Flagship course
Retention Grid Lab
Six weeks to put one feature on a Sticky Feature Performance grid, defend the empty cells, and leave with language a sceptical partner can test.
What you will actually do
Each week you bring evidence from a product you own. We do not supply a toy app with perfect events. If your logs are incomplete, that becomes part of the work, not a reason to skip the drawing.
You will present twice: a draft grid in week three, and a defended grid in week six. Presentations are short. The interesting part is the question you cannot answer yet.
Outcomes
- A one-page grid with named return windows and owners for two cells
- A written limitation that would survive a finance review
- An event patch list ranked by whether the grid can exist without them
- Language for “not adopted” that does not sound like failure theatre
Modules
- Week 1 — Jobs and cadences. Write the recurring job without using the epic name. Decide whether weekly actives are even a fair clock.
- Week 2 — Event honesty. Inventory what you log today. Mark events that conflate view, intent, and completion.
- Week 3 — First grid. Plot depth against return. Present the noisiest cell and the emptiest cell.
- Week 4 — Counter-stories. Stress-test whether a campaign, a prompt, or a contract renewal created a fake return.
- Week 5 — Experiments that match cells. Design one change aimed at a specific cell, not at “engagement”.
- Week 6 — Defence. Deliver the grid, the limitation, and the next thirty days of work. No scoreboard.
Instructor
Rhea Calder taught research operations in two GB SaaS firms before opening the Cadder studio. She facilitates the lab, marks grids for hidden uncertainty, and will stop a session if the room starts treating a percentage as a personality trait.
Informational fee
A Bench Seat on this lab is listed at £1,240 per person. Team blocks sit under Grid Fellowship on the pricing page. Payment is arranged after we confirm a place. This page cannot take a card.
Next public cohort is scheduled on weekdays, 10:00–12:00 UK time. Recordings are available for two missed sessions, not as a fully async course.
FAQ
Do I need a data analyst in the room?
It helps. It is not required. You do need export access or a colleague who can pull counts. If you cannot get even rough weekly returns, wait for Instrumentation Weekend first.
Will you tell us the feature is sticky?
No. The lab produces a picture and a limitation. Stickiness is a pattern you keep earning, not a badge we award in week six.
What is a real limitation of this lab?
Six weeks is too short to observe a genuine 90-day return for most B2B jobs. We work with the windows you can honestly see now, and we will say when a “sticky” claim would need another quarter of evidence. If your company demands a lift number for the next board pack, this course will feel incomplete — because we will not invent one.
Can I bring a consumer app?
Yes, if you own the surface. The habit language differs; the grid still applies. We have fewer consumer alumni than workplace-tool alumni.
“Week four’s counter-story session was the part I used on Monday. We had been counting password resets as ‘returns’ to security settings.”
The critiques ran long. I missed two office hours because of a launch. The recordings were dry. I still finished a grid I could put in a steering pack without flinching.