Nine Comp Plan Mistakes We See in Every Diagnostic, Ranked by Damage
A field-ranked list of the comp design errors that show up in nearly every org we baseline — and the one that costs more than all the others combined.
Every comp plan review starts the same way. Someone slides a spreadsheet across the table — base, OTE, an accelerator table, three years of "this has worked fine" — and I ask three questions: what happened to the rep who used to own this patch, how long did the person who replaced them take to hit full quota, and what did finance quietly change the night before rollout. By the third answer the story usually falls apart. Comp plans rarely fail because the maths is wrong. They fail because the maths was built for someone who no longer holds the job.
We've run enough of these diagnostics to know the failure modes repeat almost verbatim across industries, deal sizes and company ages. What changes is which combination shows up, and in what order. Below are the nine we see most often, ranked by the damage each one does — not by how loudly reps complain about it, not by how uncomfortable it makes finance, but by a rough sum of ramp time lost, attrition triggered, and revenue left on the table in the first twelve months the plan is live.
How we're measuring damage
Some of these mistakes are annoying. Some are expensive. One of them is a business-model problem wearing a spreadsheet's clothes, and it's worth separating from the rest before you read the list, because it changes how you should prioritise fixing it.
The nine, ranked
| Rank | Mistake | Core damage |
|---|---|---|
| 1 | Territory and quota inherited from the departed rep, unadjusted for the new incumbent | New hire measured against relationship equity they don't have; six-month attrition spike |
| 2 | Ramp exists in the onboarding deck, not in the payout formula | Reps miss plan in month one by design, then keep missing it for real |
| 3 | Quota set top-down from a revenue target, no bottoms-up capacity check | Territories that were never winnable in the first place |
| 4 | One plan spanning segments with different sales cycles | SMB reps punished for a plan built around enterprise deal velocity |
| 5 | Accelerator curve with a comp cliff | Marginal rate drops as attainment rises; top performers throttle their own back half of the quarter |
| 6 | New logo and renewal paid at the same rate | Hunting behaviour disappears once a rep's book has enough renewals to coast on |
| 7 | Draws and guarantees with vague recovery terms | Termination disputes and clawback fights, occasionally worse |
| 8 | SPIFFs that contradict the core plan's incentive direction | Reps chase the SPIFF and quietly ignore what the plan is meant to reward |
| 9 | Mid-cycle plan changes with no grandfather period | Trust damage that outlasts the plan year itself |
A word on the gap between rank one and everything below it: it isn't a gentle slope. It's a cliff, and it earns its own section.
Why rank one dwarfs the rest
Picture the territory as it actually exists the day a rep leaves. Three years of quarterly business reviews. A referral flywheel that took eighteen months to spin up. Champions who pick up the phone because they trust the person, not the logo in the email signature. None of that is in the CRM. None of it transfers with the account list.
Now picture what actually gets handed to the replacement: the same forty accounts, the same quota number, sometimes the same OTE, occasionally a "should be easier, the accounts are already warm" from the hiring manager. The new rep isn't inheriting a territory. They're inheriting someone else's trailing twelve months of relationship capital, minus the relationships.
This is why the standard ramp assumption — three months at partial quota, back to full target by month four — falls apart specifically on inherited territory. A brand-new patch of cold accounts at least comes with honest expectations attached. An inherited patch looks easy on paper and is quietly much harder, because the rep first has to unwind the previous rep's promises, half-finished proposals and champion relationships before they can build their own. Boards and finance teams almost never model this distinction. They see "existing book of business" and assume "easier ramp." It is frequently the opposite.
The fix isn't complicated. It's just rarely done: rebuild the quota from the new incumbent's actual starting position, not the account list's historical performance. Run it through something closer to a New-Hire Ramp Quota Schedule Template before the offer letter goes out, and separately check whether the territory itself was ever fairly carved to begin with — a Territory Design Fairness Scorecard catches a surprising number of "warm" territories that were only ever warm for one specific person.
The rest of the list, briefly
Ranks two through four share rank one's root cause — someone modelled the plan around an abstraction (a revenue target, a segment average, a generic ramp curve) instead of the actual person expected to hit it. They cost less mostly because they surface faster: a rep 40% behind quota in month two gets a coaching conversation, whereas a rep quietly failing against an impossible inherited book often doesn't get flagged until the exit interview.
Ranks five and six are the ones sales leaders assume are the biggest problem, because they're the ones reps actually complain about in Slack. A comp cliff — where the rate on units above 100% attainment is somehow worse than the rate below it, usually thanks to a badly drafted tier boundary — genuinely does throttle your best performers' back half of the quarter. Run any accelerator table through an OTE Calculator (On-Target Earnings Breakdown) before you ship it; if total earnings per unit sold ever falls as attainment rises, you've built a cliff, not an accelerator.
Ranks seven through nine are hygiene problems: real, expensive, entirely avoidable, but plan administration failures rather than plan design failures. You fix them with clearer documentation and a grandfather clause, not a redesign.
The uncomfortable bit
Fix only mistake one and leave the other eight exactly as they are, and you'd still see meaningfully better ramp times and lower first-year attrition than an org that fixed everything except mistake one. That isn't an argument for ignoring the other eight. It's an argument for retiring the habit of treating comp design as a spreadsheet exercise when the actual defect sits upstream of the spreadsheet, in how the territory got handed over in the first place.
Most comp plan post-mortems spend their energy on accelerators and caps because those are the visible, arguable-over-Slack parts of the plan. The one that actually decides whether your next hire survives to their first renewal cycle gets settled before anyone opens the compensation model at all.