I Fired My Agency. Here's What Happened Next.

The first agency talked a great game. Beautiful deck, impressive client logos, a founder who spoke in frameworks. Three months in, the reports arrived like clockwork: impressions up, engagement up, reach up. Revenue flat. When I asked why, I got a new deck explaining that brand building takes time.

The second agency was cheaper and worse. Half the retainer, a quarter of the thinking. I briefed them on everything. I wrote the creative direction. I basically hired an employee who worked somewhere else, answered my Slacks a day late, and billed me monthly for the privilege.

The third one actually knew paid social. Smart people, decent media buying, real dashboards. But they needed so much hand-holding that I spent more time managing them than I would have spent doing it myself. That was the moment I did the math, and the math ended the relationship.

Signs your marketing agency isn't working

Founders ask me how to know when it is time. It is not one bad month. Every agency has a bad month. It is a pattern, and the pattern has three tells.

Tell one: the reports measure the wrong things. Impressions, reach, engagement rate, click-through rate. These are activity metrics. They tell you the agency did things. They do not tell you whether the business grew. A good report leads with revenue, contribution margin and payback period. If your agency's report would look identical for a brand that grew 50% and a brand that shrank 20%, the report is decoration.

Tell two: you are doing their job. You write the briefs. You come up with the angles. You chase them for the deliverables. You explain your customer to people you pay to understand your customer. An agency should reduce your workload. If your calendar got fuller after hiring them, the arrangement is backwards.

Tell three: nothing compounds. Each month starts from zero. New campaigns, new creative, new tests, and the learnings evaporate. There is no growing asset, no system getting smarter, no email list getting bigger in a way that pays off next quarter. You are renting activity by the month instead of building an engine.

If all three are true, it is not a rough patch. It is the model working as designed.

Why agency retainers misalign incentives

This is the part that is not the agency's fault, exactly. It is structural. A retainer pays a fixed fee for a bundle of activity: so many posts, so many campaigns, so many reports. The agency's incentive is to deliver the activity at the lowest cost that keeps you paying. Your incentive is revenue growth. Those are different goals wearing the same contract.

Think about what the retainer rewards. It rewards busyness. More campaigns look like more value. It rewards reporting, because thick reports justify the fee. It rewards retention of the client, not growth of the client, because a client who grows fast eventually asks why they are paying a percentage or a flat fee for work that now runs itself.

The deepest misalignment is around speed. Your business needs fast iteration: test, learn, compound. But fast iteration is expensive for an agency. It means senior people, quick turnarounds, constant creative production. The retainer economics push toward the opposite: junior staff, templated work, monthly cycles. You pay for a growth partner and receive a content subscription.

None of this means agencies are useless. A great agency, tightly managed, with incentives tied to outcomes, can work. But that describes almost no retainer I have ever seen signed. The default outcome of the default contract is the default result: motion without movement.

Ecommerce agency vs in-house vs a system

When founders fire their agency, they usually swing to one of two options. Bring it in-house, or hire a freelancer. Both can work, and both have the same failure mode: you replace the agency bottleneck with a hiring bottleneck. Now you are recruiting, onboarding, managing and retaining, which is a full-time job you did not apply for.

There is a third option that most founders miss because it did not exist until recently. Instead of renting people by the month or hiring them by the year, you install a system that runs the work continuously. A specialist SI bot team is software specialists led by a chief of staff bot, shaped around your business, running growth, audience, customers, operations and numbers around the clock. It drafts, you approve what goes out, and every task is measured against revenue because the system has no incentive to look busy. It only has the incentive you give it.

The difference is structural, not motivational. An agency optimises for the retainer. An in-house team optimises for keeping their jobs, which is better but still not the same as your P&L. A bot team optimises for whatever you tell it to optimise for, and you told it revenue. The playbook behind SIBT recorded 518% average sprint growth across client sprints. One brand reached 27x ROAS in 14 days. Those are system outputs, not heroics.

What actually changed after I fired the agency

The first change was speed. Decisions that took a week of back-and-forth took an hour. Creative that waited for a monthly content calendar shipped daily. Tests that needed a statement of work ran continuously. The growth loop went from a monthly cycle to a daily one, and the compounding started almost immediately.

The second change was honesty. No more decks. No more impression reports. Every task showed its cost before it ran and its result after. When something did not work, we knew in days, not at the quarterly review. Failure got cheap, which meant testing got aggressive, which is where growth actually comes from.

The third change was the quietest and the biggest. I stopped thinking about marketing as something that happens to my business and started thinking about it as something my business does. An agency is a vendor relationship. A system is an asset. Assets compound. Vendors invoice.

What to do in the 30 days after you fire them

The fear that keeps founders in bad retainers is the gap. If I fire the agency, who runs the ads next Tuesday? The answer is to sequence the exit, not to leap.

Week one, you take inventory. Every login, every ad account, every pixel, every audience, every piece of creative. Agencies hold more of your infrastructure than you realise, and access transfers are where exits get messy. Get admin on everything before the termination email goes out, not after.

Week two, you stabilise. Turn the best-performing campaigns into evergreen holds. Pause the experiments that were really just activity. Cut the spend that exists to justify the retainer rather than to make money. Most brands discover that 70% of their ad account was noise the moment someone looks at it with fresh eyes.

Week three, you install the system. This is where the sequenced growth system earns its keep: offer, conversion, email, content, ads, outreach, in that order. You fix what is broken before you scale what works. The brands that do this usually find their first win in email, because it is the channel the agency neglected while chasing new customer metrics for the report.

Week four, you measure honestly for the first time in months. Revenue, contribution margin, payback. No impressions. The numbers are sometimes uncomfortable, but they are yours, and they move when you work them. That is the whole point.

If you are sitting in the agency cycle right now, thick reports and thin results, you already know what you need to do. The question is what replaces the retainer. Here is the system that replaced mine.