# Sales commission structures for a small team: pick one, then test it on last year’s deals

https://salesbond.app/blog/sales-commission-structures-small-team/
Published: 2026-09-29 · Updated: 2026-10-25

For a small team, start with a flat percentage of closed-won revenue on a base salary, add one accelerator above quota only if reps clear quota regularly, and before announcing anything run the plan against last year’s won deals to see what it would have paid.

That last step is the one most teams skip. They pick a structure from a list of twelve, agree on a rate in a meeting, and find out six months later that one rep earns twice as much as the others because of a single deal that happened to land in the right month. You can find that out in an afternoon instead, with a spreadsheet and an export from your CRM.

Below: the choice, the back-test, and the part nobody warns you about. Once commission depends on CRM data, your close dates and deal amounts become payroll inputs.

## Three structures are enough for five reps

Commission guides love catalogues: straight commission with no base, gross margin, residual, draw against commission, territory volume, multiplier, splits. Most of them exist for a specific problem, like reps who can’t live without a base yet, margins that swing deal to deal, or big territories. A team of three to fifteen people selling one or two products usually has none of those problems yet, so you can skip them for now. If you do take on a rep on straight commission, pay is half the question, and [what a commission-only rep should see in Pipedrive](https://salesbond.app/blog/commission-only-reps-pipedrive-access/) is the other half.

Here is what we’d actually consider.

Flat rate on base: reps get a salary plus a fixed percentage of every closed-won deal. It’s the easiest plan to explain, the easiest to calculate, and the hardest to argue about. If your deals are similar in size and price, this is where we’d start. Its weakness is that it pays the same rate for the deal that makes the quarter and the deal that comes after it, so it does little to push a rep who is already comfortable.

A tiered rate keeps the same base, but the rate steps up once a rep passes a threshold, usually quota. Below quota they earn the standard rate; on revenue above it, a higher one. That higher tier is what people call an accelerator. It only works if your reps hit quota often enough to see the tier as something they can reach. If nobody has cleared quota in a year an accelerator costs you nothing and motivates nobody.

The third option is base plus variable against a target. You start from on-target earnings (OTE: base plus variable at 100% of quota), and the variable half is paid out in proportion to quota attainment. We think it suits bigger sales orgs better, and it’s where plans start to need a spreadsheet to explain. We cover the OTE side in [how to set OTE for your first sales hire](https://salesbond.app/blog/ote-sales-first-hire/), and quota-setting in [how to set sales quotas](https://salesbond.app/blog/how-to-set-sales-quotas/), so we won’t repeat either here.

Our default for a five-person team on one product: flat rate on base. Move to a single accelerator tier once you have a year of data showing that at least some reps clear quota in most quarters. One tier. Three is too many.

## Back-test the plan: calculate what last year would have paid

A commission plan is a formula. You can run a formula on history before you run it on people. Here’s how we do it.

1. In Pipedrive, open the Deals list view and filter for Status = Won and Won time within the last four full quarters. Add the Owner, Value, Won time and Pipeline columns, then use “…” > Export filter results.
2. Remove anything that wouldn’t be commissionable under the new plan: renewals if you exclude them, internal test deals, deals owned by the manager.
3. Add a column for the quarter of each deal, based on Won time, not on expected close date.
4. Sum won revenue per rep per quarter. This is the table your plan will run against.
5. Apply the draft plan’s formula to each rep-quarter and add up the payout.
6. Compare three numbers: total payout as a percentage of won revenue, the spread between your highest-paid and lowest-paid full-year reps, and how much of the payout came from one or two deals.
7. Change one parameter (rate, quota, or a quarterly vs annual measurement period) and run it again.

Hypothetical example below. The numbers are illustrative, not from a real team.

The draft plan: 8% of closed-won revenue up to a quarterly quota of $120,000, and 12% on revenue above quota. Rep E joined in Q3.

![A table demonstrating the payout structure for a sample sales commission plan.](https://salesbond.app/wp-content/uploads/2026/09/sales-commission-structures-small-team-2.png)

| Rep | Q1 | Q2 | Q3 | Q4 | Won revenue | Above quota | Commission | % of won revenue |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| A | $150k | $140k | $160k | $130k | $580k | $100k | $50,400 | 8.7% |
| B | $110k | $125k | $100k | $115k | $450k | $5k | $36,200 | 8.0% |
| C | $60k | $210k | $70k | $80k | $420k | $90k | $37,200 | 8.9% |
| D | $90k | $95k | $85k | $100k | $370k | $0 | $29,600 | 8.0% |
| E | — | — | $40k | $70k | $110k | $0 | $8,800 | 8.0% |
| **Team** | | | | | **$1,930k** | **$195k** | **$162,200** | **8.4%** |

Total cost is 8.4% of won revenue, against 8.0% under a plain flat rate. The accelerator added $7,800 across the year. Is that worth it? Your call, but at least now it’s a number you can discuss.

Look at Rep C. They were below quota in three of four quarters, and yet they collected $3,600 of the accelerator money, almost as much as Rep A, who beat quota every single quarter. One large deal in Q2 did that. Rep B, steady and slightly below quota most of the time, got $200 of it.

Now switch the measurement to an annual quota of $480,000. Rep A is still $100,000 above it. Rep C, at $420,000, is below, and so is Rep B. The accelerator now pays only the rep who was consistently over. Same rates, different period, very different message to the team.

