Jesus is a Saviour. That's not a theological claim, it's a casting note. Saviour is a specific role in a specific diagram, and once you've seen the diagram you can run a church off it the way you'd run a Shopify store.
The pinnacle of evangelism is a single sentence: you are a sinner, and you need a hero. Look at what that sentence does to a person. It makes them the one who did the harm and the one who was harmed, both at once, guilty and wounded in the same breath. You cannot stay neutral inside it. There's nowhere to stand.
Then the hero arrives, and the sale closes.
Now the operational problem. A saved person can't stay a victim, because victims burn out and stop being useful. So once you're saved you're promoted. You get to move between two jobs: persecuting sinners, which is to say reminding them what they are, and rescuing sinners, which is to say offering them the hero. You're always doing one or the other. If you're doing neither you aren't being productive for the gospel, and everybody in the room knows it. Every relationship becomes a transaction, with help flowing one way or the other, and friendship is what's left over when the ledger balances.
Which leaves one awkward case. What do you do with someone who doesn't need help?
You persecute them until they do. You don't need saving? Didn't you realise you're a sinner? What do you mean, you're a good person? No, no, no. You can be a sinner just of the mind. If you've ever had an impure thought, you need saving. It's absolutely not good enough. And when you die there will be a judge.
Feeling tired? Weary? That's the devil's persecution. Do not worry. The hero is here.
In May 2009 I typed a small-group lesson for a church of about eighty people, and one line of it reads: you might get them saved, but you'll never see them again. I meant it as a warning about being unfriendly. It was a retention memo. I've told that story once and I'm not going to tell it again here, except for the part I didn't understand at the time, which is that I'd written down a sales funnel and called it a vision.
Three seats, and you're in one of them
The diagram is Stephen Karpman's. In 1968 he published a short paper in a transactional analysis bulletin called "Fairy Tales and Script Drama Analysis", with an upside-down triangle in it and a role on each corner: Victim, Persecutor, Rescuer. He used it to count the switches in Cinderella and the Pied Piper. His point was never the roles. It was the switch. There's no drama, he wrote, until somebody changes corner, and the whole energy of the thing comes from the rotation.
Read a sales funnel with the triangle open beside it and the rotation is the product.
The hook names the pain and seats the reader as Victim. The next line, in nearly every long-form sales video made since 2006, is it's not your fault, which is the most generous sentence in direct response and also the most load-bearing, because it lifts the blame off the reader and leaves it lying around looking for somewhere to go. The next few minutes find it somewhere: the algorithm, the big players, the agency that burned you last time, big food, big pharma, the old way of doing things. That's the Persecutor, now with a face. Then the founder walks in. I was just like you until I found this. Rescuer. The offer is the way out of the seat, for a price, and the renewal is the way back in.
None of those steps is wrong on its own. Plenty of small businesses really were hurt by an update they had no say in, and naming the real cause is a kindness. Eugene Schwartz wrote in 1966 that copy cannot create desire, only channel the desire that's already there, and Justin Jackson says the same thing with a surfboard: businesses respond to demand, they don't make it. The pain is the wave. You didn't make the wave.
So here's the sentence I used to tell myself, and I've seen it in a dozen other marketers' mouths since. We don't create anyone's drama triangle. We just know where it is.
It's an alibi. A triangle has three seats. You can't find a triangle you're sitting in a corner of, and the moment you wrote it's not your fault you took a seat. The wound was the reader's. The triangle is yours.
The wound is theirs. The triangle is yours.
It was written down in 1835
The church didn't borrow any of this from marketing. Marketing borrowed it from the church, and the church wrote the manual first.
Charles Finney was a lawyer before he was a revivalist, and he wrote like one. His Lectures on Revivals of Religion, published in 1835, opens with a heading nobody at the time expected: a revival of religion is not a miracle. It is, in his words, a purely philosophical result of the right use of the constituted means. Use the means correctly and the result follows, the way a crop follows the plough. He then lists the means. Protracted meetings, night after night, so nobody gets to go home and cool down. Plain language. Immediate demands for repentance, because unbelief is a decision and decisions can be made tonight. And the anxious bench: a seat at the front where the person under conviction sits in full view of the room, visibly not yet saved.
He defended the bench in print. Sitting on it, he said, breaks the sinner's pride by getting them willing to be seen as serious, which is an important point towards conversion. Persecute until victim, in public, then rescue. A Princeton reviewer called Albert Dod replied the same year that the bench overwhelms the timid, rewards the showy, and produces converts who then have to save face. He was describing churn.
