I did in six hours what our marketing agency needed two months for. So we fired them.
Not out of spite. The math made the call for me.
Three months on retainer. $42,000 for a two-month campaign. Eight rounds of feedback and twelve meetings just to get to launch day. Then the campaign went live, broke in seven different ways, and cost us close to $100,000 before anyone could ship a fix.
Plenty of agencies are good, and later on I’ll get into exactly when hiring one still beats building in-house. What follows is the real numbers from one specific relationship, what it took to unwind it, and a framework for making this call before it costs you six figures instead of six hours.
The short version
A company I worked with paid an agency $42,000 over two months to launch a multichannel campaign: ads, landing pages, copy, visuals. Planning took 12 meetings and 8 rounds of feedback. After launch, a run of preventable mistakes cost roughly $100,000 in lost monthly revenue. I rebuilt the broken pieces myself in six hours. Moving the work in-house afterward saved $4,250 a month, cut planning-to-launch time in half, and lifted campaign ROI by around 20%.
The setup: three months, $42,000, and a rocky start
The agency had already been on retainer for three months when this campaign started. The brief wasn’t complicated: a multichannel push covering ads, landing pages, copy, and visuals, budgeted at $42,000 for a two-month sprint from planning to launch.
Planning was rocky from the first call. We kept having to re-explain who our product was for, well into month three of paying them. An agency still asking “wait, who’s this for again?” in its fourth month isn’t in an onboarding phase anymore. It never built a real model of the business in the first place.
It took 8 rounds of feedback to get the materials shippable and 12 meetings to reach go-live. Turns out that’s not just us. Setup®, which has tracked client-agency breakups for eight years, found that in 2025 “the agency didn’t understand our business” was the fourth most common reason clients left, cited by 44% of them, right behind dissatisfaction with delivery and value, both at 61% (Setup®). We were living that stat and didn’t have a name for it yet.
What went wrong after go-live
Within days, the cracks weren’t cosmetic:
An ad went out with grammar errors. The CEO caught it and forwarded a screenshot.
The URL on our most important landing page pointed to the wrong place.
Some ads got served to existing clients instead of the new-prospect audience they were built for.
Landing pages carried images that had never gone through approval.
The A/B tests, the entire point of a controlled launch, weren’t set up correctly. No clean read on what was working.
An influencer video posted early, out of sync with the rest of the launch.
Our client success manager was out of office through all of it and hadn’t told us.
Three days to fix everything. By the time it was patched, we were staring at a $100,000 hit to monthly revenue, more than double what the whole two-month campaign had cost to make.
None of these were hard problems. They were carelessness. Nobody on the other side was treating our launch like it was theirs.
The real number isn’t the $4,250 a month
This story leads with the monthly savings: $4,250 once we moved in-house, about 20% more ROI, planning-to-launch time cut in half. Real numbers, and a clean headline.
But they weren’t what mattered most in the decision. The retainer savings were a rounding error next to the $100K we lost from three days of a broken campaign. Most agency-vs-in-house comparisons price the relationship by the invoice. The real cost shows up later, when something breaks and nobody senior is paying close enough attention to catch it before it ships.
That tracks with a wider pattern. A 2026 analysis of the CMO-agency relationship found dissatisfaction with delivery is the #1 reason clients leave (48%), while agencies rank it 7th on their own list of concerns. Six positions apart, same relationship. Agencies keep bracing for the budget conversation. Clients are walking over a broken URL nobody checked (Everything-PR).
The same research puts retainer churn at 18% a year, project-based churn at 42%, and the highest risk of all in the first 90 days. We were three months in. Right in that window, not past it.
Why this happens more than founders realize
I started my career at an agency. The irony isn’t lost on me: I’ve now fired more agencies than I’ve hired.
But this wasn’t one bad vendor. It’s structural. Full-service retainers typically run $3,500 to $10,000 a month, with most agencies still enforcing minimums around $1,000 to $1,500 even for smaller scopes (HawkSEM). At that price, a small account rarely gets a senior team’s full attention. It gets whoever has bandwidth that sprint. Agencies also routinely spend $400,000 or more just pitching for new business (Everything-PR), a cost structure built around landing new logos, not protecting the ones already signed.
Agencies aren’t inherently bad. The incentives on a mid-size retainer just don’t reward the kind of obsessive, would-I-ship-this-on-my-own-product attention a founder assumes they’re paying for.
