What we did:
Working together, Pythian helped the grocer to:
- Set the right incentive strategy. Pythian’s data scientists helped Schnucks develop a machine-learning model to predict shopper likelihood to buy certain products. These insights drove other campaign parameters.
- Launch in the cloud. The modern, scalable GCP platform gave Schnucks computing power and ease-of-use. It offered far more flexibility than creating a custom, on-premise solution.
- Test and learn. The 12-week proof of concept gave the Schnucks team the opportunity to learn the platform and absorb Pythian’s analytic approach in a collaborative environment.
With the subject in place, Pythian and Schnucks worked quickly to launch the proof of concept. This included:
- Building an algorithmic, machine learning model in GCP. The model predicts the likelihood a customer will buy certain products, based on their individual shopping history.
- Identifying Own Brand products to recommend to each customer, based on their unique results in the model.
- Distributing coupons, special offers and other promotions to the test group during each week of the 12-week campaign, via Schnucks Rewards, the company’s loyalty program. (The control group received no recommendations or incentives.)
Technologies used:
- Anthos
- App Engine
- Apigee
- BigQuery
- Cloud Logging
- Cloud Run
- Cloud SQL
- Composer
- DataFlow
- DNS
- Google Cloud Storage (GCS)
- Google Compute Engine (GCE)
- Google Kubernetes Engine (GKE)
- IAM
- Natural Language API
- Pub/Sub
- Vertex AI
- VPC