Why we stopped writing 40-page audits

We used to open every engagement with a big audit. Forty pages, sometimes more. A cover slide, a table of contents, an executive summary that nobody read, then thirty-something pages of findings organized by channel. It looked like consulting. Clients paid for it and told us it was thorough. We kept making them for about two years before we admitted the obvious: the document was a performance, not a plan.

Here is what actually happened after we handed one over. The client read the first four pages, skimmed the rest, forwarded it to someone on their team, and asked us what to do first. The forty pages did not answer that question. They described the terrain. They did not tell anyone where to walk.

The audit was solving the wrong problem

An audit answers "what is wrong." That is a real question, but it is rarely the bottleneck. Most growth-stage companies and local service businesses already have a rough sense of what is wrong. Their conversion rate is low. Their cost per lead crept up. Their organic traffic flattened. They do not need a stranger to confirm it in a slide deck with a red, yellow, green scoring rubric.

The bottleneck is sequencing. Given twelve problems and a finite budget, which three do you fix this quarter, in what order, and how will you know if it worked? A forty-page audit is almost hostile to that decision. It presents everything as roughly equal because a scoring rubric flattens everything to a color. A page on your meta descriptions sits next to a page on your entire paid acquisition strategy, both marked yellow, as if they carry the same weight.

They do not carry the same weight. Pretending they do is the core dishonesty of the format.

A long audit signals effort. A short plan signals judgment. Clients are paying for the second one, even when they ask for the first.

Length was a proxy for trust, and it was a bad one

We told ourselves the page count built confidence. If the client can see how much we looked at, they will trust the recommendation. That logic is backwards. Volume is what you produce when you are not confident enough to prioritize. Anyone can list forty findings. It takes actual judgment to say "ignore thirty-five of these, here are the five that move revenue, and here is why the other thirty-five do not matter yet."

The trade-off is uncomfortable. A short plan is falsifiable. If we say "your paid search is bleeding budget on branded terms you already rank for, and fixing that frees roughly a fifth of spend for prospecting," we can be wrong about that, in public, on the first page. A forty-page audit hedges. It never commits hard enough to be wrong, which also means it never commits hard enough to be useful.

We would rather be wrong and correctable than thorough and inert.

What we do instead

We still look at everything. The analysis did not shrink. What shrank is the deliverable. The work now produces a document that fits the way a decision actually gets made.

• One page on the constraint. Not a list of problems. The single thing holding growth back right now, stated plainly, with the evidence that points to it. Sometimes it is the offer, not the marketing. We say so.
• Three moves, ranked. What we do first, second, third. Each with the mechanism it improves and the metric it should move. If we cannot name the metric, the move does not make the list.
• What we are deliberately ignoring. The stuff that scores poorly but does not matter yet. Naming it out loud stops the client from getting nervous about it in month two.
• A rough read on time and money. Not a fake forecast. The order of magnitude and the assumptions behind it, so we can revise together when reality disagrees.

That is usually three to five pages. It takes longer to write than forty, because compression is the expensive part. Deciding what to leave out is the whole job.

The grader does the inventory so the plan does not have to

Part of why we could drop the page count is that we moved the inventory work into software. A free grader can enumerate the technical and structural issues on a site faster and more consistently than a human writing paragraphs about them. Let the tool produce the checklist. That frees the human deliverable to do the thing software is bad at: deciding what the list means and what to do about it in what order.

This is a general pattern worth stating. When a machine can produce the inventory, a person padding a document with that same inventory is adding negative value. They are re-typing what a scan already knows, and charging for the typing.

The objection we hear, and our answer

"But how do I know you looked at everything if you only show me five pages?"

Fair. Two answers.

First, we will show you the full inventory on request, or point you at the grader output that generated it. The raw findings exist. We just do not confuse the raw findings with the plan. One is a database query. The other is a decision.

Second, and more honestly: you do not hire us to prove we looked. You hire us to be right about what to do next, and to be on the hook for the outcome. Our compensation is tied to results, not to the weight of the paper we hand over. A forty-page audit is a great deliverable for a firm that gets paid whether or not the recommendations work. It is a strange deliverable for one that gets paid when they do.

That alignment changes what a good document looks like. If we get paid on outcomes, every extra page we write that does not change a decision is pure cost to us. The incentive runs straight toward the shortest thing that produces the right action. Length stops being a feature and becomes waste.

The rule we kept

We did not throw out rigor. We threw out the theater of rigor. The rule now is simple: every page has to change a decision the client is about to make. If a section does not change what they do on Monday, it comes out. Most of the forty pages came out. What is left is the part that was always the point.

If you want a straight read on your own constraint instead of a scoring rubric, start with the free grader, or if you would rather have someone on the hook for the outcome, apply for partnership and we will send you three moves, not forty pages.