Flagship course

Cohort Signal Studio

Eleven weeks of App Analytics as a writing and measurement craft. You leave with a defended event map, a family of retention curves, and a Signal Ledger you can keep without us.

Abstract network visualisation suggesting event streams
Students work from anonymised UK consumer-app exports, not toy CSVs.

Modules

  1. 01

    The event that actually happened

    Naming verbs, killing container events, and writing a dictionary a new engineer can read on a Monday.

  2. 02

    Identity without over-stitching

    Anonymous to known, device resets, shared tablets, and the damage of a greedy merge key.

  3. 03

    Clocks, windows, calendar artefacts

    UTC versus device time, bank holidays, DST, and why “month” is a political word.

  4. 04

    Retention as a family

    N-day, unbounded, bracketed, and resurrection. One curve is never the whole household.

  5. 05

    Funnels that hide drop-off

    Ordering, uniqueness, and time-travel. Conversion theatre versus leakage you can assign an owner to.

  6. 06

    Crash-free as a product metric

    Placing stability beside activation so quality is not a separate religion in engineering.

  7. 07

    Store listing to first frame

    The path store consoles summarise too politely. Impressions, installs, and the silent fail before UI.

  8. 08

    The weekly Signal Ledger

    Five numbers, one limitation, a date. Defence in front of the cohort in week eleven.

What you should be able to do

  • Rewrite an event dictionary so every name is a speakable verb.
  • State the window, clock, and exclusions for any retention chart you publish.
  • Separate listing tourists from resident users without insulting either group.
  • Place crash-free sessions on the same page as activation.
  • Produce a ledger a sceptical engineer will sign.

Informational fee

£1,240 per seat, 2026

Twin Channel (two seats) is listed on fees. No payment is collected here. Enrolment is by correspondence after we read a page on the worst metric you currently own.

Instructor

Portrait of course instructor Meredith Ainsworth

Meredith Ainsworth

Meredith ran product analytics for a London payments app through two store-policy shocks and one painful SDK migration. She still writes event names as if a barrister might read them. She does not teach paid-acquisition bidding.

From people who sat the studio

I wanted Mixpanel recipes. Meredith made me define a session first. The recipes came later and were shorter.

Owen Blake, Android lead, Cardiff

★★★★☆

Homework collided with a release week. I skipped one critique and felt it. Four stars because the ledger still sits in our Monday doc.

Platform-style slip · Instrument Bench 2025

We stopped quoting a single D7 number in board packs. That alone justified the fee, though the identity week still feels unfinished in my notes.

Anonymous client in travel, London

Questions we actually get

Do you teach a specific analytics vendor?

No. Exports arrive as tables. If your warehouse is BigQuery or a CSV in despair, both appear. Vendor UI tours are not the syllabus.

What is a real limitation of this programme?

We do not cover paid user-acquisition bidding, SKAdNetwork postbacks, or media-mix models in depth. Those remain with media teams. We stop at product-side interpretation of attributed traffic, which is a narrower — and frankly more honest — skill.

Can I join from outside the United Kingdom?

Seminars run on UK time. Several alumni join from Ireland and the Netherlands. Residencies on-site are quoted from Worplesdon.

Is there a certificate?

A letter confirming hours, not a badge for LinkedIn theatre. We care that your ledger survives contact with engineering.

Ask for the next cohort