We have a 24-month cohort on the books now. Thirty-eight ecommerce founder clients who signed between May 2024 and May 2025. We ran the numbers last week: 19 are still active at month 18+, 11 churned between month 7 and month 12, and 8 churned inside the first 6 months.
What we wanted to know: is there anything a client does in their first 30 days that predicts whether they're still here at month 18?
There is. Five behaviors. Every retained-18-months client did at least four of them. Every month-7-to-12 churn missed three or more. It is not subtle.
The five month-1 behaviors
We are not talking about anything mystical. These are observable in our shared Notion, in calendar invites, and in Slack DM history. We back-coded all 38 onboardings against this list.
1. They forwarded a customer DM or sales call note inside the first 14 days
Not asked. Unprompted. Some version of "this came in today, might be a post." The retained-18-months cohort did this a median of 3.2 times in the first 30 days. The churned cohort did it a median of 0.4 times.
This is the single strongest signal. A founder who sees their day-to-day operator life as the input to their content engine within two weeks of signing is a founder who will still be sending us pipeline anecdotes in month 18.
2. They booked their voice-sync calls themselves and showed up on time
We schedule weekly 30-minute syncs in week 1, then biweekly. The retained cohort booked all four weeks of month-1 calls inside the first 3 days of onboarding. Churned cohort: 6 of 19 rescheduled their week-2 call at least once.
A rescheduled month-1 call is not a calendar problem. It is a "this is not on the priority list" signal. Almost every client who rescheduled twice in month 1 had churned by month 9.
3. They told us something they were wrong about in the last year
We ask this in the kickoff intake: "Where did you change your mind in the last 12 months?" Retained clients answered with a specific operational reversal — a SKU they killed, a channel they exited, a hire they regretted. The answers were 80-200 words.
Churned clients gave us generic answers — "I underinvested in brand" — or no answer at all. We can ghostwrite around generic answers. We can't ghostwrite good posts around them. The founder who can articulate a specific wrongness in month 1 is the founder whose drafts will not need 4 rounds of revision in month 6.
4. They edited a draft with a substantive operational note, not a copy edit
The first or second draft we deliver, every client edits. Retained clients sent back edits like "this number is wrong, we hit 14% TACoS not 18%, here's the screenshot" or "this framing isn't how I'd say it to another founder — it's how a consultant would." Operational substance.
Churned clients sent back edits like "can we make this shorter" or "I don't love the word 'leverage.'" Surface-level. We can ship surface-level edits. They do not produce drafts that compound over 18 months because the founder is not in the writing.
5. They repurposed a ghostwritten post into a sales asset inside 30 days
The retained-18-months cohort, every single one, pulled a ghostwritten post into a deck, a proposal, a sales email, or a podcast pitch inside their first 30 days. We saw this in shared Notion comments, in forwarded "btw I used this in a deck today" notes, in pipeline-call mentions.
If a founder is not pulling their own content into their commercial life within a month, the content is decoration. Decoration churns.
What this tells us about onboarding
We rewrote our onboarding playbook in March 2026 based on this data. The change: we now engineer these five behaviors in week 1.
Specifically:
- Day 1, we send a one-line ask: "Forward us the next customer DM or sales call note that lands today."
- Day 3, we send the wrongness prompt as a written assignment, not a verbal one — written answers are 4x longer in our data.
- Day 7, after first draft, we ask the founder to mark one factual error and one framing miss, even if they have to invent the framing miss. Forces operational engagement.
- Day 14, we send a "where could you use this?" prompt with three concrete suggestions (deck slide, cold email, podcast intro).
- Day 21, we ask for the next two months of operational decisions they're sitting on. Pipeline of post material.
Since the rewrite (8 clients onboarded), we have zero month-1 dropouts and a measurable 41% increase in operator-detail density in month-3 drafts vs the prior cohort.
What we stopped doing
We used to lead with portfolio walkthroughs — show the client a polished example, build excitement. The data killed this. Polished-portfolio onboardings correlated with higher month-1 satisfaction scores and lower month-9 retention. Founders who got sold the dream churned. Founders who got handed homework on day 1 stayed.
We also stopped sending "welcome packets." A 14-page welcome packet is a churn signal we were generating ourselves. The retained-18-months clients did not need it. The churned clients did not read it.
What this means if you are picking a ghostwriter
Ask the ghostwriter what their month-1 onboarding looks like. If the answer is heavy on deliverables to you (welcome decks, content calendars, voice analysis docs), be suspicious. The behaviors that predict 18-month retention all involve the founder doing the work of feeding the system. A ghostwriter who is not asking you to do homework in week 1 is a ghostwriter who will be guessing at your voice by month 6.
The same rule runs both ways. If you are a founder who reads month 1 and thinks "this is too much work, I want them to just write" — you are probably a 6-month client. That is fine. Just price the engagement accordingly.
FAQ
How long is the typical ghostwriting engagement before retention patterns show? Our churn distribution clusters at month 7-12. That is when most founders re-evaluate. Month 1 behaviors are the earliest predictor we have found. Pipeline signals (inbound DMs from ICP) usually appear month 4-6 and become a co-predictor by month 9.
What if a founder is genuinely busy in month 1? Genuinely-busy founders who later renewed asked for a 2-week start delay before signing. They came in ready. The pattern that kills retention is signed-but-distracted, not signed-and-delayed.
Can we predict churn for in-flight clients using these signals? We now run a 30-day audit on every new client. Two or fewer of the five behaviors triggers an escalation call where we name the pattern directly. Of the four clients we've had this conversation with since January, three corrected and one chose to pause — which is the right outcome.
If you are evaluating a LinkedIn ghostwriting engagement and want to see what month 1 actually looks like, reach out. We will send you the homework before we send you the proposal.