Rakuten
prediction of future buyers for ad targeting
In 2025, Rakuten's Future Purchase Prediction ad service predicts future buyers from the data of more than 70 services in its ecosystem and cuts advertisers' customer acquisition cost by 50% on average.
Key points
- AI prediction of future buyers for ad targeting.
- Proprietary predictive model fed by data from more than 70 Rakuten services.
- Customer acquisition cost cut by 50% on average across the cases cited.
- Evidence B, status confirmed in Rakuten's October 2025 communication.
Objective
Find buyers that classic ad targeting misses, by predicting who will buy before they show it, and lower the acquisition cost of advertisers who buy space at Rakuten.
The deployment
Future Purchase Prediction is a Rakuten ad service that uses AI to forecast which users will buy. It draws on cross-service data from more than 70 services in the Rakuten ecosystem (Ichiba marketplace, mobile, banking, travel, content) plus advertiser data. The AI identifies demand that human targeters do not see and goes after potential customers missed by conventional targeting campaigns. Rakuten highlights concrete cases, including a home manufacturer and a beverage manufacturer, with an average halving of the customer acquisition cost.
Results Proof B
Figure published by Rakuten in its own innovation communication (interested official source, T2) and picked up by established specialist press (PYMNTS, T4). The 50% CPA gain is an average communicated by Rakuten from client cases, not a result audited by a third party, which places it at B.
How it works
Inferred typical approachThe internal detail is not public. Here is a proven approach that leads to the same result, to adapt to your stack.
The stack in detail
- plateforme Future Purchase Prediction Rakuten ad service that predicts future buyers and targets them for advertisers.
- outil Modele predictif proprietaire Rakuten Machine learning on behavioral data crossed with advertiser data; the exact model architecture is not public.
- infra Ecosysteme de donnees Rakuten Data from more than 70 services (Ichiba marketplace, mobile, banking, travel, content) under a single ID.
- plateforme Regie publicitaire Rakuten Delivery of ads to the predicted profiles across the Rakuten network.
How it runs, concretely
For ops teams-
1Campaign definition customer
The advertiser specifies the product, the target, and the acquisition cost goal.
-
2Buyer prediction AI
The model crosses Rakuten ecosystem and advertiser data to score future buyers.
-
3Targeted delivery AI / ad sales
The ad sales operation serves ads to the predicted profiles, including those missed by classic targeting.
-
4Measurement and adjustment data team
The team tracks the real CPA against the goal and adjusts the targeting scope.
The purchase signals crossed across the more than 70 Rakuten services. If the user is not in the ecosystem or refuses data crossing, the model loses its advantage.
How your customers perceive this type of use
Sourced studiesC'est la famille la moins acceptee : 68% des Americains jugent inacceptable un score financier personnel calcule par algorithme et 67% l'analyse video automatisee d'entretiens d'embauche (Pew Research, 2018). La demande d'explication et de recours est massive : 83% veulent savoir quelles donnees l'IA utilise et 91% veulent pouvoir corriger des donnees erronees (Consumer Reports, 2024). A l'echelle mondiale, seuls 46% se disent prets a faire confiance aux systemes d'IA et 70% jugent une regulation necessaire (KPMG / Universite de Melbourne, 2025).
Acceptance conditions
- Transparence sur les donnees utilisees : 83% des Americains la reclament (Consumer Reports 2024)
- Droit de correction des donnees erronees : 91% le demandent (Consumer Reports 2024)
- Explication de la logique de decision : 44% des consommateurs sont plus enclins a utiliser un agent IA si sa logique est clairement expliquee (Salesforce 2024)
- L'acceptabilite depend du contexte de la decision : 50% des Americains jugent equitable un score de risque criminel pour la liberation conditionnelle, contre 32% pour un score financier applique aux consommateurs (Pew Research 2018)
Red lines
- La decision opaque et sans recours sur l'emploi, le credit ou le logement : 45% tres mal a l'aise pour l'embauche, 39% pour le pret, 39% pour le logement (Consumer Reports 2024)
- Le scoring des personnes a partir de donnees comportementales : 68% le jugent inacceptable pour les offres financieres (Pew Research 2018)
Sources: Pew Research Center 2018 · Consumer Reports 2024 · KPMG / Universite de Melbourne 2025 · Salesforce 2024
How to replicate
Inference, not sourcedData prerequisites
- Broad-coverage behavioral and purchase signals
- Unified customer ID across services
- Advertiser conversion data for training
Org prerequisites
- GDPR legal basis for cross-service data crossing
- Ad sales operation or ad activation platform
- Per-campaign CPA measurement loop
Possible stack
- Predictive model on first-party data
- CDP unifying the IDs
- Ad activation platform
The plan, step by step
- Step 1Audit the data: unified customer ID, usable conversion volume, legal basis for crossing (GDPR in the EU).Deliverable: Data mapping and compliance opinion.
- Step 2Consolidate behavioral and purchase signals in a CDP or feature store under a single ID.Deliverable: Unified profiles activated for training.
- Step 3Train a purchase prediction model and backtest it on the conversion history.Deliverable: Scoring model with offline measured lift.
- Step 4Activate an advertising pilot on the predicted profiles, with a CPA goal and a classic targeting cell for comparison.Deliverable: Test campaign vs conventional targeting.
- Step 5Read the real CPA against the goal, adjust the targeting scope, and industrialize per-campaign scoring.Deliverable: Documented CPA comparison and a reproducible scoring process.
First step: Verify that you have a unified customer ID and a sufficient volume of conversions to train a purchase prediction model.
The same pattern, proven in other industries
Bayer Consumer Health
machine learning demand forecasting on search trends to drive media
Kellanova
predictive scoring of creative effectiveness before delivery
Rovio
LTV prediction to steer acquisition (pLTV)
Sources
- S1 Rakuten's Ad Business Utilizes AI to Identify Future Buyers Interested party archive pending
- S2 Rakuten Puts 'Brain Twin' AI at Center of Retail Ecosystem Secondary 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.