Introduction — The Month Engagement Rose and Everything Got Worse
Twelve posts, thirty-four drafts, and a rate that rose because reach halved.
Rafa ran six clients on a flat monthly fee. Twelve posts each, two platforms, captions and graphics included, replies "as needed". It read well on a rate card and it had won every one of the six.
Month one produced twelve published posts for the first client. It also produced thirty-four drafts. One client rewrote every caption — not badly, not unreasonably, just differently, in a voice that was theirs and had never been written down anywhere. Another approved nothing for eleven days and then approved everything at once on a Sunday night, which meant a week of scheduled content had to be rebuilt because three of the posts referred to an event that had already happened. Across all six clients the average post went through two and a bit rounds before it went live, and none of that was in the fee, because the fee had been calculated on twelve posts rather than on thirty-four drafts.
The content bank held about two days. That was the number that made everything else worse. With two days of approved material in reserve, every approval delay was an emergency, every sick day was an emergency, and every campaign that consumed the bank left nothing behind it. Rafa was producing content on the day it was needed, every week, for six clients, which meant there was never a moment where the work was ahead rather than behind.
In month four a reel used a trending audio track — the one everybody was using, taken from the platform's own library, on the platform's own tools. It was taken down, and the client's account received a strike. Nobody could say precisely why that use and not the thousand others. There was no record of where the audio had come from, because there had never been a record of where anything came from.
And in month five, one client's engagement rate rose by about forty per cent. They were delighted. Rafa put it in the report because it was true. It was also true that reach had fallen by roughly half that month, and that a smaller number of people seeing the content while the same committed few interacted with it is exactly what produces a rising engagement rate. The number went up because the denominator went down. The client read the same figure as evidence the content was working, and asked for more of the same — more of what had, in fact, coincided with half as many people seeing anything at all.
By month six Rafa was doing the work of a full-time job in evenings, on a fee set for something much smaller, with no way to explain any of it. There was no voice document to check a rewrite against, so every rewrite was a matter of taste. There was no bounded revision round in any agreement, so extra rounds were free. There was no rights record, so the takedown could not be traced. And there was no baseline of the client's own figures with the definitions copied alongside them, so a rising rate over a falling reach looked like success to everybody in the conversation.
Nobody behaved badly. The clients asked for reasonable changes. The audio was on the platform. The engagement figure was real. The fee had seemed generous when it was quoted.
Every part of it was ordinary, and all of it was expensive.
What this book is about
This book is about the four things underneath that story, because every one of them is a property of the trade rather than a mistake anybody made.
The stream never ends. There is no version of this work that reaches a finished state. Clear the backlog and it regenerates the same week. A feed that stands still does not stay level, it decays — which means stopping is itself a visible act. That is why the content bank is the central operational fact in this book, and why it appears in Chapter 5 rather than somewhere near the end.
The voice is borrowed and the audience is silent. You write as somebody else, to people who overwhelmingly never respond. The client is not the audience. The loud handful in the comments is not the audience either. Which leaves you producing work in a voice that is not yours, judged by somebody who is not who it was for.
Distribution is rationed, and you will be blamed for it. How many people see a post is allocated by a system that changes without notice and explains nothing. A well-made post with poor distribution is, from the client's chair, indistinguishable from a badly-made one. And the derived figures move in ways that mislead everybody honestly — engagement rate rises when reach falls.
The most valuable work leaves no artifact. The complaint handled quietly in the replies. The comment removed before anybody else saw it. The post that was not published. The crisis that did not become one. None of it appears in a deliverable, none of it shows in a report unless you put it there, and all of it is the first thing a client cuts when they are looking at what they are paying for.
What this book will give you
A voice document built from what the client already wrote, so a rewrite becomes a check against a standard rather than an argument about taste. A claims register, so nothing gets published that the business cannot stand behind. A revision round that is counted, bounded and written into the agreement, so the second and third pass are priced instead of absorbed. A content bank measured in days, and a plan for building one from a standing start. A rights record covering every image, track, font and clip, captured at the point of use rather than reconstructed after a takedown. A reply library with a pre-approved set, which is what makes a holiday possible. And your own dated baseline with every metric definition copied alongside it, so that a rising rate over a falling reach can be explained rather than celebrated.
The economic unit is revenue per published asset — what a piece of finished, approved, live content actually returns against what it cost to produce, revisions included.
The leading indicator is revision rounds per published asset. It is the number that decides whether a fee works, and almost nobody measures it.
The master variable is bank depth in days — how much approved content sits ready. It is the difference between a schedule and a permanent emergency.
And the risk indicator is the share of published assets whose rights you cannot document, because that is the one that ends accounts.
Who this is for
Somebody who has priced a month of content on the number of posts rather than the number of drafts. Somebody who has never once counted a revision round. Somebody producing on the day of publication, every week, with nothing in reserve. And anybody who has watched a client celebrate a rising rate in a month when fewer people saw anything.
There is one more person in this, and they have no seat at the table: the reader. The customer scrolling past. They did not agree to be marketed to, they will never know your name, and they are the one who encounters a wrong claim, an undisclosed ad, a stolen image or a complaint left standing in public. Chapter 9 and Chapter 15 exist for them.
©2026 James Henderson / https://localhandyman.work