You want to see this before you announce the plan. After the first payout is too late.

## Where plans break: cliffs, splits, and deals that slip a quarter

Most commission arguments on small teams aren’t about the rate. They’re about edges.

A cliff is any point where one dollar of revenue changes the payout by a lot more than one dollar’s worth of commission. Pay a bonus for hitting 100% of quota and nothing at 99%, and you have built a cliff. Reps will notice it before you do. Deals get pulled into the quarter with discounts, or pushed out of it when the quarter is already lost, because a deal that won’t clear the cliff this quarter might help clear it next quarter. Tiered rates on the marginal revenue (like the example above) are gentler than bonuses on thresholds, because the rate only changes on the revenue above the line.

If you want to push one behavior for a few weeks, run [a short SPIFF instead](https://salesbond.app/blog/sales-spiffs-small-team-pipedrive/) of bending the commission plan, and check afterwards that it did not just pull deals forward.

Splits come up the first time two reps work the same deal. One sourced it, the other closed it, or a founder stepped in for the final call. If the plan doesn’t say how that is paid, the answer gets made up under pressure, and whoever loses the the argument remembers it. Write down a default split before it happens. Something like 50/50 between the owner and one named contributor, with the manager deciding exceptions in writing, is fine. The exact ratio matters less than having one. Recording that split in Pipedrive then takes [two fields instead of an argument](https://salesbond.app/blog/shared-deal-credit-pipedrive/).

Splits get more frequent when one rep opens accounts and another grows them, and [whether a small team needs both roles](https://salesbond.app/blog/hunter-vs-farmer-sales-small-team/) is a decision of its own.

Then there are deals that slip a quarter. Under a quarterly plan, a $40,000 deal that closes on April 2nd instead of March 30th can move a rep from below quota in Q1 to above it in Q2, and change the payout on both quarters. You can’t prevent slips; you can decide which date counts. We’d count the Won time Pipedrive records when the deal is marked Won, and nothing else. If your team has a habit of moving close dates around, [pushed close dates in Pipedrive](https://salesbond.app/blog/pushed-close-dates-pipedrive/) covers how to spot it.

Refunds and cancellations belong here too. If a customer cancels in the first month, does the commission come back? A common answer is a clawback: commission taken back if the customer cancels within an agreed window. Whatever you choose, put it in the plan, along with the length of that window. This plan covers your reps; a referral partner fee is a different agreement, paid only when the deal is won and the first invoice is paid, as in [referral partner deals in Pipedrive](https://salesbond.app/blog/referral-partner-deals-pipedrive/).

## The CRM fields your payout now depends on

Before a commission plan, a wrong close date in the CRM was a forecasting annoyance. After it, the same wrong date is a pay dispute.

In Pipedrive, four fields carry the weight. Value is the base of the formula, so the plan has to say whether it means the Value at [the moment the deal was won](https://salesbond.app/blog/pipedrive-won-not-paid-prepayment-invoice/) or the amount on the invoice, whether discounts, taxes, and one-off setup fees count, and, if deals come in more than one currency, [which currency and rate the plan uses](https://salesbond.app/blog/pipedrive-multiple-currencies-totals/). Won time decides the quarter. Pipedrive sets it when the deal is marked Won, and [only users with deals admin permission can edit it](https://salesbond.app/blog/pipedrive-won-time-edit-month-end/) afterwards. Owner decides who gets paid, so reassigning a deal in the last week of the quarter is really a payroll change. And Status itself: a deal marked Won that is later reopened or lost needs a rule, otherwise the same revenue can be paid twice or clawed back without warning.

So limit deals admin permission to the sales manager. On a five-person team that’s one person, and it’s the right one. Everyone else can request a change, and the request gets a line in the deal’s notes.

We’d also run a quick check before each payout: won deals with no Value, won deals with Won time outside the period, deals that changed Owner in the last two weeks. If you already track a handful of numbers weekly, those checks fit next to them; the [sales manager dashboard with six numbers](https://salesbond.app/blog/sales-manager-dashboard-six-numbers/) is a reasonable place to start.

Nobody on the team is being accused of anything here. If there’s an argument, you want it to be about the rule, and a rule can be read. Memories of a call can’t.

## Rollout: what to tell the team, and when to look again

Show the back-test to the team. Skip the rep-by-rep table with everyone’s names on it and share the shape of it: here is what last year would have paid under the new plan, here is how it compares to the old one, here is who it favors and why. Reps trust a plan more when they can check the arithmetic on thier own deals.

Put the plan in writing, one page. Rate, quota and measurement period, what counts as commissionable revenue, which date and value decide the payout, how splits and clawbacks work, when payouts happen, and who can edit the fields that feed them. If a sentence in the plan starts with “usually”, rewrite it.

Start it at the beginning of a quarter, never in the middle. A plan that changes mid-quarter turns every open deal into a negotiation about which rules apply.

Then put a date in the calendar to look again. We’d rerun the back-test after two quarters on the new plan, using the real data this time, and compare it with what the team expected. If payout as a percentage of revenue has drifted well away from what you planned, or one rep’s pay depends mostly on a single deal, adjust the rate or the period at the next quarter boundary. Small teams change fast. A plan that fit four reps selling one product may not fit seven reps selling two, and it’s easier to adjust a plan everyone understands than to defend one nobody does.