Show me a sales team's behavior and I will show you their comp plan, because they are the same document written two ways. Reps do not do what you say in the kickoff. They do what the plan pays them to do, and they find the shortest path to it by the second week.
Sales compensation is the system of base pay and variable pay you use to fund and direct a sales force. The word that matters is direct. A comp plan is not a payroll cost to minimize. It is a steering wheel. Every rule in it, the split, the quota, the accelerator, the gate, tells a rep which behavior earns the most money, and a good team will optimize for that behavior whether or not it is the one your revenue plan needs. I have built revenue and forecast models for B2B SaaS companies for two decades, and the most expensive plans I have seen were not the ones that paid the most. They were the ones that paid handsomely for the wrong thing, hitting their payout target while missing the revenue target by a mile.
So this is not a survey of comp structures. It is a method for making the plan pay for the behavior you actually want, plus the failure modes that quietly do the opposite.
The four levers, and what each one is for
A comp plan has exactly four levers worth tuning. Most teams obsess over the commission rate and leave the other three on inherited defaults. Each lever does a specific job, and the job is the point.
| Lever | What it controls | What it is for | What it breaks when wrong |
|---|---|---|---|
| Base / variable split | How much pay is guaranteed vs at-risk | Matching risk to cycle length and role | Too much variable on a long cycle starves reps and drives churn |
| Quota | The bar that defines 100 percent of OTE | Setting the line between miss, hit, and beat | A quota nobody hits demotivates; one everyone hits is a giveaway |
| Accelerators | The rate on revenue above quota | Making overperformance irresistible | Caps tell top reps to stop selling at 100 percent |
| Gates and modifiers | Conditions on payout, like deal quality | Steering toward the right kind of deal | The wrong gate produces gaming, not quality |
The split matches at-risk pay to how much control a rep has and how long they wait. A transactional role closing in three weeks can run 40/60. An enterprise role on a nine-month cycle needs 60/40, because no one survives three quarters of pipeline-building on commission. Quota sets the line that defines on-target earnings, so it comes from the same model that sets the sales plan and the capacity plan, not from last year plus ten percent. A quota disconnected from real coverage either pays out on a bar nobody clears or sits so low attainment is automatic, which is salary wearing a commission costume. Accelerators and gates I take up below, where the sharpest gains and the deepest self-inflicted wounds both live.
The Behavior Ledger: design the plan from the behavior backward
Here is the method I use, and it inverts how most plans get built. I call it the Behavior Ledger. Most teams start from the structure, a split, a rate, a quota, and hope the right behavior falls out. The Behavior Ledger starts from the behavior you want and works back to the rule that pays for it. Three columns, filled left to right, and you do not write a single number until the first two are done.
Column one: the behavior the revenue plan needs. Not a goal, a behavior. "Close more enterprise logos." "Protect average deal size." "Sell the multi-year contract." Each is something a rep does or does not do on a Tuesday. Column two: what the current plan actually pays for. Trace the money honestly. Flat commission across deal sizes pays for deal count, so reps chase small fast deals over large slow ones. No multi-year incentive pays for one-year deals, and that is what you get. This column is almost always different from column one, and the gap is the whole problem. Column three: the lever that closes the gap. Now, and only now, you reach for structure. Bigger deals? A deal-size modifier or tiered rate. Multi-year? A bonus on contract length. Discount discipline? A gate that pays full commission only above a margin floor. Each rule exists to drag column two into line with column one.The discipline is the sequence. Design from the structure forward and you get a tidy plan that points reps at whatever the defaults reward, usually deal volume, because flat commission is the lazy default. Design from the behavior backward and every rule has a job you can name. If you cannot say which behavior a rule pays for, it should not be in the plan. That single test strips more dead weight out of a comp plan than any benchmarking exercise.
A worked example: Hollanby uncaps the wrong year
Numbers here are illustrative, not a benchmark. They exist to show the mechanism.
Hollanby is a mid-market B2B SaaS company. Last year's plan was simple: 50/50 split, 800K quota per account executive, flat 10 percent commission on all new ARR, and a cap at 150 percent of quota. Bookings came in on plan, leadership called it a good year, and they rolled the plan forward.
Run it through the Behavior Ledger and the year looks different. Column one, the behavior the plan needed, was twofold: land larger enterprise deals, because the model showed mid-market saturating, and sell more multi-year contracts to firm up retention. Column two, what the plan paid for, was neither. Flat 10 percent paid identically for a 30K mid-market deal closed in a month and a 120K enterprise deal closed in a quarter, so per hour worked, reps were paid to avoid enterprise. Nothing rewarded contract length, so reps sold the easy one-year deal every time. The plan paid for small and short. The team delivered small and short, precisely as designed.
Worse, the cap throttled the two best reps. Both hit 150 percent by mid-November and went quiet, because every deal past the cap paid zero. The plan had trained its strongest closers to stop closing for six weeks during the highest-intent stretch of the year, overperformance the company paid nothing to capture and a competitor was happy to.
The redesign followed the ledger. Hollanby adds a deal-size modifier paying 12 percent on new ARR from deals above 75K, a flat bonus for every multi-year contract, and removes the cap, replacing it with an accelerator that pays 15 percent on every dollar above 100 percent of quota. Same target OTE at 100 percent attainment. Completely different behavior around it, because the plan now pays for big, long, and relentless instead of small, short, and capped.
The point is not the rates. It is that Hollanby's first plan was not broken in any way the bookings number could show. It hit its payout target and its bookings target. It was broken in what it taught reps to do, which surfaced only when someone traced the money behavior by behavior instead of reading the total at the bottom.
