What is a superlinear agency?
The short answer
A superlinear agency is one whose output grows faster than its headcount. Traditional agencies are sublinear: every client they add dilutes the attention yours gets, because they scale by hiring juniors. A superlinear agency scales by building systems that compound, so senior judgment stays on every account while capacity grows.
A superlinear agency is an agency whose output grows faster than its headcount… it scales by building compounding systems rather than hiring juniors. Here’s the longer version.
Where the term comes from
Superlinear is a math word. A function is superlinear when output grows faster than input… double the input, more than double the output. Sublinear is the opposite: double the input, less than double the output.
Agencies, as a business model, are sublinear. The pitch meeting has the partners. The kickoff has the seniors. Month three has whoever they hired last quarter. It’s not malice, it’s arithmetic: agencies grow by adding clients, clients require staff, and the staff that’s available to add is junior. Every new logo dilutes the attention every existing client gets.
The other failure mode arrived more recently: the “AI-powered” agency, which flipped the problem. Automation with nobody senior watching. Faster output, same slop as their other hundred clients, because it came from the same prompts.
| Model | How it scales | What happens to your account |
|---|---|---|
| Traditional agency | Adds clients, staffs juniors | Attention dilutes; month three has whoever they hired last quarter |
| “AI-powered” agency | Automates output, removes oversight | Faster slop from the same prompts as their other hundred clients |
| Superlinear agency | Builds compounding systems, keeps only seniors | Senior judgment on everything, and the systems improve every account they touch |
A superlinear agency is the third thing. It scales the way software scales: by building systems that do the repetitive work… monitoring, analysis, reporting, production drafts… so the humans it keeps are all senior, and their judgment covers more ground every year. The systems compound. Every account benefits from every system the agency has ever built.
How to tell if an agency is actually superlinear
Most agencies now claim some version of “AI-enabled.” The test isn’t the claim, it’s the structure:
- Who touches your account in month six? If the answer is a different, more junior person than month one, the agency is sublinear regardless of its tooling.
- Do the systems run the agency’s own business first? Tools bought off a shelf help a person type faster. Systems built in-house against real workflows remove the work entirely.
- Does output get refined by a human before it ships? AI output is a starting block. If it goes out untouched, you’re buying slop with extra steps.
- Is the roster small on purpose? Superlinear economics let an agency make more money from fewer, longer-tenured clients. A roster that grows faster than the team is the old model wearing new words.
Why we call it that
The math term is old… Paul Graham wrote a whole essay on superlinear returns. Nobody had applied it to agency structure, so we did. We ran the sublinear model for years and built our way out of it: monitors that check every dollar of client spend daily, analysis that reads every search term and creative weekly, reporting that drafts itself for a human to refine. Somewhere along the way the label “agency” stopped describing the structure, so we started calling the structure what it is.
Whether other shops adopt the term or argue with it, the underlying question is the one worth asking any agency: when you grow, does my service get better or worse? For most, the honest answer is worse. The whole point of the superlinear model is that the honest answer becomes better.
More on how we run it: the Rise model and the systems themselves.