When to Fire Your Marketing Agency: 7 Signals Owners Miss Until the Number Stalls

By Avi Vatsa, CEO, Exchange Four Agency Last updated: September 6, 2026

Fire your marketing agency when it cannot connect its work to revenue, cannot explain a falling number, or has quietly moved you onto junior staff. Fire it slowly — after a written 30-day accountability test. Most owners fire too late, on frustration rather than evidence, and then repeat the same hire six months later.

What's the fastest way to tell if the problem is the agency or the market?

Ask one question: can they tell you what happened to the number, and why? An agency that owns the outcome will bring you the bad month before you notice it, with a diagnosis attached. An agency that owns a scope will bring you activity — posts shipped, impressions up, "engagement trending well" — and wait for you to raise the revenue question.

That distinction does most of the diagnostic work. The market is genuinely harder in some quarters; leads do get more expensive. But a partner who is actually running your system knows which stage broke — traffic, conversion, booked calls, close rate — because they instrument all four. A vendor knows only the stage it was hired for, and everything upstream or downstream is somebody else's problem.

We run this test in every takeover engagement. The first thing we ask a new owner-led client for is not access to their ad accounts — it's the last six months of reporting from the incumbent. In most cases, we can find every metric except the one that matters: what a dollar in produced in dollars out.

That gap isn't unusual. An industry survey distributed via Businesswire found 71% of brands are frustrated with their ability to demonstrate marketing ROI effectiveness — a majority of the market is paying for work it cannot evaluate. If you're in that 71%, you don't have an agency problem yet. You have a measurement problem, and you can't make a firing decision until it's fixed.

"They know exactly how to connect marketing execution to real business outcomes." — Riggs Eckleberry, Chairman, OriginClear

That is the whole bar. Execution connected to outcome. Everything below is a signal that the connection has been severed.

---

Signal 1: The reporting has stopped tying to a dollar

If your monthly report leads with impressions, reach, follower growth, or "engagement rate," and revenue appears nowhere or appears as an afterthought, you are being reported to rather than reported on. This is the single most common failure, and it's not always malice — many agencies genuinely cannot see past the last click they own.

What to check this week: open the last three reports side by side. Can you trace a single closed customer back through booked call → qualified lead → source? If the chain breaks anywhere, the reporting is decoration.

Vanity dashboards are a specific, diagnosable defect, not a vibe. We've broken down why most marketing reporting doesn't prove anything and how to calculate marketing ROI when the agency dashboard never ties to a dollar — run those before you run a termination conversation.

Signal 2: Your project manager has changed twice in a year

Account-team churn is the most reliable early warning available to you, and it shows up in the public record. In G2 and Clutch reviews of underperforming agencies, the same pattern repeats verbatim: "constantly changing project managers and the issues with communication made it difficult to work with them."

Every handover resets institutional knowledge about your business — your margins, your seasonality, your good-fit customer, the thing you tried in 2024 that failed. You pay for that reset in months. Two PM changes in twelve months means the seniority you bought at signing is not the seniority delivering the work.

What to check: who was on your kickoff call, and who is on your calls now? Name for name. If none of the original names remain, ask directly who owns your account internally and what their tenure is.

Signal 3: Nobody can explain a falling number

A good operator never explains away a falling number. "The algorithm changed," "seasonality," "the market's soft," "it's a long game" — these are all sometimes true and all useless without a mechanism attached. The acceptable version sounds like: cost per booked call rose 40% because the top-of-funnel creative fatigued in week three; here's the replacement and here's the date it goes live.

Ask for the mechanism. If you get a category instead of a cause three months running, the team is not diagnosing — it's narrating.

Signal 4: You're the integration layer between your own vendors

If you have an SEO shop, a paid shop, and a design shop, and you are the person making sure the landing page matches the ad which matches the offer — you're doing the general contractor's job for free.

Vendor-management research puts coordination overhead alone at 8–15% of annual vendor spend in hidden labor that never appears on an invoice, with businesses reporting roughly 30% higher total cost versus one integrated partner. Your own hours are the largest line in that hidden bill, and they're the ones you can least afford. We've costed this out in detail in what using multiple marketing vendors actually costs.

Important nuance: this signal often means consolidate, not fire. One of the vendors may be perfectly good and simply mis-scoped.

Signal 5: The system stops when they stop

Here's the test that separates a vendor from an installed system. Ask: if we ended this engagement on Friday, what would still be running on Monday?

