when to fire your agency.

switching agencies costs three months minimum — notice period, procurement, onboarding, and the first month of the new relationship being mostly setup. sometimes that is money well spent. often it is a very expensive way to move an unsolved problem to a new supplier, which is why it is worth diagnosing carefully before you act.

the useful test is whether the problem lives in the work, the relationship, or your own process. only the first two are reasons to leave.

the reason this is hard to judge from inside is that all three feel identical day to day. output is late, results are flat, calls are frustrating. the cause is invisible unless you look for it deliberately, which is what the questions below are for.

— 01four reasons to leave.

dishonesty about results. reporting that hides a bad month, attribution that credits the agency for revenue it did not create, or a number that changed definition without anyone saying so. this is not recoverable, because everything else you are told becomes unverifiable.

quality that does not improve after clear feedback. everyone ships something weak occasionally. what matters is whether specific, written feedback changes the next round. if the same note has to be given three times, the capability is not there.

no measurement at all. if nobody can tell you what an enquiry costs after six months, the account is being administered rather than managed. that is a structural failure and it will not fix itself.

the team changed without a conversation. you hired named people. if the senior person you met is now on a different account and nobody mentioned it, the thing you bought no longer exists.

note what is not on this list: a single bad campaign, an idea you disliked, or a month where the numbers went the wrong way. those are normal. an agency that never has a flat month is either working in an unusually forgiving category or is not reporting honestly, which returns you to reason one.

— 02three reasons that are usually yours.

a slow quarter. marketing has variance. one flat month with a clear explanation and a plan is not a firing offence, and switching resets the learning that would have produced the recovery.

approvals. if content routinely waits a week for sign-off, the agency is not slow — your process is. we have watched brands fire an agency for missed deadlines that were entirely caused by a founder who approved in batches every ten days.

strategy you never agreed to. if you cannot state the plan in a sentence, the failure is upstream of execution. that is worth one honest conversation before it becomes a termination, because a new agency will need the same thing from you and will fail the same way if it does not get it.

a fourth borderline case: results that are fine but the relationship feels flat. that is a reason to renegotiate scope or people, not necessarily to rebuild from zero.

it is worth asking for a change of people before a change of supplier. account chemistry is real, and most studios would far rather move a strategist than lose a client. the request is much less awkward than it feels, and the response tells you how the agency is run.

— 03the conversation to have first.

put it in writing and be specific. three things that are not working, with examples. what you need to see in sixty days, in numbers where possible. and what you are prepared to change on your side — approval speed, access, a decision-maker who attends the call.

a good agency will respond with a revised plan and, often, a couple of things they need from you. a struggling one will respond with reassurance and no changes. you will know which you have within a week, and that answer is more reliable than any amount of internal debate.

give it one review cycle. if the sixty-day review shows no movement on the specific items you named, you have a documented case and a clean conscience.

— 04if you do leave, leave properly.

get the handover in writing before you serve notice: account ownership transferred, raw files delivered, tracking documentation, the content calendar, and any templates built with your money. do this first, because goodwill declines sharply after notice is given.

then brief the next agency on what actually happened, including your own part in it. the fastest-improving relationships we start are the ones where a client says "here is what went wrong last time and here is what we changed about how we work". the slowest are the ones where the previous agency is described as simply incompetent.

and give the new team the same sixty-day structure from the start: what you need to see, in numbers, by when. the discipline that would have saved the last relationship is the discipline that makes the next one work, and it is entirely within your control.