheaper spend."

Adjacent benchmarks worth pulling. If you are designing SDR comp in isolation you will create internal-equity problems. Pull the AE plan, the customer success plan, and the partner or channel plan at the same time. A common and corrosive failure: a top SDR discovers that a first-year AE with lower measurable output earns 60% more, and concludes the promotion path is the only route to fair pay. That is fine if your promotion pipeline is fast and transparent; it is poison if promotion is slow and opaque. The SDR plan should be legible as the first rung of a ladder with visible rungs above it, including the timing and criteria for the next rung.

Risks, edge cases, and failure modes

Mid-year plan changes. The fastest way to destroy trust is to change the plan in Q3 because Q2 attainment ran hot. Reps remember every change, and they interpret it — correctly — as the company reserving the right to move the goalposts whenever they win. Hold the plan for a full twelve months even when the margin impact exceeds the model. Absorb the overage, learn from it, and adjust at the plan-year boundary. The one legitimate exception is a genuine structural break — a pricing change, a segment reorganization, a product line sunset — and even then, the right move is to grandfather in-flight opportunities and communicate the change with weeks of notice.

Capping the top. Caps are a tax on your best people and a subsidy to your competitors' recruiting teams. A top-quartile rep frequently produces several times the pipeline of the median. Paying that rep well above OTE in a strong quarter is among the best-returning spend in the sales budget. Finance teams push for caps because uncapped variable is hard to forecast; the correct answer is to model the tail explicitly and reserve for it, not to eliminate it.

Quarterly instead of monthly payout. SDRs are early-career, and the psychological distance between effort and reward matters enormously. A commission that arrives eleven weeks after the work does not function as an incentive; it functions as a surprise deposit. Monthly payout with next-day statement visibility — the rep books a qualified meeting today and sees the credit tomorrow — is the single cheapest retention improvement available, and it costs nothing except commission-system configuration.

Plan complexity. If a rep cannot compute their own commission on a napkin, the plan has failed regardless of how elegant the model is. Multipliers stacked on modifiers stacked on kickers produce reps who stop optimizing because they cannot tell what optimizing means. Three components is a reasonable ceiling: the per-opportunity rate, the accelerator tier, and one periodic bonus or SPIF. Everything else belongs in recognition programs, not the comp plan.

The AE-acceptance gate weaponized. The most common failure of the accepted-opportunity model: AEs realize they can suppress SDR commission by slow-walking dispositions or aggressively disqualifying. Guard against this with a hard SLA — unactioned opportunities auto-accept after the window closes — and with a disqualification audit where a manager samples rejected opportunities monthly. If one AE's rejection rate is a large outlier versus peers, that is a management conversation, not a comp problem. Without these guardrails, the gate transfers pay decisions to people who have an incentive to say no.

Territory inequity. Two reps on identical plans working unequal territories will produce unequal outcomes that have nothing to do with skill, and the disadvantaged rep will leave. Review territory balance every plan cycle, and rotate list assignments periodically so no rep permanently owns the graveyard. If territory quality genuinely cannot be equalized, adjust quota rather than pretending it is equal — reps can accept a harder patch with a lower number; they cannot accept a harder patch with the same number.

AI-augmented capacity without a base adjustment. When prospecting tooling absorbs most of the research and drafting work, per-rep capacity rises substantially. The temptation is to raise quota and leave base flat, which top reps read exactly as it is: significantly more expected output for the same guaranteed pay. The defensible move is to raise both — a meaningful base increase alongside the quota increase — so the rep's earnings rise even as revenue-per-headcount improves. Both sides win, and the plan survives contact with a rep who does the arithmetic. Skip the base adjustment and your best people will do that arithmetic anyway, then take it to a competitor.

Over-indexing on comp as the retention lever. Pay fixes pay problems. It does not fix a manager who runs no coaching cadence, a promotion path with no published criteria, or a role with no visible next step. Exit-interview data consistently shows career-path opacity ranking alongside compensation as a departure driver. A plan that pays well inside a role with no future retains for about four quarters and then stops working. Pair the comp redesign with a published promotion rubric and a documented enablement track; the two together do far more than either alone.

