How to Calculate Sales Commission: Formulas, Examples, and the Mistakes That Cost You

The base formula

Start with the one everyone knows:

commission = commission base × commission rate

A $10,000 deal at 10% pays $1,000. Nobody disputes that. Disputes start one layer down, at the three questions the formula quietly assumes you have answered.

Question one: what is the base?

There are three defensible answers, and they suit different situations.

Gross revenue

The full contract value. Simple, transparent, and easy for a salesperson to predict — which matters more than it sounds, because a commission model nobody can calculate in their head does not motivate anyone.

The weakness: it is indifferent to cost. A deal sold at a heavy discount pays the same percentage as one sold at full price, so if the salesperson controls pricing, gross revenue quietly rewards discounting.

Net revenue

Revenue after pass-through costs. For an agency, this usually means excluding media spend that flows straight to Meta or Google. If a client pays $50,000 and $35,000 of that is ad spend, the agency’s net revenue is $15,000, and commissioning the full $50,000 would pay out on money the agency never kept.

For any agency that bills media spend through its own accounts, this is usually the right base.

Gross profit

Revenue minus delivery cost. Correct in theory, and the right answer when the salesperson sets both price and scope, because it makes them care about margin rather than volume.

The practical objection is transparency: paying on profit means the salesperson must be able to see the cost figures, or the payout is unverifiable. Plenty of agencies are unwilling to expose delivery costs to the sales team, and then choose gross profit anyway. That combination produces exactly the disputes you would expect.

Question two: when is it earned?

This is the question that decides whether commission ever has to be taken back. Three common triggers, in increasing order of safety for the agency:

• On close. The deal is marked won. Fast and motivating, and it pays out on revenue that has not arrived and might not.
• On invoice paid. The client has actually paid. Slower for the salesperson, dramatically safer for the agency, and it removes most clawback situations before they can happen.
• On cash cleared, after the refund window. The safest, and the least motivating. Reasonable where refunds are common; overkill where they are not.

Most agencies land on invoice paid, and it is usually the right call. It aligns the payout with the money without pushing the reward so far from the work that it stops functioning as an incentive.

Question three: what happens when money comes back

A refund, a chargeback, or a client who cancels in month two of a twelve-month contract. If commission was paid on close, that money is already gone.

A workable clawback rule needs three parts, all agreed in advance:

• A window. Commission is recoverable if the refund happens within, say, 90 days. After that it is the agency’s risk.
• A method. Deducted from the next commission run, not invoiced back. Invoicing a salesperson for returned commission is a good way to lose them.
• A cap. No single run goes negative. If the clawback exceeds what is owed, it carries forward.

Worked examples

Flat rate

The simplest model: one rate on all closed revenue. At 10%, a closer who books $40,000 in a month earns $4,000. Easy to run, easy to predict, and it treats the first deal of the month exactly like the twentieth.

Tiered

Rates rise as cumulative volume passes thresholds. Say 8% on the first $20,000, 12% from $20,001 to $50,000, and 15% above that.

A closer at $65,000 for the month:

• First $20,000 × 8% = $1,600
• Next $30,000 × 12% = $3,600
• Final $15,000 × 15% = $2,250
• Total = $7,450

The mistake to avoid: applying 15% to the whole $65,000, which gives $9,750. Each rate applies only to the revenue inside its own band. This single error is the most common commission miscalculation there is, and it is expensive in both directions — it overpays if the agency makes it, and it destroys trust if the salesperson catches it.

Setter and closer split

Where one person books the meeting and another closes it, define a single pool first, then split it.

On a $12,000 deal with a 10% pool, the pool is $1,200. At a 30/70 split, the setter earns $360 and the closer $840. Total commission cost stays 10% of the deal regardless of how many people touched it — which is the entire point of pooling.

The alternative, giving each role its own independent percentage, means commission cost per deal changes depending on who was involved. That makes forecasting harder for no benefit.

Recurring revenue

Retainers need a decision about how long commission continues. Three common shapes:

• First month only. A $3,000/month retainer at 10% pays $300, once. Cheap, and it gives the closer no reason to care whether the client stays.
• Fixed term. The same 10% for twelve months, so $3,600 total, paid monthly as the client pays. Ties the payout to retention without creating a permanent liability.
• Declining. 10% for the first six months, 5% thereafter while the client remains. Rewards long-term accounts and tapers the cost.

Five mistakes worth avoiding

1. Applying a tier rate to the whole amount. Covered above, and worth re-checking in whatever spreadsheet you currently use.
2. Leaving “closed” undefined. Signed contract, first payment, or verbal agreement? Write it down, because everyone assumes the definition that favours them.
3. Commissioning pass-through media spend. Paying 10% on $35,000 of ad spend the agency never keeps is a $3,500 mistake per deal.
4. No clawback rule at all. The rule is easy to agree before the first refund and nearly impossible after it.
5. Running it in a spreadsheet only one person understands. The calculation stops being auditable, and when that person is on holiday, commission is late.

Why the spreadsheet eventually breaks

Almost every agency starts with a commission spreadsheet, and it works fine for a while. It breaks for a structural reason rather than a scale one: the spreadsheet does not know when a deal closes or when an invoice is paid, so a human keeps it in sync. That human is the single point of failure, and every disagreement becomes an archaeology exercise across the CRM, the invoicing tool and the sheet.

The fix is not a better spreadsheet. It is having commission calculated from the same records that hold the deal and the invoice, so the number is derived rather than maintained.

That is how Openbiznis handles it: the deal, the invoice and the commission are the same data, so marking a deal won produces the commission entry, and payment status drives when it is earned. Nobody reconciles anything, and a disputed figure can be traced back to the call that produced it. You can see how the finance and sales roles see it differently, or what it costs.