Product Signal Intelligence
An eight-module programme for product, growth, and analytics partners who want AI-assisted product performance analytics grounded in accountable judgement.
Learning outcomes
- Design an event dictionary tied to product questions, not vendor defaults
- Run a weekly signal review that separates noise from decision-worthy drift
- Prompt AI assistants as sceptical second readers of performance summaries
- Produce a stakeholder narrative with evidence, caveats, and a clear ask
Amira Holt
Lead instructor. Former product analytics lead for UK SaaS firms; focuses on instrumentation ethics and narrative clarity.
What you work through
- Module 1 — Questions before trackersMap decisions to measurable signals and refuse orphan events.
- Module 2 — Schema honestyVersioning, naming, and GDPR-aware property hygiene for UK products.
- Module 3 — Cohorts that mean somethingDefine entry events carefully so retention curves stay interpretable.
- Module 4 — Drift vs dramaStatistical common sense for product reviews without becoming a data scientist overnight.
- Module 5 — AI-assisted readingPrompt patterns that surface anomalies while documenting model limitations.
- Module 6 — False positive clinicLab: break AI summaries that overclaim causality from correlational charts.
- Module 7 — Narrative craftWrite the one-page performance memo stakeholders actually finish.
- Module 8 — Ritual designInstall a lasting signal review cadence with your team’s calendar realities.
How seating is listed
Pulse Seat from £420 · Signal Circle from £890 · Observatory Team from £4,200. Full detail on the pricing page. No checkout here.
Practical answers
Do I need a specific analytics vendor?
No. Exercises are tool-agnostic. You need exportable events or warehouse access for the richer assignments.
Will this teach me to build machine learning models?
No — and that is intentional. We focus on reading and prompting existing assistants around product performance, not training models.
What is a real limitation of this course?
We do not deep-dive warehouse modelling or SQL optimisation. If your bottleneck is pipeline engineering, pair this course with a data engineering programme; our strength is judgement and narrative around already-available signals.
How large are live cohorts?
Signal Circle cohorts are capped near twenty learners so critique sessions stay specific.
Module 6’s false positive clinic was uncomfortable in the best way — our AI summary had been blaming a release that never shipped.
Short take: the workbook’s signal review template is now our Monday ritual. Wish there were more SQL examples, but the syllabus warned us.