Firing a provider feels like a decision about them. It is mostly a decision about you: whether you can tell a bad month from a bad fit, and whether you have somewhere better to put the work once they are gone. Most businesses get the timing wrong in both directions — they stay a year too long, then leave in the one week that costs them the pipeline.
The buyer who already has an outbound provider and quietly doubts it is the hardest decision on this list, because inertia is cheap and switching looks expensive. This is how to make the call with your eyes open, and what to check before you write the email.
First, separate the dip from the decline
A quiet quarter is not evidence of anything on its own. Outbound moves in cycles — a segment saturates, a message ages, a buyer’s budget cycle turns over — and a competent provider is already correcting for it before you notice. What you are looking for is not a bad month. It is a pattern the provider cannot explain and does not seem to be fixing.
The distinction that matters: a dip comes with a diagnosis, and a decline comes with a reassurance. When results soften and your provider can tell you which segment cooled, which message stopped landing and what they are changing next week, that is a working relationship having a hard month. When results soften and you get a slide about “market conditions” and a promise that next month will be stronger, the thinking has left the account. The first is worth staying for. The second rarely recovers, because nobody is looking.
The signals that are really about you
Before you blame the agency, rule yourself out — because half the terminations that should have been fixes fail again with the next provider. Three things sit on your side of the line.
Your brief. If the provider never got a clear picture of who you sell to and why they buy, the targeting was guesswork from day one, and that traces back to what you handed over rather than what they did with it. Your offer. If the thing you are asking cold buyers to say yes to is vague or unproven, no amount of outbound craft rescues it — the inbox exposes a weak offer faster than any other channel. And your own capacity to hold a standard: an unsupervised account drifts towards volume, which is often less about the agency’s intent and more about the fact that senior attention on their side is spread thin across too many clients. It is worth reading the arithmetic of an over-full practice before you assume the drift was aimed at you specifically.
If all three are sound and results are still declining with no diagnosis, the problem is genuinely theirs. If any one is shaky, changing providers just resets the same problem with a new invoice.
A provider does not fail on the month you fire them. They fail on the month you stop arguing with the report — and keep paying anyway.
Three tests before you write the email
You do not need a new spreadsheet to make this call. You need the last three monthly reports and an honest read of them. We have set out the tests a report has to survive in what an honest outbound report contains; the shortest of them settles most switching decisions on its own. Has any report your provider sent in the last six months ever told you something you did not want to hear? A report that only ever reassures is not measuring your pipeline. It is managing your perception of it.
Read the last three reports through that lens and the answer is usually already visible. If every trend runs comfortably upward, if no headline figure resolves into something you could act on this week, and if nothing in half a year of reporting has changed a single decision you made, you have your finding — and it predates the month you started doubting them.
Switching has four costs nobody prices in
Switching is not free, and pretending it is leads to the second timing mistake: leaving badly. Four costs get underpriced.
The gap. However fast the next provider starts, there is a ramp — learning your market, warming domains, building sequences — and for six to eight weeks the pipeline you were relying on is thinner than it was. That is an assumption from how these transitions typically run, not a guarantee, but budget for a lull rather than a handover. The knowledge. Everything the outgoing provider learned about which segments respond and which messages died walks out with them unless you extract it deliberately. The infrastructure. If the outbound ran on their domains, their tooling and their data, you may be starting the technical setup from zero — and domain reputation, once you are rebuilding it, takes weeks to earn back. And the live conversations: half-warm replies and booked meetings mid-sequence are the easiest thing to drop in a messy exit, and they are the most valuable thing you own.
None of these is a reason to stay with a provider who has stopped working. They are reasons to leave on a plan rather than in a fit of frustration.
How to leave without losing the pipeline
Time the exit to protect the pipeline, not to punish the provider. Before you give notice, get three things out of the relationship while it is still live: the full contact and reply data in a format you own, a written account of what worked and what was retired, and a clean list of every live conversation with its current state. Most retainers run on a notice period — use it to run the handover, not to coast.
Then sequence the change so it overlaps rather than gaps. Line up the replacement before the old arrangement ends, so warming and briefing happen while the outgoing provider is still holding the live conversations. The worst version of this is a hard stop on the last day of a contract with nothing ready to catch what is in flight. The pipeline does not care whose fault the gap was.
The provider you replace them with is the real decision
Firing the agency is the easy half. The half that decides whether you are better off in a year is what you put in its place — and if you replace like with like, you have bought a change of logo and kept the failure mode. The question is which model actually fits the problem you have now, which is a different question from the one you answered when you first hired: we set out the four ways to structure that work, and the honest answer is often not another agency.
The reason accounts drift towards unsupervised volume is that most models split the thinking from the doing and leave you to bridge the gap. A senior-led practice that sets the strategy and runs it with the same hands removes the bridge, because there is no junior to hand your brand to and no diagnosis sitting in a drawer while the sends go out regardless.
So do not fire your agency on a bad month. Fire them on a bad pattern, after you have ruled yourself out, with the data extracted and the replacement ready. The month you finally decide is rarely the month the relationship broke — it broke the quarter you stopped reading the report, and kept paying anyway.