Why I will not cap a comp plan
This is the contrarian stake, and I will plant it firmly: do not cap commission on net-new revenue. Not at 150 percent, not at 200 percent, not anywhere. Uncap it and let your best reps run.
The objection is always the same, and it always comes from finance: an uncapped plan creates "windfall" risk, a rep stumbles into a giant deal and out-earns a director. The logic is backward. The rep who can close the next deal after quota is the single cheapest unit of revenue you will ever buy. The pipeline is already paid for, the marketing spend is sunk, and closing one more deal costs nothing incremental, so paying a higher rate on it is the best margin trade you have. A cap takes your most productive closer at the most productive moment and tells them, in the only language a comp plan speaks, to go home. That is not prudence. It is leaving revenue on the table to avoid the discomfort of a rep getting rich by selling, the one thing you hired them to do.
This matters more in a hard market than an easy one. Median B2B win rates have fallen to 19 percent (First Page Sage, 2025) and sales cycles have lengthened 22 percent since 2022 (Digital Bloom, 2025), so the pipeline conversion math is tighter than it was three years ago and the rep who can actually push a deal across the line is rarer and more valuable. Capping that rep is sabotage dressed as cost control.
If windfall genuinely worries the board, fix it where it belongs, in the quota, not the cap. A quota set from real coverage makes a true windfall mathematically unlikely, because the bar already reflects what the territory can produce. Set the right quota and pay every dollar above it without flinching. The dishonest version is a soft quota with a cap bolted on to claw back the overage, which reps read instantly as a bait and switch and never forgive. Caps do not manage windfall. They manufacture sandbagging, because a rep nearing the cap simply parks deals for next period, and now your forecast is wrong on top of your comp plan.
Where comp plans quietly go wrong
Most broken plans break in one of four ways, and each is a behavior the plan paid for without meaning to.
Paying for bookings instead of the right bookings. Flat commission rewards count and speed, so reps optimize for small and fast. Deals close, the dashboard looks busy, and average deal size and win quality erode underneath. The fix is a modifier or tier that pays for the deal shape the model needs, a sales process question before it is a payroll one. Quota set from the org chart, not the pipeline. "Last year plus a percentage" or "the target divided by headcount" loses contact with what territories can produce. Some reps draw an impossible number and disengage; others draw a gift and coast. Both wreck attainment distribution and your forecast. Tie every quota to modeled coverage, the same discipline that drives capacity planning. Capping the people you most want uncapped. Covered above, and common enough to repeat. A cap tells your best closers to stop. Replace it with an accelerator and a defensible quota. Changing the plan mid-year. Reps build their working life around the plan, and every mid-year change resets their trust and their behavior at once. Even a change that improves the math can cost a quarter of momentum while the team re-learns what pays. Reset annually, communicate early, and hold the line, treating comp as part of the broader performance management system rather than a knob you turn whenever a number wobbles. The only justified exception is a single-variable fix, like uncapping an accelerator, where the change only ever helps the rep.Pay for the behavior, not the average
If you take one idea from this, take the inversion. A comp plan is not built from rates and splits with the right behavior hoped for as a byproduct. It is built from the behavior the revenue plan needs, traced back to the rule that pays for it, every rule able to name its job. The total payout can look perfectly healthy while every line above it pays reps to do the opposite of what the model requires, and you will not see it until you read the plan as behavior instead of as cost.
So before you roll the plan forward another year, run it through the ledger. Write down the behavior you need, write down what the plan pays for today, and stare at the gap, because that gap is your real comp problem and the bookings number is hiding it. The plans that work reconcile intended behavior against modeled revenue line by line, the same forecast-grounded discipline ORM builds, so the plan you approve is the one your revenue model actually needs.
Frequently Asked Questions
What is sales compensation?
Sales compensation is the system of base pay and variable pay a company uses to fund and direct its sales force. The plan is not a payroll line. It is a steering input. Every rule in it tells reps which behavior earns the most money, and the team optimizes for exactly that, whether or not it matches the revenue plan.
What is a good base-to-variable split for sales comp?
For a closing account executive in B2B SaaS, a 50/50 base-to-variable split is the common anchor, meaning half of on-target earnings is salary and half is commission. More transactional, high-velocity roles lean toward more variable, like 40/60. Longer enterprise cycles and technical sales lean toward more base, like 60/40, because a rep cannot survive a nine-month cycle on commission alone.
What is OTE in sales?
OTE stands for on-target earnings, the total a rep makes when they hit 100 percent of quota. It is base salary plus variable pay at full attainment. OTE is the number you recruit on and the number the plan is built around. A rep who beats quota earns above OTE through accelerators, and a rep who misses earns below it.
How do sales accelerators work?
An accelerator raises the commission rate on revenue booked above quota. A rep might earn a base rate up to 100 percent of quota, then a higher rate on every dollar past it. Accelerators exist to make overperformance irresistible, because the rep who can close the next deal is the cheapest revenue you will ever buy. Capping them tells your best reps to stop selling once they hit quota.
What is the most common mistake in sales compensation?
Paying for activity or bookings the plan never intended to reward, then acting surprised when reps deliver exactly that. A plan that pays flat commission on any deal will produce small, fast, low-fit deals. A plan that caps earnings will produce sandbagging. The plan is always working. The question is whether it is working toward the revenue you actually want.
How often should you change a sales comp plan?
Reset the plan once a year, aligned to the annual planning cycle, and resist mid-year changes unless something is structurally broken. Reps need stability to chase a number, and every change resets their trust and their behavior. The exception is a clear, single-variable fix, like uncapping an accelerator that is throttling your top performers.
Five free days of implementation
Start with ORM before the end of June and your first five days of implementation are free. We build your forecast model on your live pipeline, then you decide.
Claim your five days