Adjacent-team spillover. Changing the SDR payable event changes marketing's sourced-pipeline number, sales ops' attribution model, and possibly the board deck. Socialize the definition change before it lands, and republish historical numbers under the new definition so the quarter-over-quarter comparison is apples to apples. Teams that skip this step spend the next board meeting explaining an apparent pipeline decline that is purely definitional.

A practical rollout plan

Run the change on a ninety-day clock with a shadow period. Rushing it is the second-most-common reason good plan designs fail in practice; the most common is launching without a backtest.

Days 1–30 — benchmark, backtest, and draft. RevOps pulls current-market compensation data for every metro or band you employ in. The comp lead builds a model showing gross-margin impact at 80%, 100%, 120%, and 150% attainment — four scenarios, not one, because finance will ask and you want the answer ready. Critically, backtest: take the last four quarters of actual per-rep activity, apply the new payable-event definition retroactively, and see what each rep would have earned. This surfaces quota calibration errors and identifies which specific reps take a haircut under the new plan, which is exactly the list you need before communication starts. Deal desk drafts the one-page acceptance definition. Finance signs the cost envelope, including the tail reserve for uncapped accelerators.

Days 31–60 — build and communicate. The commission platform team maps CRM stage IDs to payable events and configures the clawback rule. Meanwhile, communication runs on two tracks. Track one is a team-wide session walking through the structure, the rationale, and the acceptance definition — with AEs in the room, because the gate is a joint contract. Track two is a private one-on-one with every rep showing *their* modeled earnings at 80%, 100%, and 120% attainment, using their own trailing data. Do not skip track two. A rep who learns their comp is changing in a group meeting spends the following week on LinkedIn; a rep who sees their personal numbers with their manager asks questions instead. Reps who have not signed by roughly day 55 escalate to leadership — not as a threat, but because unsigned means unresolved and unresolved means a resignation you have not received yet.

Days 61–90 — shadow and switch. Run both plans in parallel for one full month and pay the higher of the two. This costs one month of modest overpayment and eliminates the overwhelming majority of rep anxiety, because it converts an abstract risk into an observed outcome. Around day 75, publish a delta report showing aggregate and per-rep differences. Address the losers individually — usually the fix is quota calibration or territory, not the plan structure. Switch live at day 90.

Ongoing. A comp committee reviews quarterly: attainment distribution, cost of pay per unit of sourced pipeline, dispute volume and resolution time, and voluntary attrition segmented by performance quartile. That last metric is the one that answers whether the plan is doing its job. Review, but do not change mid-year. Feed everything you learn into the next plan-year design.

One adjacent note worth carrying into the rollout: the same ninety-day structure works for redesigning AE, customer success, and partner compensation, and if you are touching more than one plan in a year, sequence them so the SDR plan lands first. The SDR payable event feeds every downstream pipeline metric, so defining it cleanly makes the subsequent plan designs easier rather than harder.

Related questions

Should SDR quota be based on meetings, opportunities, or revenue?

Accepted opportunities. Meetings reward calendar entries; revenue sits too far downstream and depends on AE execution the SDR does not control. Accepted opportunity is the last outcome the SDR meaningfully owns, which makes it the fair and behaviorally correct payable event.

How do you compensate SDRs during ramp?

Pay a declining ramp guarantee — a percentage of target variable that steps down over the first three to four months — regardless of attainment. This prevents new hires from taking a de facto pay cut for a learning curve you designed, and it materially reduces first-year attrition.

Do SPIFs help or hurt SDR retention?

Short, targeted SPIFs on a specific segment or product help. Constant SPIFs hurt: they train reps to wait for a bonus before working, and they obscure the base plan's signal. Cap SPIF spend at a small fraction of total variable and run them for defined windows only.

How often should SDR compensation plans change?

Annually, at the plan-year boundary. Mid-year changes correlate strongly with voluntary attrition because reps read them as goalpost-moving. The only justified exception is a structural business change, and even then grandfather in-flight opportunities and give weeks of notice.