A century later the method got a department. Donald McGavran's Bridges of God argued in 1955 that mission strategies should be judged by one thing, whether churches grew, and the Church Growth Movement he founded taught that as a discipline with its own journals. In 1980 a young pastor who had studied under McGavran's colleagues spent twelve weeks going door to door in Saddleback Valley, skipping any house that already went to church, and collected the four complaints people had about churches: boring, unfriendly, always asking for money, bad childcare. He built a church that answered all four, mailed fifteen thousand households before its first Easter, wrote up a target customer called Saddleback Sam, and took Peter Drucker as a mentor. Rick Warren's Purpose Driven Church sold a million copies, and its central diagram is five concentric circles, Community to Crowd to Congregation to Committed to Core, with the job of moving people inward. It's a funnel drawn as a target.
So when I typed win, connect, train and send into a cell lesson in 2009, I wasn't quoting scripture. I was quoting a franchise. The four words are Larry Stockstill's, from Bethany World Prayer Center in Baton Rouge, Louisiana, whose book The Cell Church came out in 1998 and whose model came in turn from a church in Bogotá that ran thousands of cells off the slogan win, consolidate, disciple, send. By the time it reached a hall in New Zealand with eighty people in it, the method had crossed two oceans, and I'd have told you it was a vision.
What the back door counts
The retention memo I wrote in 2009 was answering a real number, and the number hasn't improved.
Lifeway asked two thousand young American adults who'd gone to church regularly as teenagers what happened next. Two-thirds dropped out for at least a year. Of those, under a third came back to regular attendance. The reasons they gave weren't doctrinal. The members seemed judgmental or hypocritical. They didn't feel connected to anyone. Pew counted it from the other direction in 2025: for every adult who converts to Christianity in the United States, six leave. And of the people still inside, 86% say they want to serve and 30% volunteered for anything in the past year, which is what burnout looks like when you survey it. You cannot keep a congregation in the Victim seat and expect them to staff the Rescuer seat on Sunday.
Now put the same numbers next to the commercial version.
In the largest study of online courses ever done, across five and a half million learners, 3% of people who signed up finished, and half never opened the course at all. The model works when you don't finish. Dan Lok's High Ticket Closer programme, at up to four thousand dollars, once offered a thirty-day guarantee on the condition that you'd attended every call and submitted every assignment, and the current version lists all digital products as non-refundable. The refund gate and the altar call are the same mechanism: your failure to complete is the proof you didn't really mean it.
HelloFresh in New Zealand made more than a million phone calls to former customers asking for feedback, offered a voucher, and reactivated almost eighty thousand subscriptions that the customers didn't know they'd agreed to. The Commerce Commission fined them $845,000 last year. Amazon's cancellation flow for Prime ran four pages, six clicks and fifteen options, and the engineers who built it named it Iliad. The settlement was two and a half billion dollars. The American regulator's category for all this, business and money-making opportunities, logged three-quarters of a billion dollars in reported losses in 2024 and is the third most-reported kind of fraud in the country.
Different industries, one shape. The customer either gets better and leaves, or doesn't get better and leaves angry. A business that needs you to stay has to make sure you do neither, and the way to do that is to keep you sick enough to need the hero and well enough to pay for him.
The agency that needs you stuck
My own industry's version is politer, and I've sold it.
Take a piece of advice that's now standard in paid media: to win on Meta you need to test hundreds of ad creatives a year. The numbers vary by who's talking. An agency called Common Thread Collective says eight to twelve new ads a month if you're small and forty or more if you're not. A creative-analytics company called Motion studied $1.3 billion of spend and found top advertisers launching a median of thirty-one creatives a week, which is sixteen hundred a year, and that winners are only 4 to 8% of what gets made. One practitioner published his own books: 1,847 ads tested across six brands in twelve months, 94 winners.
Every one of those numbers comes from somebody who sells ad creative, creative analytics or media buying. Meta itself used to say six or fewer ads per ad set was plenty, deleted that advice quietly in early 2025, and has published no number since. The only independent voice I can find argues that volume as such isn't the point at all.
Notice what the advice does, whether or not it's true. Nobody got into business to spin up sixteen hundred ads a year. It isn't a skill an owner can learn on a Tuesday, and it isn't meant to be. The capability gap is the product. The business is cast as a victim of an algorithm that changes faster than anyone can follow, the algorithm is cast as a persecutor with no face and no address, and the agency arrives as the only party who can keep up. The persecutor is kept abstract on purpose, because an abstract persecutor never has to be held to account and never goes away.
Family therapists have a plainer name for the arrangement. Murray Bowen called it reciprocal functioning: one person over-functions, the other under-functions, and each keeps the other in place. As the under-functioner does less, the over-functioner directs more, and the criticism that comes with the directing makes the under-functioner feel less capable, which makes them do less. Bowen warned his own profession about it. When the therapist becomes the healer, he said, the family goes into dysfunction to wait for the therapist. Swap in agency and client and the sentence doesn't need another word changed.