If the honest answer is "nothing" — no documented CRM automations, no follow-up sequences you own, no reactivation engine, no tracking you can log into — then you never bought a system. You rented activity. That's the structural distinction between a marketing vendor and an installed revenue system, and it determines how much leverage you actually have in the exit conversation.

The corollary matters too: an agency that has genuinely installed things in your business is harder to fire and more valuable to keep, which is exactly the incentive alignment you want.

Signal 6: Everything you receive feels like it was written for someone else

The recurring review-site complaint is "cookie-cutter" — accompanied by "processes are confusing" and teams that lack the capacity to change direction. You know this one when you see it. The strategy deck could have your competitor's logo on it. The email sequence references a pain point that isn't yours. Requests to adjust course get absorbed into a process rather than acted on.

What to check: ask for one change to the current plan, with a date. The response tells you more than any audit. A team that owns the outcome reprioritizes. A team that owns a scope quotes you a change order.

Signal 7: The follow-up is broken and they're selling you more leads

If the standing recommendation every quarter is "increase budget" or "we need more top of funnel," and nobody has audited what happens to the leads you already have, the agency is selling the easy product rather than the right one. Most owner-led companies are leaking money between lead and close, not before lead.

Run a follow-up process audit and look at your dormant database first — reactivating old leads routinely beats buying new ones on both cost and speed. If your agency has never proposed this, ask why.

---

What should you do before you fire them?

Run a written 30-day accountability test. Firing on frustration is how owners end up hiring the same agency with a different logo. Firing on evidence is how you avoid it.

The 30-day test — send this in writing:

1. Name the number. One metric, tied to money: qualified booked calls, cost per acquisition, or revenue by source. Not three. One.
2. Name the baseline. What it was for the last 90 days, agreed in writing by both sides.
3. Ask for the diagnosis. Where in the funnel do they believe the loss is occurring, and what is the mechanism?
4. Ask for the intervention and the date. Specific change, specific ship date, inside 30 days.
5. Ask who owns it. A named senior person, not a role.
6. Ask what you keep. Full list of assets, accounts, automations, and documentation that remain yours on exit.

Two outcomes, both useful. Either the agency responds like an owner — in which case you may have just fixed the relationship for the price of one email — or the request exposes that there is no diagnosis, no owner, and nothing to keep. That's not frustration anymore. That's evidence.

When is it not the agency's fault?

Three cases where firing is the wrong move:

• You changed the goal. If the brief was awareness and you're now judging on booked revenue, that's a re-scope, not a failure. Rewrite the mandate.
• You haven't given it time. Real revenue systems compound over months. The common consultant objection — "is this just someone who leaves when the contract's up?" — is often self-fulfilling: owners exit at month four because results weren't instant, then start over, and never get past month four with anyone. Building the machine genuinely takes time.
• Your offer or your sales process is the constraint. If your close rate is 4% and your competitor's is 20%, no amount of traffic fixes that. Diagnose before you plan.

Related reading before you decide: why a marketing plan isn't the same as a marketing system, and why a fractional CMO doesn't fix what's actually broken.

How do you exit without losing the machine?

The exit is an asset-recovery exercise. Before you give notice, secure these in your own name:

Get all of it confirmed before the termination email. An agency that resists any line in this table has told you everything about how the last engagement was structured — and confirmed you were right to leave.

What should you look for in the replacement?

Not another vendor. The next hire should be judged on one thing: whether they will own the number or own a scope. That's the difference when an agency owns the outcome instead of the scope, and it's the question that should dominate every sales conversation you take.

Ask them to show you what remains installed in a former client's business. Ask what number they report in. Ask who — by name — will be on your account in month nine. Then run the full pre-signature diligence checklist, and consider whether the answer is an agency at all — there's a third option between in-house and agency most owner-operators never evaluate.

The short version

You don't fire an agency because a month was bad. You fire it when it can't tell you why the month was bad, can't point to anything installed that survives its departure, and can't report in the only unit that matters to the person signing the cheque.

Run the 30-day test first. Secure the assets. Then decide — on evidence, not on the slow accumulation of frustration that made you search this in the first place.

---

About the author: Avi Vatsa is CEO of Exchange Four Agency, where he leads the team that installs and runs AI-leveraged revenue systems for owner-led companies. His background spans law, technology, and marketing; he also co-founded Dialora, an AI voice-agent platform for automated lead capture and booking. Sources: Marketer of the Day #1411 · The Jeremy Ryan Slate Show · LinkedIn