Zalando
increased by 40% — basket sizes
Independently reportedRetail & e-commerce · Personalization · Classical ML · In production
Personalization
Mindvalley · Google Cloud
Vendor-reportedEducationMalaysiaIn productionLLM
Vendor-reported. The customer is named and the numbers are quoted from the source page, but the account comes from the vendor. No independent confirmation.
Reported by google.com (vendor self-report). Checked against the source page on 2026-09-06. 5 of 5 figures below appear on that page word for word.
The problem
Mindvalley faced technology challenges hindering scale and performance, with scattered infrastructure impacting product releases and making business metrics difficult to see.
What was deployed
- Vendor
- Google Cloud
- Products
- Vertex AI, BigQuery, Cloud Run, Cloud Storage, Cloud Composer, Cloud Monitoring, Gemini for Google Cloud, TensorFlow, LangChain
- Models
- Gemini Pro, Gemini 1.5 Pro
- Technique
- LLM
- Build or buy
- Bought off the shelf
- Deployment
- API
- Data used
- demographic information and behavioral data
- Scale
- platform can handle spikes in user demand and scale down as traffic slows
What changed
Each row is quoted from the source. Figures we could not find on the page in those words are marked — they are kept, not deleted, so you can judge them.
increased by 30%Click-through-rate (CTR)year-over-year
“These APIs deploy collaborative filtering and content-based filtering with TensorFlow, Cloud Run, and Vertex AI, and together have increased click-through-rates by 30% year-over-year.”
Quoted word for word from the source · round percentage, no baseline given
nearly 99%ML API endpoints service availability
“On the technical side, for example, Mindvalley has all but eliminated downtime on ML API endpoints and maintains a service availability of nearly 99%.”
Quoted word for word from the source
99.9%ML pipelines deployment success rate
“Additionally, it has a 99.9% success rate in deploying ML pipelines—including both model training and prediction.”
Quoted word for word from the source
reduced by approximately 20%ML infrastructure costs
“By revamping the entire MLOps process, Mindvalley has reduced ML infrastructure costs by approximately 20%, savings it is able to reinvest in more AI-driven technologies.”
Quoted word for word from the source · round percentage, no baseline given
eight percent reductionOverall churn
“The model identified potential leavers, resulting in an eight percent reduction in overall churn.”
Quoted word for word from the source
“Thanks to the seamless interplay of advanced Google Cloud features, ranging from BigQuery to Cloud Run functions and Vertex AI, we are able to spread our mission of lifelong learning and life transformation.”
Difficulties and limits
- technology challenges that hindered its scale and performance
- scattered infrastructure technology directly impacted its ability to release product features
- managing multiple, disconnected learning platforms made it difficult to clearly see its overall business metrics
- increasingly difficult to manage its growing ML infrastructure
Sources
- google.com — vendor self-report
Similar deployments
Intuit
over 65 billion — Machine learning predictions generated
Independently reportedTechnology & software · Personalization · LLM · Scaled
Kingfisher Group
more than 10% — e-commerce sales originating from product recommendations
Independently reportedRetail & e-commerce · Personalization · Classical ML · Scaled · 2023
BABOR · Revieve
Vendor-reportedRetail & e-commerce · Personalization · Computer vision · Scaled
JCPenney · Revieve
Vendor-reportedRetail & e-commerce · Personalization · Computer vision · In production
Paula’s Choice Taiwan · Revieve
104% — achievement on new conversion target
Vendor-reportedRetail & e-commerce · Personalization · Computer vision · In production · 2024