The two-sided name is codependency, and it's a business model. The test for it is simple and most agencies would fail it: if your client became fully capable tomorrow, would they still want you?
There's a version of this that runs inside one head, and it's worth a paragraph because it's where the funnel gets its leverage without supplying anything. In Richard Schwartz's model of the mind there's a wounded part that carries the old pain, a managing part that works to keep that pain out of sight, and a part he calls the firefighter, which arrives when the pain breaks through and offers immediate relief, usually something impulsive. A funnel names the wound, borrows the reader's own critic to do the agitating, and arrives as the relief. It doesn't have to build a triangle. The reader walked in with one, and the funnel just sells to the firefighter.
The honest objections, including the one about me
Naming someone's pain is often a kindness, and the copywriters who defend it are right. A business owner who's been hurt by an algorithm update deserves to hear the real cause, in plain words, from someone who knows. Diagnosis isn't the triangle. The triangle starts one step later, when the diagnosis needs a villain to be complete, and the villain gets appointed on the customer's behalf.
Enemies can be real, too. Andy Raskin, who's built the strategic narratives for a hundred tech companies, is careful never to say enemy. His villain is the old way of doing things, a shift in the world that's happening whether you like it or not, and that's a Challenger, not a Persecutor. The difference is whether the enemy existed before the brand needed one. Liquid Death appointed an enemy for its customers and got 34% awareness and 16% trust, which is what polarisation buys you: the room, not the belief.
The believer's objection is the one I've been waiting for. That's not my church. Conviction of sin is doctrine, not a tactic. And you left, so of course you'd say this. The third one lands, and I've put it in the assumptions below. The first two run into Finney, who wrote the technique down as a technique, defended it as one, and would have been surprised to hear anyone call it a doctrine. This is a critique of a method that described itself. Belief is out of scope.
And then there's the objection the reader should be making by now. You're reading an essay by a man who sells marketing strategy, which has spent two thousand words casting you as the victim of manipulative funnels and is about to arrive as the clear-eyed guide. Every critique of this industry I found while writing this has become its own funnel. The consent-based marketers sell courses on consent. The people telling you to fire your agency sell fractional retainers. The Substacks explaining that coaching is a pyramid scheme sell coaching. I'm in a seat. The only honest thing to do about it is apply the test to myself: does this make you calmer, or more capable, and do I need you to stay stuck?
Rescuer or coach
David Emerald redrew Karpman's triangle in 2005 with the same three corners doing different jobs. The Victim becomes a Creator, who focuses on what they want rather than what's being done to them. The Persecutor becomes a Challenger, the real difficulty that makes you better. And the Rescuer becomes a Coach, who still turns up when you're stuck, and asks what you want instead of saying poor you.
A Rescuer and a Coach look identical in the first week. Both arrive when you're hurting and both have an answer. The difference shows up around month six. The Rescuer still needs you stuck, because every renewal depends on the problem still being there, and the enemy gets a little bigger in every email, because a shrinking enemy means a shrinking invoice. The Coach wants you to need them less, hands over the files and the logins and the playbook, and would count it a win if you didn't renew because you didn't need to.
Rescuer marketing converts fast and churns. Coach marketing converts slow and compounds.
The church that grew knew its back door existed. I know, because I wrote the memo about it. What it never considered was leaving the door open on purpose. The lesson I typed was about how to keep people from leaving. The coach's version would have been about how to make sure that, if they left, they were better for having come. You might get them saved and never see them again. Yes. That's what saving them was supposed to mean.
Four questions, then, for any offer, including this one. Who is this casting me as? Whose voice is the agitation borrowing? When the work is done, am I more capable or more dependent? And does this business need me to stay stuck?
Ask them of the next thing that tells you it's not your fault. Then go and find out whose it is.
Assumptions, stated plainly
I've made a few and I'd rather list them than have them found.
- That a model built for people works on a system. Karpman drew the triangle for families, and I've used it on funnels and churches. I think the fit is exact. It's still a borrowing.
- That I'm a reliable witness. I left the church in the story, and the people who stayed would tell it differently. Chapter 3 is my receipt; it isn't theirs.
- That the ad-volume numbers mean what I've said. Every one of them is vendor data. I've read that as evidence of the model. A vendor would read it as evidence the model works.
- That coaching churn has a number. It doesn't, not a clean one. The course-completion figure and the refund terms are the two ends of the same shape, and I've let you draw the line between them.
- That this essay isn't doing the thing it describes. It might be. The test is above. If it passes, you'll be able to run those four questions without me.
I'd like to be wrong about point 2 in particular. Tell me where it breaks.