Powering a Grocery Price Comparison App
French grocery price comparison startup
We built the data engine behind a price comparison app that helps consumers find the cheapest groceries near them. 15 million observations per month, collected from heavily protected retail sites across all of France.
The Challenge
The client’s previous scraping provider couldn’t extract data from heavily protected grocery delivery sites. Without fresh data, the app’s development had stalled for six months. The entire product depended on a data pipeline that didn’t work.
The client builds a consumer app that answers a simple question: which store near me has the cheapest groceries? The app lets users build a shopping basket and compare prices across local retailers. The whole product depends on one thing: fresh, accurate price data from grocery delivery sites across France. Their previous provider had been blocked by anti-scraping protections and couldn’t deliver. For six months, the app had stalled. No new data meant no new features, no expansion, no growth.
Our Approach
We built scraping agents capable of collecting product data from every major French grocery drive and delivery platform, including geolocation-aware queries. An AI classification layer organizes products into the client’s taxonomy. Data is delivered continuously via API, feeding the app in near real-time.
What We Built
Nationwide grocery coverage
Scraping agents collect from every major French grocery drive and delivery platform.
Geolocation-aware collection
Prices vary by location. Our agents query stores by geographic zone so the app can show users what’s actually available near them.
AI-powered product classification
An automated classification layer maps products into the client’s category tree, keeping the catalogue structured as it grows.
API delivery
Observations flow to the client via API in near real-time. No batch files, no manual imports.
Continuous expansion
New sites, new delivery conditions, and scope changes are handled as the product evolves.
Results
The system now generates 15 million product observations per month across French grocery delivery platforms. The client resumed product development within 6 to 8 weeks of engagement, and the scope has continued to expand with new sites and delivery conditions added over time.
Before & After
| Metric | Before | After |
|---|---|---|
| Data provider status | Blocked, no deliveries | Continuous collection |
| Monthly observations | — | 15 million |
| Coverage | Partial | All major French grocery platforms |
| Time to first results | — | 6–8 weeks |
| Project status | Stalled for 6 months | Active development, ongoing expansion |