Machine Learning for Growth Marketing
  • Home
  • Start here
  • Schedule
  • Days
    • Before Day 1 · Module 0 (pre-work)
    • Day 1 · Customers and churn
    • Day 2 · What a customer is worth
    • Day 3 · Incrementality
    • Day 4 · Where the next dollar goes
    • Day 5 · Targeting and the plan
    • Extensions · Beyond the core week
  • Notebooks
  • References
  • What’s new
  • Teach

Schedule

Five days, one question each

Each day opens with a warm-up and closes with a wrap-up. A module is a briefing, a lab and a debrief; on Days 2 and 3 the afternoon is an applied clinic tied to that day’s decision. The intro slides open Day 1.

Module 0 · Pre-work: environment, warm-up and coding agents is optional pre-work, done on your own before Day 1. The optional extensions after the timetables are self-paced: they have no times of their own, and a facilitator can swap one into a custom schedule.

Briefing Lab or clinic: hands-on Debrief, warm-up, wrap-up: the whole room Break or lunch

Before Day 1 · Pre-work

Is my environment ready, and can I count purchases per customer in pandas, SQL or dplyr?

On your own, before the workshop: Module 0 · Pre-work: environment, warm-up and coding agents (about 60 minutes).

Day 1 · Customers and churn

Who are our customers, and who is still with us? 09:00–16:20

Time Session Module
09:00–09:15 Warm-up
09:15–10:00 Briefing 1 · Transaction logs to customer-base analysis
10:00–10:50 Lab Module 1
10:50–11:00 Debrief Module 1
11:00–11:15 Break
11:15–12:00 Briefing 2 · Buy-till-you-die models: BG/NBD and Pareto/NBD
12:00–12:55 Lab Module 2
12:55–13:05 Debrief Module 2
13:05–14:00 Lunch
14:00–14:45 Briefing 3 · Covariates in CLVTools, and BTYD versus machine learning
14:45–15:40 Lab Module 3
15:40–15:50 Debrief Module 3
15:50–16:00 Break
16:00–16:20 Wrap-up

Day 2 · What a customer is worth

What is a customer worth, in money, with uncertainty? 09:00–16:00

Time Session Module
09:00–09:15 Warm-up
09:15–10:10 Briefing 4 · Monetary value and dollar CLV
10:10–11:05 Lab Module 4
11:05–11:15 Debrief Module 4
11:15–11:30 Break
11:30–12:25 Briefing 5 · From CLV to decisions
12:25–13:25 Lunch
13:25–14:20 Lab Module 5
14:20–14:30 Debrief Module 5
14:30–15:30 CFO memo clinic: peer review of the Module 5 decision memo
15:30–15:40 Break
15:40–16:00 Wrap-up

Day 3 · Incrementality

Did the marketing cause the sales? 09:00–16:00

Time Session Module
09:00–09:15 Warm-up
09:15–10:10 Briefing 6 · Experiments done right
10:10–11:05 Lab Module 6
11:05–11:15 Debrief Module 6
11:15–11:30 Break
11:30–12:25 Briefing 7 · Geo-experiments and quasi-experiments
12:25–13:25 Lunch
13:25–14:20 Lab Module 7
14:20–14:30 Debrief Module 7
14:30–15:30 Geo-test design review, with the GeoLift market-selection demo
15:30–15:40 Break
15:40–16:00 Wrap-up

Day 4 · Where the next dollar goes

Where should the next dollar go? 09:00–16:35

Time Session Module
09:00–09:15 Warm-up
09:15–10:05 Briefing 8 · Bayesian marketing mix models
10:05–11:00 Lab Module 8
11:00–11:10 Debrief Module 8
11:10–11:25 Break
11:25–12:15 Briefing 9 · Calibrating and validating an MMM
12:15–13:10 Lab Module 9
13:10–13:20 Debrief Module 9
13:20–14:15 Lunch
14:15–15:00 Briefing 10 · Budget allocation under uncertainty
15:00–15:55 Lab Module 10
15:55–16:05 Debrief Module 10
16:05–16:15 Break
16:15–16:35 Wrap-up

Day 5 · Targeting and the plan

Whom should we target, and what is the plan for next quarter? 09:00–17:00

Time Session Module
09:00–09:15 Warm-up
09:15–10:05 Briefing 11 · Uplift and heterogeneous treatment effects
10:05–11:00 Lab Module 11
11:00–11:10 Debrief Module 11
11:10–11:25 Break
11:25–11:40 Briefing 12 · Capstone: one decision pipeline for the retailer
11:40–13:00 Lab · Pair work: CLV forecasts and the geo test Module 12
13:00–13:50 Lunch
13:50–15:30 Lab · Pair work: calibrated MMM, allocation, targeting, the 5-slide brief Module 12
15:30–15:40 Break
15:40–16:40 Debrief · Presentations and scoring Module 12
16:40–17:00 Wrap-up

Extensions · Beyond the core week

What changes when the data are messy, observational, path-shaped or drifting, or when the channel is one you own and can fire every day?

Optional and self-paced, with no timetable: they are not part of the five days. Do them on your own after the week, or swap one into a custom schedule in place of a core module. Each is one module, a briefing, a lab and a debrief:

  • Module 13 · Messy real data: cleaning a real log for BTYD and spend models: 110 minutes (45 briefing, 55 lab, 10 debrief)
  • Module 14 · Survival analysis for subscription churn: 110 minutes (45 briefing, 55 lab, 10 debrief)
  • Module 15 · Causal effects without an experiment: double machine learning: 110 minutes (45 briefing, 55 lab, 10 debrief)
  • Module 16 · Multi-touch attribution: Markov chains, Shapley values and incrementality: 110 minutes (45 briefing, 55 lab, 10 debrief)
  • Module 17 · MMM with drifting effects and a Prophet baseline: 110 minutes (45 briefing, 55 lab, 10 debrief)
  • Module 18 · Notifications: volume, targeting and a policy for the long run: 110 minutes (45 briefing, 55 lab, 10 debrief)

Machine Learning for Growth Marketing · Genial Labs · Built with Quarto

  • Setup

  • FAQ

  • Readiness

Text CC BY 4.0 · Code MIT