Engineering Blog
Data Intelligence
Ecommerce product data for AI shopping agents — building the corpus
AI shopping agents need real-time data but scraping for real-time is expensive. What is the minimum viable freshness for a useful agent?.
Estimating inventory from public listings — methods and caveats
Inventory estimates from public data are inherently uncertain. At what point do they become reliable enough to act on?.
From one-off scrape to live API — productizing your ecommerce data feed
The step from a batch file delivery to a live API is enormous in engineering cost. When does it actually justify itself?.
GTIN, UPC, EAN, ASIN, MPN — product identifiers explained for data projects
Which identifier should I use when a product exists on multiple platforms each with their own ID system?.
How much does a one-time ecommerce scrape cost?
Why does one project get quoted $50 on Fiverr and $3,000 from an agency for apparently the same job?.
How to brief a scraping vendor — a one-page spec that gets accurate quotes
I sent three vendors the same vague brief and got quotes from $200 to $4000. How do I write a brief that gets consistent accurate quotes?.