Compensation design

I once shipped a plan with a clean 2x accelerator over 100% and watched it teach my best rep to sandbag. She closed a $180K deal on October 2nd that had been ready to sign in the last week of September. She was not lazy. She was rational. Landing it in the new quarter, where she was already past 100% and sitting in the 2x band, paid her nearly double what it would have paid in the old quarter, where she was still climbing to target. The plan rewarded the behavior and I paid for deal-timing games. I got exactly what I paid for.

Comp is the most expensive behavior-modification tool a company owns, and reps read the plan more carefully than any strategy deck. Every clause is a lever. Every accelerator is an instruction. The plan you write is the plan they run, whether or not it matches the plan you meant. So this is what I actually build now: pay mix mapped to control, a quota-to-OTE ratio checked against a hard red line, a commission rate that falls out of the math instead of getting picked, and a payout curve I have modeled at the quarter boundaries before finance ever sees it.

Pay mix tracks how much of the outcome the role controls

On-target earnings is base plus variable at 100% attainment. Pay mix is the base-to-variable split, and it should map to how much of the result the role actually controls. A rep who owns the whole deal carries more variable, because more of the outcome is in their hands. A CSM whose renewal depends on product quality and onboarding as much as their own effort carries less, because punishing them for a churn caused by a broken product is not an incentive, it is a tax.

That single principle sets the whole table. The more of the number a role can personally move, the more of their pay you can put at risk without it feeling arbitrary. Here is where the market sits in 2026:

Role	Pay mix (base/var)	OTE band	Quota (annual)	Commission on ACV
SDR / BDR	65/35	$70-90K	8-12 SQLs/mo	$150-400 per SQL
AE, SMB	50/50	$130-170K	$500-750K	8-12%
AE, mid-market	50/50	$180-240K	$750K-1.2M	9-12%
AE, enterprise	55/45	$240-320K	$1.2-2.0M	8-10%
Sales engineer	75/25	$150-200K	overlay	2-4%
CSM, SMB	80/20	$80-110K	GRR/NRR book	3-6% expansion
CSM, mid-market	75/25	$110-150K	GRR/NRR book	4-8% expansion

Source: GetGangly 2026 compensation guide citing Bridge Group, Everstage, and RepVue.

Read the pay-mix column top to bottom and the logic is visible. The SDR sits at 65/35 because they influence pipeline creation but do not close, so a smaller slice rides on an outcome they only partly own. The AE sits at 50/50 because they own the deal end to end. The enterprise AE tips to 55/45 because the deals are larger and the reward for landing one is worth putting more at risk. The SE and CSM ride gentler because they influence rather than own. Get this backward, put a CSM on 50/50 tied to logo retention, and you have built a plan that pays out on outcomes the person cannot control, which reads to them as a lottery, not a comp plan.

The comp signal chain

Every clause in a plan is downstream of one decision: what behavior do you need more of. Name that first, and the rest of the plan falls out of it in order. Skip it, start with a number, and you end up with a plan that pays for the behavior you happened to encode instead of the one you needed.

The rate is the last thing you compute, not the first. Once behavior, OTE, pay mix, and the quota-to-OTE check are set, the commission rate is arithmetic:

Commission rate falls out of the plan

commission rate = variable comp / quota

The median AE earns about 11.5% of bookings in commission at 100% quota (Everstage 2026). If your derived rate lands far off that band, the problem is upstream in your quota-to-OTE ratio, not in the rate itself.

The payout curve is the message

The shape of the curve tells the rep where to spend their last hour of the quarter. A flat rate says every deal is worth the same, so coast to target and stop. An accelerator over 100% says the last dollar is the most valuable dollar you will book all year, so keep going. That is the entire point of the accelerator, and it is why capping commission is an anti-pattern: a cap tells your best rep to stop selling in November.

Toggle the two curves below. The flat plan pays a straight 10% at every attainment level. The accelerated plan is the same up to 100%, then steepens. The gap between the two orange bars past target is the money you are using to buy over-performance, and it is real money, so it has to buy real behavior.

Accelerated (2x over 100%)Flat 10%

View as table

Stage	Value
80% ($480K)	48K
100% ($600K)	60K
120% ($720K)	84K
140% ($840K)	108K

At 140%, the accelerated plan pays $108K against the flat plan’s $84K. That $24K difference is what you are spending to keep your best rep pushing after they have already cleared target. Spend it deliberately. If the extra bookings past 100% are worth more than $24K in margin, the accelerator prints money. If they are timing games dressed up as over-performance, you are paying a premium for pipeline theater.

The quota-to-OTE red line

Quota-to-OTE is the ratio of a rep’s annual quota to their total on-target earnings. It is the single number that tells you whether a plan is even reachable. The market lives in a narrow band, and stepping outside it breaks the plan in a predictable way.

Quota-to-OTE	Read	What it does to behavior
Under 4x	Rich	Comp as a share of revenue is high; finance will push back
4-5x	Standard	The healthy band for most SMB and mid-market roles
5.5-6x	Enterprise-normal	Larger deals justify a thinner comp ratio
Over 6x	Structurally broken	Accelerators are unreachable; reps stop chasing over-performance

Source: GetGangly 2026 quota benchmarks; the 6x red line is a standard RevOps guardrail.

The over-6x row is the one that quietly kills plans. When quota is set so high that a rep can barely touch 100%, the accelerators above target become decoration. Nobody reaches them, so they incentivize nothing, and you have paid your comp designer to build a curve that never bends. Worse, reps do the math on day one, conclude the number is unreachable, and mentally check out of the over-performance game before Q1 closes. The plan reads as a message, and the message is: we do not believe you can do this.

