Month two is too early. Month six is late. The decision belongs somewhere in months three to five, and what makes it a decision rather than a mood is knowing which numbers are supposed to be missing at each point.
Before any of that: find out today, in writing, what you get back if you leave. Who owns the domains, the mailboxes, the contact list, the reply history, the sequences. If the answer is "we do", the switching decision is already harder than you thought, and that is worth knowing in month one rather than month five.
This all assumes you already decided to buy outbound rather than build it. If that part is still open, settle build versus buy first.
So here is a normal engagement walked through month by month, with what should be true at each point and what is not yet evidence of anything.
Month one: nothing you see is evidence
Domains bought, mailboxes created, warm-up running, list criteria argued over, copy drafted, first sends going out late in the month at low volume.
Zero meetings in month one is the expected outcome. What is a warning is silence from the agency. In month one you should be arguing about the list, seeing the actual messages before they go out, and getting told which criteria they could not fill. If month one is quiet on their side, you are not being run, you are being invoiced.
Ask for the list. Not a count, the list. Ask for it in the sales process too, before you sign anything.
Month two: first replies, no verdict available
Replies start. Some positive, most not. Maybe a meeting or two.
This is the month founders decide the agency is failing, and it is the worst possible month to decide anything. Volume has barely built and the sequences have not yet completed a full pass. A campaign this early can tell you whether the emails arrive. Whether the offer works is a month three question.
What to look at in month two: bounce rate, reply rate, and whether the replies sound like your buyer or like somebody else's. A pile of "wrong person, try marketing" replies in month two is a list problem, and a list problem is fixable inside the same contract.
Month three: the first honest read
Now the numbers mean something. Enough volume has gone out, the sequences have completed at least once, and the first meetings have happened.
Three questions, in order.
Did the list get built to the criteria you agreed, or to whatever the data vendor happened to have? Did the messages go out as approved, or did something generic get swapped in at volume? And did the meetings that happened contain people who could actually buy?
A failure on any one of those is a fixable failure, and month three is the right time to raise it as a fixable failure. Put it in writing, name the specific gap, give it a month. An agency that responds with a plan is worth keeping. An agency that responds with a dashboard is not.
Month four: where the patience argument actually lives
This is the month the hold decision gets made or lost, and it hinges on something most founders never account for.
People do not make decisions fast in every market. Some pipelines are simply getting started. Mid-market B2B deals have crept into a 60 to 120 day band, up from 45 to 90 days before 2021, part of a lengthening the same source puts at 20 to 30% since 2021. Do the arithmetic against your own calendar. A meeting booked in month two, in a market whose cycle runs two to four months, can close anywhere from month four to month six. If you cancel in month four, you are cancelling before most of that range has had time to produce a single signature, and then you carry the same result into the next agency and start the clock again.
So the month four question is whether the pipeline is shaped right, whatever the revenue line says. Real companies, real titles, real problems, moving. If that is true and nothing has closed yet, wait.
The part that might be on your side
Uncomfortable, and often true. Before you switch, check whether the campaign got fair conditions.
Did you pause it for two weeks because the team got busy? Did you change the offer mid-flight? Did anybody attend the calls that were booked, and did they run those calls well? Did replies sit unanswered for days?
A campaign that was paused, rerouted and half-attended has not been tested. Swapping vendors will reproduce the result exactly, on a new invoice.
Month five: the switch decision
By month five you have the full picture: two to three months of steady sending, a cohort of meetings that has had time to move, and one written attempt at fixing whatever went wrong in month three.
Switch if the list still does not match the criteria after you raised it, if the meetings are consistently with people who cannot buy, if the copy was never adjusted despite the evidence, or if nothing in the pipeline moved at all across three months of real volume, in a market whose sales cycle is shorter than that.
Stay if the pipeline is real and slow, if the problems you named in month three were fixed, or if the honest cause sits inside your own company.
Month six is late, and the calendar makes it later
Waiting past month five costs more than the extra invoice, because of when in the year you are standing.
A replacement needs its own setup month, and the year decides what that month costs. Fire an agency in November and the new one spends December on setup at best, starts sending in late January, and the first quarter is gone before anything moves. August is the other quiet window, and the better month for a replacement's setup if you want it sending ahead of the fourth quarter.
So if you are in month five in September, decide now. If you are in month five in April, you have more room.
Pick the next one differently
Picking the replacement is where the same mistake usually gets made twice, because the biggest promise tends to win the pitch. An agency that tells you plainly what month one will look like, and what it will not produce, is describing the same reality this post describes.
So whoever comes next, ask the ownership question before you sign, and write the month three review into the engagement from the start. Hold them to what the first ninety days actually look like. That is the same yardstick the agency you are about to fire should have been measured by, and the reason you know now that month two proved nothing.