Ask a membership owner why people cancel and you get the usual list. Too busy. Too expensive. They didn't do the work.
Nathalie Dorémieux has helped build over 300 membership sites, and her read on that list is blunt: it's mostly a cover story. The cancel click is the last step of a process that started weeks earlier, and most of what drives it sits inside the owner's control.
This is a breakdown of how she finds those drivers and what she does about them.
What The Membership Lab Builds
Nathalie spent ten years as a software engineer in Silicon Valley biotech before co-founding The Membership Lab with her husband Olivier. The agency designs and builds membership sites for coaches and course creators, mostly on WordPress, and moves clients off hosted platforms when they outgrow them.
The engineering background shows up in how she talks about retention. When members stall, drift, and cancel, she treats it the way an engineer treats a failing system: the output tells you where the design broke. Her position is that retention problems are design problems. The member is rarely the bug.
The Differentiator: Results Are the Retention Strategy
The industry line used to be that people come for the content and stay for the community. Nathalie calls that “BS,” with one honest caveat: it worked once, and she's still in a few memberships purely for the community. But as a strategy it's finished.
“Now people want results. They are not getting results, they're out.”
Her framework for this is the CSA model: content, support, accountability. Content on its own gives members a library. The support and accountability layers are what turn the library into progress, and progress is the only thing that reliably renews a subscription. She points to a client as proof that the shape matters less than the fit: hundreds of members, one call a week plus a replay, no community at all, and the membership thrives because it gives people exactly what they came for and nothing more.
The Mechanism: Finding the Stuck Spot
Her recent newsletter claim is that every program has at least one spot where members get stuck, and most owners don't know where theirs is. The fix starts with instrumentation, not intuition.
Track what members actually do: how long each step takes, whether they did the work after a lesson asked them to. When someone stalls, ask them why. When the same question comes up on calls week after week, that's the spot, and the answer is usually a tool or resource built for that exact moment, not more content. A lot of stalls, she's found, come down to decisions: members treat each decision as permanent, so they make none. Designing for smaller, faster decisions keeps them moving.
She has a personal stake in this. Years ago she paid about $15K a year for a coaching program and flew from France to the US four times a year to attend. At one retreat, the coach told the whole room: “I should charge more, because you guys are not doing the work.” When Nathalie later spoke to her directly, the coach's parting line was worse: “I hope you make it.”
That experience is why she thinks about the members nobody sees. Big programs parade 40 or 50 video case studies, she notes, while thousands of buyers behind them stay stuck. Those thousands are the people she's designing for, and they're the ones whose lifetime value disappears when the program blames them instead of looking at itself.
The Insight: The Cancel Click Is an Autopsy
Put yourself in the member's shoes the month before they cancel. Nathalie lists what's usually going on. The payments have gone through two or three times and they still haven't logged in, so canceling feels like stopping waste. Or they joined for the right reason but the delivery format doesn't fit them: she gives her own example of being unable to raise her hand on a group call with a thousand people, no matter how good the coach is. Or they're active in all the wrong ways, consuming, taking notes, posting in the community, but never sending the email or doing the step that moves their business. Activity feels like progress. It isn't.
Her point: “If you don't send the email, it's a no by default.”
Then there's involuntary churn. Failed payments and expired cards need automated reminder sequences, and some members let an expired card die on purpose because it's a cancel that requires no conversation.
Which brings up the exit itself, the part most owners get wrong. Her rules: let people cancel instantly, don't make them email anyone, don't put save-offers between them and the button. Offer help on the way out, offer an optional open-ended feedback box, and take the data.
“People should not have to tell you a no is a no.”
The reason to do this: the cancel experience is the last thing a member remembers about you, and members who leave on good terms come back.
How It's Engineered: The Right Members In
Retention starts before anyone joins, and her filter is controversial on purpose: price.
Will you pay $47 a month for a membership and do nothing with it? Plenty of people do, she admits she has. At $500 a month, nobody pays to let it sit. The price you'd be okay losing every month is the price that attracts members who never log in. Price to the level of transformation you're selling, and accept that it changes who shows up.
Her second filter is the interest list. Before building anything, ask people to raise their hand for free: here's the idea, here's who it's for, here's the problem it solves. No price, no dates. Her challenge to owners who skip this: if you can't get people to raise their hand for free, how do you expect them to pay later?
Then co-create the first version with the people on that list. In her (now closed) First Members program, participants converted 40 to 60 percent of their interest lists into paying members, because those members helped build the thing and already felt ownership when the doors opened. And the first launch itself should be a capped beta: limited seats, unscalable personal attention, members told upfront that the format will evolve. You're not maximizing revenue in round one. You're learning where people get stuck, so round two can cost more and run leaner.
Support That Scales: Narrow AI Tools
The obvious objection to real support is that it doesn't scale. Her answer, five or six years in practice by now, is narrow AI tools attached to specific actions. Not one assistant that knows everything. A tool that helps draft the email the lesson just told you to send. A role-play bot for the hard conversation you're avoiding. Something that gets a member past the blank page or the “is this good enough” stall, at the moment it happens, not three days later on the group call.
She calls the design pattern the learning loop: every training paired with an action. The fillable PDF workbook was the old version of this idea, and she's blunt about it: downloaded plenty, filled in none. The AI version acts more like a coach sitting next to the member, asking questions and brainstorming, with one narrow job and no need to know the whole program.
The usage numbers back it up. One client's members logged 60,000 conversations with their AI tool across 500-something people in under a year.
The deeper reason this works, she argues, is belief. People don't join for three reasons: they don't believe in you, in your offer, or in themselves. A paying member already cleared the first two. The remaining job, the one that decides whether they get results and stay, is helping them believe in themselves.
What Membership and eLearning Operators Can Take From This
Nathalie doesn't run a course business. She builds these systems for a living, and most of what she describes is implementable on any serious membership platform. Three things transfer directly.
Your cancel reasons are probably cover stories. Start tracking what members do and don't do, watch for the questions that repeat on calls, and ask the members who succeeded where they got stuck. The stuck spot is findable, but not by guessing.
The exit is part of the product. Instant cancellation, no email requirement, an optional feedback box, and an automated “sorry to see you go” cost almost nothing and decide what departing members say about you. Some of them come back.
Support is a design decision before it's a headcount decision. Narrow tools attached to specific actions (drafting, role play, implementation prompts) scale the feeling of being coached without scaling your calendar. Start with the one action your members stall on most.
The pattern underneath all of it: members who make progress stay. Everything in the program either moves them forward or it doesn't, and the owner is the only one who can see the difference.
Nathalie Dorémieux is co-founder of The Membership Lab, where she and her husband Olivier design and build membership sites for coaches and course creators. Find her at themembershiplab.com and on LinkedIn.
Running a Membership or eLearning Platform on WordPress?
MemberFix provides managed WordPress operations for membership, eLearning, and community platforms — so your team never has to worry about the tech breaking during a launch or losing members to site issues. We handle development, support, updates, and monitoring with SLA-backed guarantees. Trusted by industry leaders like Echelon Front, TechServe Alliance, and hundreds of others since 2014.
About MemberFix Radio:
We interview experts and share insights for membership site owners and subscription business leaders building sustainable, profitable recurring revenue models.
