Studio week 3 finally made our call-centre assists usable in the path. Still wrestling with latency, but the schema is settled.
Flagship course
Marketing Attribution Studio
A six-week programme for marketers and analysts who need a defendable answer to “which spend actually moved revenue?”
Learning outcomes
- Select attribution models with explicit trade-offs for your funnel shape.
- Reconcile platform-reported conversions with warehouse truth tables.
- Design a stakeholder readout that separates pacing metrics from planning metrics.
- Document offline assists, brand search inflation, and cookie loss assumptions.
- Propose budget shifts with confidence language finance can audit.
Who it suits
Performance marketers, marketing analysts, and CRM leads who already pull exports — and need a shared method for interpreting them. Prior SQL is helpful but not required; we provide query sketches and spreadsheet alternatives.
Informational fee: £1,150 via Signal Desk · no checkout on this site
Modules
- Week 1 — Signal inventory Catalogue sources, join keys, and known tracking breaks. Produce a health scorecard.
- Week 2 — Model literacy Compare last-click, linear, position-based, time-decay, and data-driven approaches on the same sample journey.
- Week 3 — Offline & assisted paths Bring call, store, and sales-assist events into the path without double counting.
- Week 4 — Incrementality bridges Link geo tests and holdouts to your ongoing model so experiments inform — not replace — attribution.
- Week 5 — Narrative & governance Write the charter: refresh cadence, owners, and when numbers may change.
- Week 6 — Capstone panel Present to faculty plus a finance-minded reviewer. Receive written critique.
Instructor
Amrita Shaw
Former analytics lead for a UK retail media network. Amrita has rebuilt attribution stacks for insurers, DTC brands, and B2B SaaS teams. She teaches model selection as a communication problem as much as a maths problem.
★★★★☆
Capstone feedback on our brand-search inflation section was blunt. We revised the deck twice — worth it before the board session.
FAQ
Do I need a data warehouse?
No. Many learners work from cleaned CSV extracts. A warehouse helps at scale, but the studio methods work in spreadsheets for mid-market volumes.
Is this a tool certification?
No. We are vendor-neutral. Examples may reference common platforms for illustration only.
What is a real limitation of this course?
We cannot grant access to your ad accounts or fix broken tags for you. If instrumentation is severely incomplete, you will spend more time on the diagnostics module and may need engineering support before a sophisticated model is useful. The course will surface those gaps; it will not magically invent missing events.
How do refunds work?
See our Refund Policy. Cohort seats have specific cut-off dates before live sessions begin.