Streaming Retention Campaign: Who Should Get the Free Month?
A streaming platform has a budget to give away 200,000 free months (cost €1.6M). The data team already has a very accurate churn model and proposes giving them to the top customers most at risk of cancelling. The CFO asks: 'How many of those would have stayed anyway? And how many will leave regardless of whether we give them the month?'. There is a history from a previous campaign with a random control group.
Where the money comes from
What to do now
- 1Ensure a random control group in every campaign.
- 2Build the uplift model with the previous campaign.
- 3Offer only to persuadables with positive value net of cost.
For your committee
One page with the decision, the impact, the risks, the steps and the KPI to track.
The print dialog opens: choose "Save as PDF".
Watch out: more volume is not always more profit
Every free month given to someone who was going to stay is revenue given away. And giving to everyone teaches customers to threaten to leave in order to get discounts.
KPI for the committee
Cost per incremental retention.
Detail for your data team
Proposed technique: Uplift Modeling
Why: A churn model predicts who leaves, not who stays BECAUSE of the offer. Uplift Modeling uses the history with a control group to estimate the causal effect of the incentive on each customer and direct the money to where it changes the outcome.
Cost of getting it wrong: Giving months to people who would have stayed anyway is money wasted; contacting Sleeping Dogs increases cancellations. Both errors are invisible if you only measure the churn rate of the contacted group.
Data needed: Previous campaign: 400,000 customers, 50% received the offer and 50% did not (random), with cancellation outcome at 60 days.; Usage: hours watched, genres, devices, last session.; Billing: plan, tenure, payment failures, plan changes.; Interactions: email opens, ratings, support contacts.