Why clients fire marketing agencies

Ask an owner why they left their last agency and you will hear “we weren't seeing results”. Ask what results they were expecting and the answer is often vaguer than the complaint. That gap is the whole subject of this page: the stated reason and the actual reason are rarely the same, and the actual reason is usually fixable by the marketer without spending a penny more on ads.

The reason given is not the reason

Marketing takes months to compound, and every client knows it when they sign. So a relationship that ends at month seven did not end because seven months was too slow — it ended because somewhere around month four the client stopped being able to tell the difference between “this is working and needs time” and “this is not working and nobody has noticed”.

Once a client cannot tell those apart, every month of patience costs them more nerve than the last. Eventually the invoice arrives on a bad week and the answer is no. What they say is “results”. What happened is that they lost the ability to judge, and judging is the one thing they cannot delegate.

Why the monthly report makes it worse

Most agency reporting is a list of what happened: impressions, clicks, reach, a rate that moved. It is accurate, it takes real work to produce, and it answers a question the client never asked.

The question they actually have is what did you decide, and was it right? A metrics report cannot answer it, and worse, it implies the numbers are the work. So a client reading a good month and a bad month sees the same document with different figures, and learns nothing either time about whether their marketer is any good.

This is why adding more charts never rescues a shaky relationship. The problem is not that the client has too little information. It is that none of the information is about judgement.

What a client is actually buying

Nobody hires a marketer for impressions. They hire one because they cannot tell which of six plausible things to do next, and they are buying somebody who can — repeatedly, and for reasons they could follow if they asked.

That means the deliverable is the decision, and the evidence is supporting material. Most reporting has this exactly inverted, which is why most reporting fails to defend the retainer it accompanies.

The four things a report has to contain

What you decided, and why.Named decisions, in sentences, with the reasoning that produced them. “We moved the budget to search because the studios you named search for the problem and do not follow brands” is a sentence a client can agree or disagree with. That is the point — a decision they can argue with is one they are participating in.

What you decided against, and why.This is the half almost nobody sends, and it is the most persuasive thing in the document. A marketer who writes “we are not doing LinkedIn ads, because your buyers are not there” has demonstrated judgement in a way no positive result can, because they have shown the work of ruling something out. Clients remember it.

What the evidence showed — including its limits. If the platform reported 212 conversions and 180 of them were link clicks, say so. A client who later discovers the caveat you omitted will re-read every number you ever sent them. A client you told first will believe the next figure without checking.

What you need from them. Every stalled marketing programme has an unanswered question inside it, usually about the offer, the price, or who the customer really is. Putting it in the report moves the delay to the correct side of the relationship.

Why this favours the independent marketer

An agency has a structural reason to avoid this. Their margin depends on the client not knowing which hours produced which outcome, and a decision trail makes that legible. So the format stays as it is.

An independent consultant has the opposite incentive. Transparency is the only advantage available against a firm with more people and a better-looking deck — and it is a real one, because the thing being transparent about is judgement, which is what the client is short of and cannot buy in volume.

If you have not been keeping the trail

Most people reading this have not, because nothing in the ordinary toolchain encourages it. Decisions live in Slack, the reasoning lives in somebody's head, and the report gets assembled from whatever the dashboards will export.

The recovery is not retrospective. Do not reconstruct six months of reasoning — a reason written after the outcome is known is not a reason, and a client can smell it. Start recording decisions as you make them, and send the first report covering the period from now. Saying “from this month, here is what I decided and why” is itself a credible act.

What this looks like in practice

Rallik was built around this shape: decisions are kept with the reasoning that produced them, checked later against what actually happened, and the limits of the evidence travel with the numbers rather than being edited out. The client report it produces leads with decisions and puts figures underneath them as support — including the decisions that did not survive contact with real data.

You do not need this product to do any of it. You need the discipline of writing down what you decided and why, before you know whether it worked. The tooling only matters because that discipline is very hard to sustain by hand across six clients, which is exactly when it starts being worth the most.