Boots
machine learning scoring of offer-usage propensity, applied to mail and the app
Boots' personalized quarterly mailing, where a machine learning model scores each Advantage Card holder's propensity to use each offer, reached 23.9% redemption and +3.2% incremental spend against a control group.
Key points
- Personalized quarterly mailing: an ML model scores each member's propensity for each offer.
- Based on the Advantage Card history (15 million active), mail production with Marketreach.
- 23.9% redemption, +3.2% incremental spend vs control, 18,583 app downloads.
- Evidence level B, confirmed active status.
Objective
Increase Advantage Card holders' spend and engagement by sending each member only the offers they are most likely to use, measured against a control group to prove the incremental effect.
The deployment
Boots leverages the data from its Advantage Card program, which has on the order of 15 million active users. For its quarterly mailing, a machine learning model scores the probability that each member will use each potential offer, and keeps only the most relevant ones for the personalized mail sent to the household. The setup is measured against a control group to isolate the incremental effect. The personalized offers reached an average redemption rate of 23.9%, with 2.4 coupons redeemed in-store per member on average. Compared to the control, the addressed population spent 3.2% more. The mailing also generated 2,745 incremental sign-ups to the Over 60s club and 18,583 incremental app downloads across four sends. In parallel, the Price Advantage personalized pricing setup saved customers more than 12 million pounds.
Results Proof B
Award-winning case study (DMA Awards 2023) documenting a machine learning model and an incremental measurement against a control group, with precise figures. Corroborated by an established press source (CX Network) on the scale of the program and the Price Advantage setup.
How it works
Documented architectureThe stack in detail
- outil Modele de propension a l'offre (maison) Machine learning model that scores, for each member, the probability of using each candidate offer; the exact algorithm is not named in the sources.
- infra Programme Advantage Card (donnee first-party) Purchase history of about 15 million active users, fuel for the scoring and for tying redemption back to the member.
- integrateur Marketreach Physical mail partner: production and sending of the personalized quarterly mailing per household.
How it runs, concretely
For ops teams-
1Building the offer universe Marketing
Marketing defines the pool of candidate offers and brands for the quarter.
-
2Propensity scoring AI
The machine learning model scores, for each member, the probability of using each offer.
-
3Selection and mail personalization Data team and agency
Only the highest-probability offers are kept and printed in the household mailing.
-
4Sending and incremental measurement Marketing and data
The mail goes out, and the addressed population is compared to a control group on redemption and spend.
The purchase history of each Advantage Card holder. Without it, the propensity model cannot sort the offers, and the mailing reverts to a generic send with no measurable incremental effect.
How your customers perceive this type of use
Sourced studiesLe paradoxe est documente des deux cotes : 71% des consommateurs attendent des interactions personnalisees et 76% sont frustres quand elles manquent (McKinsey, 2021), mais 75% declarent ne pas acheter aupres d'organisations auxquelles ils ne confient pas leurs donnees (Cisco, 2024). La « creepy line » est localisee : messages recus quelques secondes apres une recherche et suivi de localisation sont les pratiques qui mettent le plus mal a l'aise (Periscope by McKinsey, 2019).
Acceptance conditions
- La confiance dans le traitement des donnees precede l'achat : 75% ne achetent pas sans elle (Cisco 2024)
- Un cadre legal protecteur rassure : 59% des consommateurs disent que des lois fortes sur la vie privee les rendent plus a l'aise pour partager des informations dans des applications IA (Cisco 2024)
- La personnalisation elle-meme est attendue quand elle est consentie : environ la moitie des consommateurs (US 55%, UK 52%) disent s'inscrire souvent ou parfois a des services personnalises (Periscope by McKinsey 2019)
Red lines
- Le message declenche quelques secondes apres une recherche ou un achat : deuxieme ou troisieme cause de malaise selon les pays (Periscope by McKinsey 2019)
- Le suivi de localisation percu comme de la surveillance : 40% de malaise en Allemagne et au Royaume-Uni (Periscope by McKinsey 2019)
- Le mesusage des donnees personnelles par l'IA, devenu la premiere inquietude des consommateurs, a 53% et en hausse (Qualtrics 2025)
Sources: McKinsey & Company 2021 · Periscope by McKinsey 2019 · Cisco 2024 · Qualtrics 2025
How to replicate
Inference, not sourcedData prerequisites
- purchase history of the card holders
- pool of offers and brands to arbitrate
- ability to tie redemption back to the member
Org prerequisites
- loyalty program with usable postal addresses
- data team for the scoring
- test protocol with a control group
Possible stack
- propensity scoring engine
- personalized mail production chain
- CDP or customer database
- loyalty app
The plan, step by step
- Step 1Define the quarter's universe of offers and brands and extract the purchase history per member from the loyalty program.Deliverable: Member x offer dataset ready for scoring.
- Step 2Train the propensity model and validate it offline: lift of the score-based selection against a random selection.Deliverable: Model with a documented lift curve.
- Step 3Build the mail personalization chain with the print partner: only the highest-probability offers are printed per household.Deliverable: Industrializable personalized mailing mockup.
- Step 4Send to a test population with a control group drawn at random before the send.Deliverable: Campaign sent, incremental measurement plan locked.
- Step 5Measure redemption and incremental spend against the control, then retrain the model with the observed redemptions.Deliverable: Quantified incremental assessment and updated model.
First step: Build an offer-propensity model on the card holders' history and test it on a mailing with a control group.
The same pattern, proven in other industries
Meta
AI feed and video recommendation (large-scale models / LLMs)
Netflix
content recommendation engine
Nike
personalized exclusive offers and access based on member engagement and history
Sources
- S1 Boots: Personalised Advantage Card Direct Mail Case Study (DMA Awards 2023) Interested party archive pending
- S2 Turning loyalty into an advantage at Boots Established press archive pending
An error, newer info, a source?
This page lives on its accuracy. If a figure has moved, if the deployment has changed, or if you have a higher-quality source, tell us. Every sourced correction is verified before publication.