From The Work
Scaling Creative Operations at Yelp: The Systems That Made It Possible
2 → 30 People | 100+ Campaigns/Year | 25% Brand Visibility Increase


The results are the easy part to tell. A 2-person creative team scaled to 30. More than 100 integrated campaigns a year. A 20% reduction in delivery time, a 25% improvement in efficiency, a 20% lift in conversion rates across multiple verticals. Those numbers are real and they matter.
What's harder to explain is the unglamorous work that made them possible. Not the campaigns — the infrastructure underneath the campaigns. The systems that had to exist before the team could grow, before the work could scale, before any of those results could happen. That's what this piece is about.

The Problem with Scaling Creative Teams
Most organizations solve a capacity problem by adding bodies. The team is behind, so you hire. The team is still behind, so you hire again. Headcount goes up and output improves — but the improvement is linear, and eventually you hit the ceiling again. The real problem was never the number of people. It was the absence of shared infrastructure.
When I joined Yelp's creative team there were two of us. Every campaign was built from scratch. There was no standard brief format, so every project started with a different set of assumptions about scope, timeline, and success criteria. There was no intake process, so requests arrived through whoever happened to be available. There was no visibility into capacity, so nobody could plan more than a week ahead without someone's schedule colliding with someone else's. There was no documentation of previous campaigns, so every retro was anecdotal rather than systemic.
This was not a resourcing problem. It was a systems problem. Adding people to it would have added chaos proportionally.

The Order of Operations
The first principle I applied was sequencing: build the system before you scale the team. Every person you add to a broken process makes the process harder to fix. So for the first several months the work was almost entirely infrastructure.
The intake system came first because it was the earliest point of failure. A single front door for every creative request: a standardized brief format that required the requester to answer the questions a creative team actually needs answered before work can begin. What's the business objective. Who's the audience. What's the one thing we want someone to feel or do. What does success look like. What's the timeline and who's approving.
This sounds simple and it is. But the discipline of completing a brief well forces a kind of thinking that most stakeholders skip when they're sending a Slack message. Brief quality improved immediately. The back-and-forth before kickoff dropped by more than half. Projects started with alignment rather than building toward it.
The resource infrastructure came second. A scheduling system that made capacity visible in real time — who was committed, who was available, what was due when. Before this existed, capacity was a guess. Someone would accept a project and discover three days later that their calendar was already full. With a live view of the team's capacity, planning became possible. You could look two weeks ahead and know where the pressure points were before they became emergencies.
The asset architecture came third. A CMS structure that made existing work findable, version-controlled, and connected to the briefs that produced it. Templates for the repeatable formats — ad units, social creative, email headers — so production didn't start from scratch every time. A brand standards system that could be handed to a new vendor on a Monday and produce on-brand work by Friday without a lengthy onboarding process.
Finally, the retro infrastructure. Every campaign cycle closed with a documented retrospective that fed directly into the next brief. What worked. What didn't. What would we change. These weren't retrospective conversations — they were structured inputs into a living document that made every subsequent campaign smarter than the previous one.

What Hiring Into a System Looks Like
Once the infrastructure existed, hiring changed entirely. Each new person joined a process they could learn rather than chaos they had to navigate. Onboarding had a defined path. Expectations were documented, not passed down verbally. A new designer could reach an independent output level in weeks rather than months because the system carried them through the learning curve.
The team grew from 2 to 8 to 15 to 30 over three years. At every stage, the infrastructure scaled with the headcount rather than lagging behind it. That's the compounding effect of building the system first: the returns increase as the team grows rather than diminishing.
The Lesson That Travels
Every creative organization I've worked with since has had a version of the same problem. Not always at Yelp's scale, but the same shape: talented people working without the infrastructure that would make them collectively more effective than they are individually. The solution is almost always the same sequence. Map the friction first. Build the intake, the scheduling, the documentation, the feedback loops. Then scale the team into the system rather than building the system around a team that's already too large to change.
The glamorous version of this story is the campaigns — the ones that ended up in the New York Times, the Awwwards wins, the 30 million impressions. Those are what the system produced. But the system is the actual work.

Key Takeaways
What Made This Work
Systems > Bodies
Didn't just hire more people—built infrastructure that made each person 10x more effective.
Creative as Partner, Not Service Org
Embedded early in planning cycles. Stakeholders trusted us to deliver.
Measured Everything
Connected creative work to business results. This earned budget and headcount.
You can't scale creative by just adding bodies. You scale by building systems, workflows, and measurement that let talented people do their best work consistently.
When I joined, Yelp's creative team was underwater. When I left this role, it was a well-oiled machine that stakeholders trusted—and that's what enabled the business to move faster.


