Public teardowns
Retention setups, pulled apart.
Unaffiliated analysis of real brands running real subscriptions — the parts anyone can observe from the outside: post-purchase email, cancel flow, reorder timing, and which brand five different models name when you ask them to recommend one. We have no client results to show yet, so there are none on this page. When there are, they will arrive with their baselines attached.
Published
Nothing published yet
The first teardowns are still being researched. Until a real brand’s flows have been subscribed to, received and documented end to end, there is nothing here worth reading — and we would rather leave the page honest than fill it with a number we have not measured.
The data we do have is published with its sources on the benchmarks pages, and the calculators will put a figure on your own numbers in about two minutes.
In the queue
These are scaffolds, not teardowns. No brand has been selected, no subscription has been bought, and every section is a list of the questions the finished piece has to answer. They are linked here so the method is reviewable before the findings exist.
-
Draft scaffold Subscription brands
The month-three supplement cliff
An unaffiliated teardown of a supplement subscription — post-purchase email, cancel flow and reorder timing, read entirely from the outside.
Brand: [Brand to select]
Read the scaffold → -
Draft scaffold Subscription brands
The second box that never ships
An unaffiliated teardown of a frozen meal subscription — the transition weeks, the second delivery, and everything the brand sends in between.
Brand: [Brand to select]
Read the scaffold →
How a teardown is done
There is no privileged access behind any of this. A teardown is what a paying customer — or anybody with a chat window — can see, documented properly and read by someone who builds these systems for a living.
- Subscribe like a customer We buy the product on the same subscription a customer would, from the same storefront, and let it run. Everything on the page comes out of that account — never out of a brand’s admin.
- Time every message Each email and SMS from order confirmation to the first reorder prompt, logged with the day it landed, screenshotted, and put back in order.
- Walk the cancel flow All the way to the exit, recording what is offered on the way: pause, skip, cadence change, smaller size, a reason picker, a discount — or nothing at all.
- Check the reorder window The date the prompt arrives against the date the product actually runs out, at the dose, portion or usage the brand itself prints on the pack.
- Ask the models The questions a buyer in that category actually types, put to ChatGPT, Gemini, Perplexity, Claude and Google AI Overviews, with each answer logged three ways: the brand named, not named, or named alongside which competitors. Every URL those answers cite gets captured too, because that list is where the mentions are actually coming from. It is the one part of a retention setup that is public by construction — the model will tell anybody who asks.
- Write it constructively The brand gets right of reply before anything goes live. These are systems problems, and almost every brand at this size has the same ones.
What a teardown can’t see
Saying this out loud is the difference between analysis and guesswork. From outside a business we cannot know:
- The retry ladder — how many times a declined card is retried and over what window. We only see the dunning email, if one arrives at all.
- Real churn, split into voluntary and involuntary. Nobody publishes it.
- Flow revenue share, list health, and everything else that lives inside the email platform.
- Whether being named by a model turned into an order. We can see which brand the answer recommends; we cannot see the conversion rate, AOV or subscription attach rate on AI-referred sessions — and that is the number that decides whether working the citations is worth anyone’s budget.
- Which manual work behind the storefront should be automated and which should not. From outside you can tell that a person is doing a machine’s job; you cannot tell what it costs, or what breaks if a system does it instead.
- Which of the gaps is worth the most money. Ranking them takes the brand’s own revenue data.
That last one is the whole difference between a teardown and the Revenue Recovery Audit. The audit reads your accounts, so every gap comes back with a dollar figure next to it instead of a question mark.
Find out what you’re losing.
Before you commit to anything, we tell you exactly what you’re losing and what it costs to stop it. Two weeks. Fixed fee. Credited in full against any build you go ahead with.
- Fee
- $1,500–$3,000, fixed
- Duration
- Two weeks
- Credited
- In full, against any build
- You supply
- Read access + one 45-minute call