Worminster · measurement school

Incrementality is a claim. We teach you how to earn it.

Techvectorbase trains product, growth, and media teams in conversion lift measurement: designing holdouts, reading noisy results, and writing the memo that survives a sceptical CFO.

Open the Lift Studio syllabus

8 weeksFlagship studio length, not a weekend webinar.

£1,840Cohort seat, billed as a course fee, not a platform licence.

126Alumni now in UK product or media measurement seats.

71Holdout designs reviewed in last year’s studio critiques.

Programmes

Four ways in, one standard of evidence

All programmes
Notebooks and calculator beside printed reports

Short course · 3 weeks

Holdout Architecture for Paid Media

Audience splits, contamination, and the unglamorous work of keeping a control group intact.

Wall of small data charts

Workshop series · 4 sessions

Geo-experiments for National Brands

Matched markets, weather, and why a pretty map is not a design.

Alumni voices

What people actually say after the studio

Longer notes

Module three forced us to throw out a national holdout that looked tidy in the brief. Painful in the room. Correct on the live account six weeks later.

Priya Nair · measurement lead, grocery retail

I still dislike how slow a clean geo test feels. The Lift Studio did not pretend otherwise.

Tom · Leeds

How we teach

Paper first. Platforms second.

Conversion lift measurement fails when teams start in a vendor UI. We start with the claim you want to make, the unit of randomisation you can actually defend, and the sample you can afford to starve.

Every studio critique uses a live-shaped brief from UK retail, travel, or subscription — rewritten so no client is identifiable, but the constraints stay awkward on purpose.

  • You leave with a one-page design, not a certificate wallpaper.
  • We mark the memo you would send to finance, including what you still cannot know.
  • Tutors have run holdouts inside brands, not only in agencies pitching them.
  • GB case law and ICO-adjacent data notes sit in the privacy module because lift tests still move personal data.

Journal

Recent notes from the desk

All articles