Skip to content
Retention AI
Back to the archive

Episode 101 - You bought personalisation. You never filled it.


Sit through any martech renewal, and you will see the same slide.

Dynamic content blocks. Liquid. Real-time triggers. Catalogue-driven recommendations. AI subject lines. Send-time optimisation. One-to-one at scale.

Every line of it is true. The platforms genuinely do this now. Braze, Iterable, Optimove, Insider, Klaviyo, Salesforce, Adobe, all of them.

Then go and look at what is actually live in the account.

I do this with clients a lot. Put the lifecycle stages down one axis and the customer states that actually matter across the other, and you get a grid. Eighty cells, give or take, depending on the business.

Nine of them are filled.

Not because nobody has worked out what belongs in the other seventy-one. Ask the CRM manager, and they will tell you in about ninety seconds. Ask what the winback should say to someone who churned after one session versus someone who churned after two years, and you get two completely different, completely correct answers.

The answers exist. The cells are empty anyway.

The constraint was never the feature

Personalisation on a platform is a container. Something has to go inside it.

A dynamic block needs eight versions written. A behavioural trigger needs a message that earns the interruption. A recommendation module needs copy that frames what it is recommending, and why now.

So the moment you buy personalisation, you have quietly signed up to a content volume problem your team has no way of meeting.

Brief. Copy. Review. Legal. Localise. Build. QA. Configure. Then again for the next variant, and again the moment one word changes.

A CRM ops lead once told me four hundred emails would take two weeks to configure. That was an honest answer, and probably optimistic.

So the grid stays at nine, the licence gets renewed, and everyone agrees that next year is the year.

Generic AI did not fix it either

Most people reading this have tried. Pasted the email into ChatGPT, read what came back, quietly stopped.

That is not a model problem. Good CRM copy is specific, and a general model holds none of your specifics.

It does not know your lifecycle stages, and what counts as good in onboarding is actively wrong in-life. It does not know which behavioural driver of churn a message is meant to counter. It does not know your voice, your constraints, your rights, your customer.

So it writes the average of everyone else's marketing. Infinite mediocre, and volume was never what you were short of.

Quality is a knowledge problem

Read back over what you would have told me about that winback. Every part of it is knowable, teachable, and already written down somewhere in your business. Lifecycle definitions. Retention principles. Behavioural science. The strategy. The voice.

Which is exactly the kind of thing you can hand to a model.

That is what I mean by a second brain. Not a clever prompt. A set of files the model reads before it writes a word, the way a very good new hire would after three months on the job.

The output is good because of what went in.

And once it exists, the seventy-one empty cells stop being a resourcing argument and start being an afternoon's work.

That is not theoretical. It is how a team of three and a half people shipped four thousand genuinely different communications in thirty-five days for a MENA sports streaming platform, on dates FIFA set and would not move. It is how a subscription hypnotherapy app now runs over seven hundred hyper-personalised messages across channels, driven by subscription status, progress and lapse risk.

Same people. Same platform. Different input.

I'm showing the whole thing on 8 October

Two hours, ten seats, 16:30 UK.

I build the brain live, connect it, and produce at volume in front of you. Then I show the route into your CRM platform, because content that cannot get ingested is not content.

You leave with my files. Thirty-two behavioural science principles, the retention techniques, the headline, explainer and CTA libraries, sixty-two case studies, fifty-one brand examples. Not a prompt pack. The actual knowledge base I use on client work.

It is $197, and ten seats because I want to answer everybody's questions properly.

Reserve your seat

If the date doesn't work, reply and I'll tell you when the next one runs. It will cost more.

One last thing

If you want to keep seeing this in your inbox, just click a link or hit reply with a hello or a question. I read and reply to every message, and it helps signal to your email provider that this is worth delivering. If it isn't for you, unsubscribe with my blessing. I only want to show up where it helps.

Until next time,

Tom

Tom Burrell is a retention strategy consultant with 27 years driving the retention P&L, formerly SVP at DAZN with senior roles at Manchester United and Publicis Groupe. He works with ExCo teams on structural retention strategy and Agentic execution across streaming, iGaming, sports and subscription apps. Founder of Retention AI, the Retention Agency for the Age of AI.

Get the next issue in your inbox

Actionable retention insights: proven frameworks, real-world case studies and Agentic techniques that stand up in the boardroom.