Skip to content
Resources

Everything I’ve made public, on one shelf.

Investigations, a class, two guides and a letter — built from the same thinking as the client work, and given away. No templates, no swipe files, nothing recycled from someone else’s playbook.

  1. 5Investigationsproducts taken apart in public
  2. 1Classthree days, by email
  3. 2Guidesone current, one from before
  4. Recent editionsLetteron the post-signup revenue layer
Investigations

Products I wasn’t hired to work on.

Sometimes I take a product apart before I ever touch it. No internal data, no invoice — everything found from the outside, the way any user could. Each one says where it stops being able to tell you anything.

54 pagesNotion · lifecycle autopsy · what I foundHow a product at that valuation fumbles user activation.A full lifecycle autopsy of Notion’s post-signup experience. Every claim traced to something you can watch happen in the product.$10BReported valuation0Internal data used54Pages, every claim traced
KitSix questions in. One email out.Kit asks six questions before you reach the dashboard — then sends everyone the same verification email. Not a missing question; a missing route. Its founder replied in public.GlovoFourteen months of going quiet.I stopped using Glovo on purpose and stayed gone for fourteen months, counting what arrived. Thirty-five reactivation attempts from Glovo. Zero from the others.Risevest67% arrived after 11pm.Thirty emails logged over eight months. Twenty of them arrived after 11pm — some at 1am, some at 2am.ChowdeckTen million orders. 2.1 million users.Chowdeck’s founder published ten million orders and 2.1 million registered users. Divide one by the other: 4.8 orders a year, roughly one every 2.5 months.
All five investigations, with how each one was found →
The class

Three days, by email.

Subscription platform founders and operators who suspect their retention numbers are lying to them.

Framework · The Cohort Inversion Equation
Blended
LTV : CAC≠Cohort
truth

The number the board sees is an average. Averages hide the cohorts quietly draining cash.

  1. 01 Day-7 Activation Bridge
  2. 02 Consumption Velocity Sequence
  3. 03 Silent Churn Deflection
Free class · 3 days · by email

The Cohort Inversion Equation

Why your blended 3:1 LTV:CAC ratio is hiding cohort-level cash leakage — and the three back-half triggers that stop it.

  • Day 1Why blended LTV:CAC metrics obscure where cash actually leaks
  • Day 2The three backend retention triggers
  • Day 3The 90-Day Retention Engineering Blueprint
Start the class ↗
Guides

Two written up, years apart.

The first is current and sits behind day three of the class. The second is from an earlier version of this practice, kept because it is still useful and because pretending it doesn’t exist would be tidier than it is honest.

Guide · 4,000+ words · currentThe 90-Day Retention Engineering BlueprintThe behavioural systems that actually shift a drop-off curve — inbox security deployment, event-driven cohort activation, second-purchase velocity compression — and a 10-point diagnostic checklist you can run against your own stack.Arrives on day three of the class. There is no standalone file.Guide · earlierContent OS: 10 ways to create ridiculously good contentContent frameworks — headline psychology, AIDA structure, USP development.From before the narrowing. Free, no email required.
The Perennus LetterLifecycle intelligence for operators who take retention seriously.Teardowns, system architectures and operator-level findings on the post-signup revenue layer. It goes out when there’s something worth sending.Read the letter →

Something feels off? Let’s find it.

Book a call →View resume →