This connects directly to how you built the number in the first place. A quota that clears the 6x line usually means the capacity model over-assigned, which happens most often when planners multiply nominal headcount by full quota and ignore ramp. Fix the capacity math and the quota-to-OTE ratio tends to fall back into the healthy band on its own.

Comp CSMs on retention, not activity

The plan I see broken most often is the CSM plan that pays on QBRs held or logos touched. That comps activity, and activity is not the outcome anyone is buying. Tie CSM variable to GRR and NRR, plus an expansion component in the 3 to 8% range on the revenue they grow.

It matters more now than it used to. Expansion is running around 52% of new revenue (gradient.works 2025), and expanding an existing account costs roughly $0.80 per dollar of ARR against $1.63 to acquire a new one (Aleph×Benchmarkit 2026). NRR at the median B2B SaaS company sits at 102%, with the top quartile at 110% and the bottom at 92% (Aleph×Benchmarkit 2026). If the CSM plan does not pay for expansion, you are leaving the cheapest revenue in the company on the table and calling it a retention role.

	The plan I shipped	The plan I ship now
CSM primary metric	QBRs held, logos touched	GRR + NRR, expansion 3-8%
AE accelerator	Steep 2x, no quarter guard	2x with quarter-boundary look-back
Quota check	Set top-down, no ratio test	Quota-to-OTE held under 6x
Modifiers	Five kickers, unreadable	One primary metric, one accelerator
Rep can self-calc	No, needs a spreadsheet	Yes, in under a minute

Does the cost pencil out

A comp plan that motivates the right behavior can still be a bad plan if it costs too much to acquire the revenue it drives. Before I sign a plan, I check what the fully loaded selling cost does to CAC payback. If layering on richer accelerators or a bigger team pushes payback past 18 to 24 months, the plan is inefficient even when it hits the number. Move the sliders and watch where the line crosses.

What richer comp does to payback

S&M spend÷( new ARR×gross margin )=10

months to payback

S&M spend / period ($)$600kNew ARR won ($)$900kGross margin80.0%

TryEfficient (10mo)Bloated (28mo)

Under 12 months is efficient, 12 to 18 is normal for mid-market, past 24 months you are buying revenue faster than it pays you back. Watch this before you scale spend.

months to payback: 10

Under 12 months is efficient. Twelve to 18 is normal for mid-market. Past 24 you are buying revenue faster than it pays you back, which is the moment to trim the plan or fix the funnel feeding it, not to keep spending.

The build order

This is the order I work in, every time. Behavior first, rate last. Each step feeds the next, so doing them out of order means redoing them.

1. 1

   Name the one or two behaviors the company needs

   New logos, expansion, multi-year, a specific product. Name them before touching a number, because the primary metric is the whole message and everything else is volume knobs on it.

2. 2

   Set OTE from market by role and geo

   Benchmark against Bridge Group, RepVue, and Everstage. Do not guess. An OTE set from a gut number is the first crack most plans hide.

3. 3

   Choose pay mix by deal control

   Aggressive for AEs at 50/50, gentler for SDRs at 65/35, gentler still for SE and CS at 75/25 to 80/20. Match risk to how much of the outcome the role owns.

4. 4

   Check quota-to-OTE against the red line

   4 to 5x standard, 5.5 to 6x enterprise. Above 6x is structurally broken and the accelerators become decoration. Fix the quota before proceeding.

5. 5

   Derive the rate, then layer accelerators

   Rate is variable divided by quota, and it should land near the 11.5% median. Add kickers above target and clear decelerators below.

6. 6

   Model 50 / 80 / 100 / 130% and the quarter boundaries

   Check total comp as a share of revenue at each attainment level, and check the incentive to hold deals across a period line.

7. 7

   Document, get finance sign-off, automate the payout

   Administer through a comp tool with a dispute process. A rep shorted once distrusts every future statement.

Audit the quarter boundaries

The sandbag I got burned by leaves a fingerprint in the data: deals that close in the first days of a quarter after sitting ready to sign at the end of the last one. You can watch for it. Pull closed-won opportunities that landed in the opening week of a quarter and had a close date pushed at least once from the prior period, and put the list in front of the manager before you cut commission.

SELECT o.Id, o.Name, o.OwnerId, o.Amount, o.CloseDate,
       h.OldValue push_from, h.NewValue push_to, h.CreatedDate pushed_on
FROM Opportunity o
JOIN OpportunityFieldHistory h ON h.OpportunityId = o.Id
WHERE o.IsWon = true
  AND h.Field = 'CloseDate'
  AND o.CloseDate = THIS_FISCAL_QUARTER
  AND DAY_ONLY(o.CloseDate) <= (FISCAL_QUARTER_START + 7)
ORDER BY o.Amount DESC

A short list every quarter is noise. A pattern from the same reps every quarter is the accelerator teaching a behavior, and the fix is a quarter-boundary look-back clause in the plan, not a conversation. Reps optimize the plan you wrote, so change the plan.

Two guardrails I never break

The two rules I hold above all the others. First, the 6x red line: at that ratio the accelerators are theoretical and reps stop chasing over-performance because they can barely reach 100%. Second, the readability test.

Comp is a set of instructions written in the only language every rep reads to the last line. Write it in that order: behavior, OTE, pay mix, the 6x check, then the rate. Model it at the quarter boundaries before finance signs it. Then watch what your best reps do, because whatever the plan pays for is exactly what you will get.