Small business marketing agency vs. in-house: when the agency wins
To be fair to the other side: outsourcing execution isn’t automatically a mistake, and the data doesn’t back a blanket “always build in-house” rule either.
Companies with serious in-house marketing muscle still keep agencies around. Among ANA member companies running both, in-house handles 61% of the work on average, but 92% still use outside agencies for the rest (Marketing Charts), usually for specialized skills that don’t justify a full-time hire, or overflow during a launch push. The shift toward in-house is real (82% of ANA members have an in-house agency now, up from 58% a decade ago). It’s rebalancing who does the work more than eliminating agencies from the picture.
A 200-person company running four campaigns a quarter across a dozen channels can get outproduced fast trying to do it all internally. What happened to us was a specific, common failure: a small account, understaffed on their side, running something complex enough to need senior attention it never got.
What changed when I brought it in-house
The six-hour rebuild wasn’t heroic. Most of it was AI doing in minutes what used to sit in a design or dev queue for days: rewriting the broken ad copy, regenerating the landing page assets, fixing the tracking and A/B test setup, checking every link and image against our approved brand set.
That’s not unique to us. Traditional production of one marketing visual used to take several hours. AI-assisted teams now do it in under 30 minutes. Full campaign launch windows have gone from a typical 2–3 weeks down to under 2 days (Luma), with teams reporting roughly 10x more creative output without adding headcount. Meanwhile, an estimated 52% of content made through traditional agency pipelines never even reaches the market.
Six hours stops looking like an outlier once you put those numbers next to each other. It’s what happens when the gap between an outsourced production queue and one person with AI-native workflows finally flips.
A simple framework for the decision
Three questions do most of the work if you’re making this call yourself.
How many people do you have? Under roughly 20, an agency retainer usually buys less senior attention than you’re picturing, at a price that could fund an in-house hire plus a solid AI toolkit with room to spare.
How complex is the launch? A single-channel, evergreen campaign is a reasonable thing to hand off. A multichannel launch with tracking, segmentation, and a hard go-live date needs someone who treats a wrong URL like their own revenue, because it is.
Can you build or check the work yourself? This is the newest variable, and most agency-vs-in-house comparisons haven’t caught up to it. If a non-technical founder can use AI tools to produce and QA campaign assets directly, the old argument for outsourcing execution gets a lot weaker. What’s left to pay for is judgment and strategy, a smaller, more defensible thing to hand to someone else.
Overpromising and underdelivering is common in this business. It’s just rarely said out loud. If a vendor keeps needing you to re-explain your own company three months in, take that as the whole story, told early.
Frequently asked questions
Is a small business marketing agency worth it, or should you build in-house? Depends on scope, but the comparison usually isn’t close once you factor in failure costs. A full-service retainer runs $3,500–$10,000 a month. One in-house marketer with modern AI tools can often match or beat agency output on execution-heavy work for a fraction of that, without the risk of a launch-day mistake going unnoticed until it’s expensive.
What’s the biggest reason companies fire their marketing agency? Delivery, not price. Multi-year survey data has dissatisfaction with delivery and dissatisfaction with value tied as the top reasons clients left in 2025, both cited by 61% of them, followed by “the agency didn’t understand our business” at 44%. Agencies tend to overestimate how much budget cuts drive departures and underestimate how much delivery does.
When does it make sense to keep working with a marketing agency? When the work needs specialized skills you don’t have and won’t need often enough to justify a hire, or when you’re running enough campaigns at once that even a strong internal team needs overflow help. Even companies with mature in-house teams keep agencies for exactly this. The goal is options: an agency for specialized or overflow work, nothing mission-critical resting on just one vendor.
How long should you give an agency before deciding it’s not working? Watch the first 90 days closely, that’s statistically when agency-client relationships are most likely to fail. If you’re still re-explaining fundamentals about your own company three months in, that’s a signal worth acting on, not waiting out.
Can AI tools really replace agency execution work for a non-technical founder? For production and QA-heavy work, increasingly yes. AI-assisted workflows have cut typical campaign production windows from 2–3 weeks to under 2 days industry-wide. What AI doesn’t replace is strategic judgment, so the more defensible use of an agency going forward is strategy and specialized expertise, not execution you can verify and rebuild yourself.