How a solo founder at CoreFlow Studio added $4K in MRR by splitting one newsletter list into 15+ segments and speaking to lapsing members differently from new ones.

CoreFlow Studio is a video on demand yoga and Pilates platform run by one person, with no marketing team and no analyst. This playbook is how that one person added $4K in monthly recurring revenue from retention alone: the behaviour signals worth segmenting on, the drift signal that predicts a cancellation two to three weeks out, and what each segment actually receives.

The constraint shapes everything that follows. Anything requiring weekly manual work was never going to survive contact with a real week, and the starting position was a basic CRM with weak tagging, no automation, and one newsletter that went to everybody.

The problem with one list

A single list forces every message to be written for the average member. The average member does not exist.

On any given month the base contained someone who signed up four days ago and has not finished a class, someone who has practised four times a week for two years, and someone who has not opened the app in three weeks and will cancel on the next billing date. One newsletter has to serve all three, so it serves none of them.

The person about to churn is the expensive one. They need a specific message about a specific problem, and they will receive a generic announcement about a new class instead.

Behaviour over demographics

The first real decision was what to segment on. The instinct is demographics: age, location, plan type, how they found you. Almost none of that predicts anything in a subscription business.

Behaviour does. The 15+ filters that matter are all about what someone actually does:

Segment axis	What it tells you
Sessions in the last 14 days	The single strongest churn predictor in the base
Change vs their own baseline	A drop from 5 a week to 2 matters more than an absolute number
Subscription age	Month one behaves nothing like month twelve
Class type preference	Determines what to recommend and what will land as irrelevant
Time since last session	The clock that actually runs toward cancellation
Completed vs abandoned sessions	Starting and quitting is a content fit problem, not a motivation one

The second row is the one most teams miss. Absolute activity is a weak signal because members have genuinely different natural frequencies. A member dropping from five sessions a week to two is in trouble. A member who has always done two is fine. Only the change against their own baseline separates them.

The drift signal

Every subscription business has a behaviour change that reliably precedes cancellation by a predictable gap. Finding yours is the highest-value analysis available, because it converts retention from reactive to preventable.

For CoreFlow it is a sustained drop against personal baseline, and it shows up roughly two to three weeks before a cancellation. That is a real window: enough time for a message to change the outcome, which is not true once someone has clicked cancel.

A cancellation flow save offer arrives after the decision is made. A message during the drift arrives while the person still thinks of themselves as a member who has been busy.

What each segment hears

The content is the point, not the plumbing:

Drifting: not a discount, and not "we miss you", which announces that their absence was noticed and does nothing about the cause. A short, specific message naming what they used to do and offering the smallest possible re-entry: one fifteen minute class in the style they already liked. The barrier to returning is the belief that they have to restart properly.

New in month one: onboarding to a habit, not to a feature set. The predictor of month-three retention is completing a first few sessions, so everything points at that.

Consistent long-term: no reactivation content at all. Referrals, new material, community. The fastest way to annoy a loyal member is a win-back email.

Lapsed past the window: low frequency, genuinely new reasons to come back. Sending weekly to someone gone for six months trains them to filter you.

Why this survived a solo founder

Segments update automatically from behaviour. Nobody maintains lists, and members move between segments as their activity changes without anyone noticing they moved.

That is not a nice-to-have, it is why the system still runs. Manual segmentation works for two weeks, degrades in week three, and is abandoned in week five, at which point everyone concludes segmentation does not work. It does. Manual list maintenance is what does not work.

What broke and how they fixed it

The first segments were too small. Fifteen filters can generate dozens of micro-segments, each needing its own copy, which is unmanageable for one person. They collapsed to four or five actionable groups, with the filters deciding membership rather than each combination becoming its own campaign.

Win-back messaging went to people who had not left. An early drift definition was too loose and caught members who were simply on holiday. Getting a "we miss you" while on a two-week break is a small insult. Tightening the definition to sustained change against baseline fixed it.

Open rate looked like the metric. Reactivation messages have lower open rates than announcements and are worth vastly more. They moved to measuring revenue retained per segment, which changed which campaigns got attention.

The numbers

$4K in added monthly recurring revenue within ten weeks, from reduced churn alone, run by a solo founder with no operations or marketing team.

The segmentation against churn playbook