A founder I worked with last year had a comp plan that looked perfect on paper. A clean 50/50 split, accelerators after 100 percent, a kicker for multi-year deals. He paid market rate. He still lost his two best reps inside six months.
When I dug in, the model was not the problem. The problem was that nobody trusted the number. Deals were getting marked closed-won in the CRM with the wrong amount. Renewals were counted as new business one quarter and not the next. Commission ran on a spreadsheet a finance analyst rebuilt by hand every month, and it was wrong often enough that both reps kept their own private tracker. By the time a payout landed, half the conversation was an argument about what actually counted.
That is the part most articles about sales compensation skip. They give you pay mix ratios and accelerator curves and quota multiples, all of which matter. But a comp plan is not a document. It is a calculation that runs on your CRM data every single month, and if that data is dirty or the logic lives in someone's head, the best-designed plan on paper still breaks. I have built and fixed comp plans for B2B teams for over ten years, and the failure is almost never the formula. It is the system underneath it.
Why most comp plans quietly fail
Start with the numbers, because they are worse than people admit. Average quota attainment sat at about 43 percent in late 2024, and across SaaS teams reps typically hit 50 to 60 percent of quota. Only about 24 percent of salespeople beat their annual number. So the median rep on your team is, by design, missing the target you set. If your plan only feels good when someone hits 100 percent, it feels bad to most of your team most of the time.
Now add turnover. Sales roles churn at roughly 35 percent a year, almost three times the all-industry average, and replacing one rep runs around 115,000 dollars once you count recruiting, ramp, and the pipeline that went cold while the seat was empty. Comp is not the only reason people leave, but it is a big one, and it is the one you control directly.
Here is the stat that reframes the whole thing for me.
Share of a rep's monthly time the Aberdeen Group found goes to shadow accounting: building a private spreadsheet to check what the commission system says they earned.
Read that again. A quarter to half of a rep's month, gone, because they do not believe the official number. That is not a comp design issue. That is a data and operations issue wearing a comp costume. WorldatWork has found that 22 percent of reps file at least one commission dispute a year, and about 9 percent of voluntary sales resignations trace back to pay transparency problems. Meanwhile teams that fixed transparency and made payouts traceable improved retention by 12 to 15 percent.
So when I audit a comp plan, I am really auditing two things. The economics of the plan, and the machine that computes it. Most teams obsess over the first and ignore the second.
Get the economics right first
You still need the model to be sound, so let me cover what good looks like before I get to the data layer.
Pay mix. For closing roles like account executives, a 50/50 base to variable split is the default and a fair one. The reasoning is simple: you want a meaningful chunk of pay tied to results, but not so much that a slow quarter wipes someone out. Support-adjacent roles like customer success or sales engineering usually sit at 70/30 or 80/20 toward base, because they influence revenue without owning the close. The 2024 Bridge Group data puts the median SaaS AE on-target earnings around 190,000 dollars with a 53/47 split, so if you are far off that for similar roles and markets, you have a recruiting problem coming.
Quota relative to OTE. A common rule is to set quota at four to six times a rep's on-target earnings. SMB reps doing higher volume and smaller deals often sit near 4x. Enterprise reps with bigger deals stretch to 6x. That math implies commission rates somewhere around 10 to 20 percent of new annual contract value. If your quota is set so high that almost nobody clears it, you do not have an ambitious plan, you have a base-salary plan with extra steps, and your reps know it.
Accelerators and caps. About 82 percent of SaaS companies pay accelerators above quota, fewer than 15 percent cap commission, and more than half use clawbacks tied to churn. I am firmly against caps. The moment a rep realizes more effort earns nothing, they stop, or they sandbag deals into next quarter. Pay your top performers uncapped. The deals they bring in past quota are the most profitable revenue you will book all year.
Ramp. New hires need a 3 to 6 month ramp with prorated quota or a guaranteed draw. Asking a rep to hit full number in month one is how you turn a 115,000 dollar hire into a 115,000 dollar write-off.
There is plenty of nuance underneath all of that, but if you get pay mix, quota ratio, accelerators, and ramp roughly right, the economics are not where your plan will fall apart. The next part is.
The part nobody designs for: the data
Every comp plan rests on a few definitions that sound obvious and are not. What counts as closed? Is it signed contract, or first payment received? Does an upsell on an existing account credit the AE who closed it originally, or the one working it now? When a deal is split across two reps, what is the rule? Does a downgrade at renewal claw back the original commission? When exactly does a deal "count," at booking or at start date?
If those answers live in a Slack thread and a finance analyst's memory, your plan is undefined. Reps will fill the gap with their own assumptions, and their assumptions will always be more generous than yours. That is where disputes are born.
I have a simple test. Ask a rep to calculate their own commission for last month from the CRM, with no help. If they cannot, the plan is too complex or the data is too messy, and you are paying for that gap in shadow accounting time and lost trust. The most common root cause of disputes is not greed. It is plan language packed with terms like "net revenue" or "qualified deal" that nobody defined the same way twice.
If a rep cannot calculate their own commission from the CRM, you do not have a comp plan. You have a comp argument waiting to happen.
Every undefined term and every dirty field is a future dispute. Clean data and clear rules are the plan, not the appendix to it.
