Every engagement we run starts with a document that names the buyer. Company size, role, channel mix, the problem they have on a Tuesday. The founder writes it with us in week one, and it is almost always wrong in a specific, predictable way.
Not wrong as in careless. Wrong as in written from the inside. The founder describes the customer they want, or the customer they had three years ago, or the customer that would make the business feel the way they'd like it to feel. Then we write to that person for a quarter, and somewhere around month four the inbound starts telling a different story.
What the founder does with that story is one of the cleanest client success patterns we track. Founders who correct the ICP in month four build pipeline. Founders who defend the original brief spend the rest of the year writing to somebody who isn't reading.
The brief describes the buyer you want. The inbound describes the buyer you have.
Here is the shape it takes on an ecommerce founder's account.
The brief says: "Amazon brands doing $2M–$10M, founder-led, in-house team, looking for a creative partner." Reasonable. Specific. Sounds like a customer.
Month four, we open the inbound log. Eleven qualified conversations. Four are agencies asking about white-label work. Three are brands under $1M with no team at all. Two are aggregators. One is a $15M brand whose marketing director found a post about return reason codes. One matches the brief.
The founder's first reaction, every time, is that the content is attracting the wrong people. Sometimes that's true. Usually it isn't. The content is attracting the people who actually have the problem you write about, and the brief was a guess about who that would be. The inbound is not a guess. It's a measurement.
Four months is roughly how long it takes for that measurement to be worth reading. Under twenty inbound conversations you're looking at noise. Past twenty, the shape is stable, and it rarely matches page one of the onboarding doc.
Why founders defend the brief
Nobody defends a wrong brief because they're stubborn. They defend it for three reasons that all feel like good judgment from the inside.
The brief was a business decision, not a content decision. The founder chose the ICP for reasons that have nothing to do with LinkedIn: margin, deal size, who they enjoy working with, who the last agency told them to target. Changing it feels like changing the business, so it gets protected the way a business decision gets protected.
The actual buyer is less flattering than the briefed one. Agencies asking for white-label help do not feel like the audience of a thought leader. Sub-$1M brands feel like a step backward. The $15M marketing director feels like an anomaly. The founder wanted to be read by peers and is being read by people with budgets, and those are different populations.
The specificity that made the content work came from the founder's real experience, which was formed with the buyer they have, not the buyer they want. So the posts land with the audience the founder actually understands. That's not a bug. That's the mechanism doing exactly what it should, and the brief being the thing that's out of step.
The two responses, and what each one costs
The correction. The founder reads the inbound log, accepts that it describes something real, and we re-aim. Not the topic. The reader. Same lane, same specificity, aimed at the marketing director instead of the founder, or at the small brand with no team instead of the mid-market brand with a department. Over the next two months the inbound gets more like the log, not less, because the platform is now getting a cleaner signal about who engages. Month six, the founder has a pipeline that matches the content. Month nine, they usually have a new offer shaped for the buyer who was there all along.
The defence. The founder concludes the content needs to be "more strategic" to attract the briefed buyer. Posts move up a level of altitude. Working detail comes out. The account starts sounding like a person who wants to be hired by a $5M brand instead of a person who has solved a $5M brand's problem. The original audience, the one with the budget and the problem, stops recognising the account. The briefed audience never arrives, because it wasn't there to begin with. By month eight, engagement is down, inbound is thinner, and the founder has a strong case that LinkedIn doesn't work for their category.
We've watched both. The second one is the more common of the two, and it is almost never framed as a decision. It's framed as "tightening the messaging."
What we actually do at month four
We run the inbound log against the brief in the month-four review, on purpose, before the founder has formed a story about it.
We count, we don't characterise. Every inbound gets a row: company size band, role, what they asked for, whether they could buy. Twenty rows, three columns. It takes fifteen minutes and it removes the "we're attracting the wrong people" conversation entirely, because the wrong people are visible and countable and usually a minority.
We separate "not the brief" from "can't buy." An agency asking for white-label work is not the brief. It is also, frequently, a buyer with recurring volume. A student asking for career advice is not the brief and can't buy. Those two get lumped together as "wrong inbound" and they have nothing in common.
We ask which offer the actual buyer would need. Sometimes the correction isn't the content. A founder who briefed mid-market brands and is getting sub-$1M brands doesn't necessarily need to change the writing. They need a productised version of the service that a small brand can afford. The content already did its job. The business didn't have a door for the people it brought.
We name it as a correction, not a failure. This matters more than it sounds. If the founder hears "the content is off," they reach for the defence. If they hear "the content found your buyer and your buyer is not who you thought," they reach for the correction. Same data. Different framing. Different year.
The tell that it's working
Around month six, after a correction, the founder stops describing inbound as wrong and starts describing it as early. "A brand that's a bit small for us reached out" becomes "another one of the small brands reached out, we should build something for them." That's the ICP updating in the founder's head, which is the only place it really lives.
The second tell is that the offer changes shape before we change the content. The founder builds the smaller package, or the agency partnership tier, or the marketing-director-facing audit. The content then gets to point at something that exists, and inbound quality jumps because there's finally a door.
FAQ
Doesn't correcting the ICP mean accepting worse customers?
Sometimes it means accepting different customers. The sub-$1M brand with no team is a worse customer for a $6K/month bespoke engagement and an excellent customer for a $1,500 productised one. The correction is often to the offer, not the standard.
What if the inbound really is wrong?
Then the log will say so: rows full of people who can't buy, in any band. That's a lane problem or a hook problem, and it's fixable, and it's a different fix. The point is to look at the rows before deciding which problem you have.
Why four months? Why not adjust every month?
Because the first eight weeks of inbound are dominated by people who already knew you. The strangers, the ones the content actually earned, show up from month three. Month four is the first time you have enough of them to read.
Can't you just write to the briefed buyer harder?
You can write to them. You can't make them have the problem. If the brief describes a buyer who doesn't experience the thing you're specific about, no amount of aiming fixes that, and aiming harder usually means going broader, which loses the audience you had.
If you want a content operation that reads the inbound as data instead of defending a document, that's the work. EcomGhosts writes for ecommerce founders and Amazon operators, and month four is on the calendar from week one.