Software for marketing agencies: reporting is the real problem

Marketing agencies are unusual buyers of software. They already live inside a dozen platforms, most of them supplied by the channels they buy, and each one produces its own numbers in its own shape. The gap is never a missing platform. It is the space between the platforms, which is where the client report gets assembled by hand at the end of every month.

That assembly is the largest recoverable cost in most agencies, and it is invisible because it is nobody's project. It appears as a busy week, not a line item.

Where the month actually goes

The pattern is consistent enough to be predictable. Someone exports from the ad platforms, someone else pulls analytics, a third person has the call notes, and one person merges all of it into a deck per client. For an agency with twelve retainers that is twelve decks, each assembled from four or five sources, each with a slightly different definition of a conversion.

The merge happens in a spreadsheet, and this is where the quiet risk sits. A 2024 study led by Pak-Lok Poon, published in Frontiers of Computer Science and summarised here, examined spreadsheets used in business decision-making and found critical errors in 94% of them. That is not a claim about careless people. It is a claim about a tool being used for a job it was never designed to do repeatedly and under time pressure, which is exactly the month-end report.

For an agency the consequence is specific: the number you defend on a client call may be wrong, and you will not find out from the spreadsheet. You will find out from the client.

What agencies buy instead, and why it does not help

• A reporting platform. Solves the export step, not the definition step. If Meta and GA4 disagree about a conversion, a dashboard shows both numbers faster without deciding which is true.
• A project management tool. Genuinely useful, and orthogonal. It organises the work of making the report; it does not make the report.
• An all-in-one agency suite. Replaces four tools with one that does each job slightly worse, and the migration costs a quarter. Sometimes right, rarely right in the year it is proposed.

None of these are bad products. They are answers to a question the agency has not asked precisely enough yet.

The question to ask first

Before buying anything, write down the definition of every metric that appears on a client report, and who decides it when two sources disagree. This takes an afternoon and it is the whole project. Most agencies discover during the exercise that three account managers have been reporting the same metric three different ways, which is a governance problem no software fixes.

Once the definitions are written, the automation is usually small. The reason it looked large was that nobody could specify the output. This is the same trap as any internal build, and the sequence for avoiding it is set out in how to scope a software project.

Build, buy, or neither

The middle row is where most agencies sit and where the least gets done, because the problem is large enough to be annoying and small enough to postpone. The relevant reading is automating manual reporting without buying a BI platform, which is the same argument at more length, and what 'real time' actually means in a dashboard, because agencies routinely pay for a freshness they do not need.

If the conclusion is that something should be built, the build-versus-buy read on custom CRM cost is the closest cost analogue, and the point where a spreadsheet stops paying for itself generalises past inventory to any recurring merge.

The honest limit

An agency under roughly eight clients should not build anything. The reporting pain is real but it is smaller than the cost of maintaining a tool, and the agency will change its channel mix at least once before the build pays back.

There is also a case that reporting should stay partly manual on purpose. The month-end assembly is when somebody actually looks at the numbers, and agencies that fully automated it sometimes found that nobody read the output at all. Automate the collection and the arithmetic. Keep a human writing the interpretation, because that is the part the client is paying for.

Worth saying too that the tooling question often arrives dressed as a cost question. Before an agency buys anything it is usually worth auditing what the current stack already does, and KF Agency's Arabic roundup of free social media management tools (in Arabic) is a reasonable inventory of what does not need paying for at all.