This is exactly where comp design and revenue operations meet, and why I treat them as one job. A comp plan is a set of business rules that run on CRM objects. If your deal stages are inconsistent, if "closed-won" gets used loosely, if amounts are entered wrong or currency is not normalized, the plan inherits every one of those errors and pays them out as real money. Fixing the plan without fixing the CRM data quality underneath is like repainting a house with a cracked foundation.
What changes when reps trust the number
Here is the contrast I see between teams that get this right and teams that do not. The plan economics can be identical. The experience is not.
The difference is not generosity. It is traceability. A rep who can see exactly which deals built their check, with the math shown, stops keeping a private ledger. That reclaims a big slice of selling time and removes the slow trust erosion that pushes good reps toward recruiters. Visibility is the cheapest retention lever you have, and most teams leave it on the floor.
How to build a plan that holds up
When I rebuild a comp plan for a B2B team, I run the same sequence every time. It is less about the curve and more about the plumbing.
A few notes on that sequence.
Steps one and two come before the economics on purpose. I have watched too many teams design a beautiful plan and then try to bolt it onto a CRM that cannot answer "what closed this month" without a human cleaning the export. Do the definitions and the data first, or you will rebuild the plan twice.
On step three, keep it to two or three metrics per role. Reps lose confidence the moment they cannot do the arithmetic in their head. A plan with six components and four conditional kickers is not sophisticated, it is a trust problem you are choosing to create. Simplicity is not dumbing it down. It is making the plan legible to the person it is meant to motivate.
On step four, you do not need expensive commission software to start. Plenty of teams run clean automated comp on top of HubSpot or Salesforce with a workflow layer that pulls the right fields and produces a per-rep, per-deal statement. The point is not the tool. It is that the number a rep sees comes from the same source of truth as everything else, not a parallel spreadsheet that drifts. When a team outgrows that, dedicated tools like CaptivateIQ or similar make sense, but the data discipline has to come first either way. This is the kind of glue work we handle in AI and workflow automation projects.
On step five, the attainment spread is your early warning system. A healthy plan has roughly 60 to 80 percent of reps in range of quota, with a real top tail pulling away on accelerators. If almost everyone is missing, the quota is set on fantasy, not capacity. If everyone clears it on day twenty, you are leaving growth on the table and overpaying for it. I tie this review to the same cadence as sales forecasting, because both run on the same pipeline data and both fall apart for the same reason when that data is wrong.
The connection to the rest of your revenue engine
Comp does not sit in a corner. It is wired into everything else, which is why fixing it usually means fixing more than the plan.
Quota depends on territory and capacity, so a comp plan built on quotas that ignore real coverage will misfire no matter how clean the math is. Crediting rules depend on consistent deal ownership, which is a CRM adoption problem before it is a comp problem. And the whole thing depends on reps logging deals accurately, which they will only do if they trust the system to pay them for it. It is a loop. Good comp reinforces good data, and good data makes comp trustworthy. Break either side and both decay.
That is the lens I bring to this work. I do not see a comp plan as an HR document or a finance spreadsheet. I see it as one of the highest-stakes calculations your revenue operation runs, and like any calculation, it is only as good as the data feeding it and the rules defining it. Get those two right and the plan mostly takes care of itself. Get them wrong and no accelerator curve will save you.
Comp plan looks fine but reps keep leaving?
Book a free 30-minute audit. We will look at your plan, your CRM data, and the gap between them, and show you the three fixes we would make first.
Book an audit →Frequently asked questions
What is a good pay mix for a B2B sales rep?
For closing roles like account executives, a 50/50 base to variable split is the standard and works well. It ties a real share of pay to results without making a slow quarter catastrophic. Customer success and sales engineering roles usually sit at 70/30 or 80/20 toward base because they influence revenue without owning the close. Match these to your market, since a median SaaS AE now sits near 190,000 dollars OTE with roughly a 53/47 split.
How high should I set quota relative to OTE?
A common rule is four to six times on-target earnings. SMB reps running higher deal volume often sit near 4x, while enterprise reps with larger deals stretch toward 6x. The real test is the attainment spread: aim for 60 to 80 percent of reps landing in range of quota. If almost nobody hits it, the quota is built on hope rather than real capacity, and you will pay for that in turnover.
Should I cap sales commission?
No. Fewer than 15 percent of SaaS companies cap commission, and for good reason. The moment a rep sees that extra effort earns nothing, they either stop or push deals into the next period. Revenue booked past quota is usually your most profitable, so pay it out fully. Use accelerators above target instead, which most teams already do.
Why do my reps keep their own commission spreadsheets?
Because they do not trust the official number, which is one of the clearest signals your comp data is broken. The Aberdeen Group found shadow accounting can eat 25 to 50 percent of a rep's month. It almost always traces back to a manual calculation process, inconsistent CRM data, or plan terms that were never defined the same way twice. Fix the data and make payouts traceable, and the private spreadsheets disappear.
How often should I change the comp plan?
Review it every quarter using your attainment data, but change the structure no more than once a year unless something is clearly broken. Reps need stability to plan their own selling, and constant tweaks read as the company moving the goalposts. Quarterly reviews are for spotting problems early. Annual changes are for acting on them, communicated clearly and well before they take effect.
If your comp plan looks right on paper but your best reps are still walking, the answer is usually under the hood, in the data and the rules nobody wrote down. That is exactly the kind of work we do at Ziel Lab. Get in touch and we will help you build a plan your team actually trusts.