Should remote SDRs be paid on location or on role?

Most companies land on two or three broad geographic bands rather than per-city rates or a single national number. Per-city is administratively brutal and creates relocation arbitrage; a single national rate simultaneously overpays low-cost markets and loses reps in expensive ones.

FAQ

What base-to-variable split retains SDRs best?

Roughly 65/35 for mid-market SaaS, adjusted by sales-cycle length. Longer, enterprise-style cycles justify a heavier base — closer to 75/25 — because the feedback loop is too slow for high variable to feel like earnable compensation. Short-cycle SMB roles tolerate 60/40 because effort converts reliably within the month. The mix should match how quickly a rep's work turns into a payable event, not what a benchmark table says the median is.

Why gate variable on accepted opportunities instead of booked meetings?

Because booked meetings are trivially cheap to produce, especially with AI-assisted prospecting doing most of the research and drafting. Paying on booked meetings pays for calendar entries and quietly funds bad-fit pipeline that erodes AE trust and inflates forecast. Accepted opportunity is the last outcome the SDR genuinely controls, and gating there aligns the rep, the AE, and the forecast on the same definition of a good meeting.

Do clawbacks damage morale?

Not when they are narrow, written down in advance, and applied consistently. A clawback limited to disqualified or clearly bad-fit opportunities inside a defined window reads as a quality standard, not a penalty. Clawbacks damage morale when they are broad, retroactive, discretionary, or applied inconsistently across reps. Publish the rule with the plan, apply it identically to everyone, and give reps a dispute path that resolves inside one pay period.

Are tenure bonuses worth the cost?

Almost always, because the comparison is not against zero — it is against the full-loaded cost of replacing the rep, which includes recruiting, onboarding, management time, and the pipeline lost during vacancy plus ramp. A bonus at month 13 and a larger one at month 24 targets the window where attrition concentrates and reps are most poachable. Present it to finance as turnover-cost offset rather than headcount inflation and it typically clears without argument.

Should top performers' earnings be capped?

No. Caps are a tax on the people who generate a disproportionate share of your pipeline, and they hand your competitors' recruiters an easy pitch. If uncapped variable makes forecasting uncomfortable, model the right tail explicitly and reserve for it. The correct response to a rep earning far above OTE is to check whether the quota was calibrated correctly for next plan year — not to cut the rep off mid-year.

What breaks a good compensation plan in practice?

Four things, in order of frequency: changing the plan mid-year, launching without backtesting against real historical rep data, making the plan too complex for a rep to compute in their head, and treating compensation as a substitute for coaching and a visible career path. A well-designed plan inside a role with no next rung retains for about four quarters and then stops working.

Sources

• Bridge Group — SDR metrics and compensation research: https://blog.bridgegroupinc.com/sales-development-metrics
• RepVue — crowd-sourced SDR compensation and satisfaction data: https://www.repvue.com/salaries/sales-development-representative
• Pavilion — GTM community research and pulse surveys: https://www.joinpavilion.com/
• SaaStr — sales compensation frameworks and commentary: https://www.saastr.com/a-framework-and-some-ideas-for-your-first-sales-comp-plan/
• Pave — real-time compensation benchmarking: https://www.pave.com/
• Everstage — sales commission and variable pay resources: https://www.everstage.com/
• Vendr — SaaS pricing and vendor cost benchmarks: https://www.vendr.com/
• Gartner — sales performance management research: https://www.gartner.com/en/sales
• Harvard Business Review — sales compensation research: https://hbr.org/topic/subject/sales

Related on PULSE

• [How to design Customer Success compensation tied to NRR in 2027](/knowledge/ra0299)
• [How to structure CRO compensation at $50M ARR in 2027](/knowledge/ra0293)
• [SDR Academy Operating Model in 2027](/knowledge/ra0449)
• [SDR to AE Handoff Architecture at $30M ARR in 2027](/knowledge/ra0407)
• [SDR to AE Promotion Criteria + Path in 2027](/knowledge/ra0261)
• [SDR Ramp Model for SaaS in 2027](/knowledge/ra0212)