We Went Looking for the Standard Partner Commission. We Found Two Operating Systems.

You are building next year's partner program.

Someone asks what commission rate you plan to offer.

You say 20%.

The next question is predictable:

"Is that competitive?"

Now you need a benchmark.

You search for the standard partner commission rate. You find affiliate-program roundups, vendor pages, and reports with clean-looking averages. Most of the numbers land somewhere between 20% and 30%.

That feels like enough evidence to make a decision.

It isn't.

Because before asking what the average rate is, there is a more important question:

How many partner programs publish a rate at all?

That is where the story changes.

We went looking for the market rate

BlueThread Research currently tracks 200 partner programs across 762 public sources.

We were looking for the same things any operator would want while building or benchmarking a program:

• What do partners get paid?
• How long do they get paid?
• What happens when they identify an opportunity?
• How much protection does a registered deal receive?
• What support exists after the introduction?

At first, this looked like a research problem. Find enough program pages, normalize the language, and calculate the midpoint.

The midpoint was easy enough to find.

Among the 42 programs that publish a numeric commission, the median headline rate is 22.5%.

That is a useful number.

It just was not the most interesting thing in the data.

The more interesting finding was how few programs allowed us to calculate it.

Only 51 of the 200 programs disclose any public revenue-share terms. Another 26 indicate that revenue share exists but keep the terms behind a partner gate.

The other 123 are publicly silent.

Not necessarily because they pay nothing.

Because they do not tell you what they pay until you are inside the program, and sometimes not until you have negotiated it.

The missing information was not random. It had a pattern.

Then company size changed the meaning of the number

We grouped the programs by company size.

Among companies with up to 1,000 employees, 34% of programs publicly disclose revenue share. The median published commission is 25%.

But only 5% disclose deal-registration mechanics, and 14% disclose any public co-sell mechanics.

The program is telling the partner what they can earn.

It is saying much less about what happens after the partner identifies an opportunity.

The pattern changes at companies with 1,001 to 10,000 employees.

Public revenue-share disclosure falls to 23%. The median published commission falls to 20%.

At the same time, deal-registration disclosure rises to 32%. Public co-sell disclosure rises to 37%.

The economics become less visible.

The operating system becomes more visible.

At companies with more than 10,000 employees, the split becomes extreme.

Only 1 of 38 programs publishes a numeric revenue-share rate.

But 11 disclose deal registration, and 11 disclose some form of co-sell mechanics.

There is no credible public commission benchmark for that largest band. One published number is not a market.

That absence is not a research footnote.

It is the finding.

As the company gets larger, the public rate does not simply decline. It disappears from view.

What replaces it is the machinery required to govern a larger ecosystem: qualification, tiers, registered opportunities, account alignment, certifications, and formal co-sell processes.

The size bands revealed two different operating systems.

Smaller vendors show the economics and hide the machinery.

Larger vendors show the machinery and gate the economics.

The machinery has a name

Co-sell is part of that machinery.

Across the mature-channel cohort, 40 of 130 programs disclose at least one public co-sell fact.

Among 44 emerging programs, only six do.

Deal registration makes the difference even clearer.

Thirty-six mature programs publicly describe deal registration.

Only one emerging program does.

That changed how I read the commission data.

The emerging vendor is often saying:

Bring us a customer and we will pay you.

The mature program is more likely to say:

Register the opportunity. Meet the qualification rules. Work through the account team. Maintain your tier. Then we will determine what protection and economics apply.

One is recruiting participation.

The other is governing participation.

This is how a mature program can become more sophisticated and less transparent at the same time.

It tells partners more about how to operate inside the system while telling them less about what the transaction will pay.

Services complicated the story

I expected services to follow the same divide.

Maybe emerging programs were mostly lightweight affiliate and referral motions, while mature programs had the real services ecosystems.

That is not what the data showed.

BlueThread Research records a service-delivery motion in 22 of the 44 programs at companies with up to 1,000 employees.

It records one in 55 of the 84 programs at companies with 1,001 to 10,000 employees.

Services exist on both sides of the maturity curve.

Both groups need partners to help customers implement, configure, manage, and get value from the product.

The difference is how that work is organized.

Smaller vendors tend to recruit agencies, experts, creators, and solution partners. They need someone close to the customer who can make the product work.

Larger vendors are more likely to add MSP paths, systems-integrator tracks, certifications, competencies, deal registration, and co-sell rules around that work.

The service may look similar to the customer.

The operating model behind it is not.

The smaller vendor is building a network of people who can help.

The mature vendor is building a system that decides who is qualified to help, what they may sell, which opportunities they can claim, and how their performance will be measured.

Both need delivery.

One is still recruiting the network.

The other is governing it.

The 22.5% was hiding a bigger story

We started with a simple question:

What is the standard partner commission?

The data gave us an answer: 22.5% among the programs that publish a numeric rate.

But by then, the number meant something different.

A 25% commission inside a lightweight referral program is not equivalent to a 20% revenue share inside a program with certifications, registered-opportunity protection, account alignment, and a formal services path.

The percentage tells you what the transaction may pay.

It does not tell you what kind of partner the vendor is trying to create.

It does not tell you how the partner gets into an account.

It does not tell you who owns the opportunity.

It does not tell you whether the partner is expected to introduce, resell, implement, manage, or retain the customer.

And it does not tell you whether the vendor has built the infrastructure to support any of it.

That became the more useful finding.

Emerging vendors tend to show the economics and hide the machinery.

Mature programs tend to show the machinery and gate the economics.

Services sit in the middle. Both cohorts need partners to deliver outcomes. Mature programs are simply more likely to build a visible operating system around that delivery.

So there may not be one standard partner program to benchmark against.

There are at least two.

One competes for partners with economics.

The other coordinates partners through governance.

One says:

Here is what we will pay you.

The other says:

Here is how you earn the right to participate.

If you are designing a program, the first decision is not whether the commission should be 20%, 22.5%, or 25%.

It is which operating system you are actually building.

Methodology note: Figures come from the BlueThread Research Program Intel corpus as of September 20, 2026. The corpus contains 200 partner programs supported by 762 cited public sources. It is a curated operating dataset, not a random sample of every partner program. Company-size comparisons exclude programs without a known employee band. "Not disclosed" means no public value was found in the recorded evidence. It does not mean the underlying benefit or mechanism does not exist.