---
title: "Comment by OnlineProxy on \"How Web Scraping Helps Businesses Match Products Across Retailers with and Without GTIN?\" | Hashnode"
description: "GTIN data is a total mess depending on what you're selling. Fashion and apparel are basically ghost towns with 70-95% of GTINs just missing, while food and CPG"
url: "https://hashnode.com/posts/how-web-scraping-helps-businesses-match-products-across-retailers-with-and-without-gtin/6a3296c033faa5b044b1ecda/comment/6a32b00a08fc44270f72d51d"
---

[Discussion](https://hashnode.com/posts/how-web-scraping-helps-businesses-match-products-across-retailers-with-and-without-gtin/6a3296c033faa5b044b1ecda)

Comment [How Web Scraping Helps Businesses Match Products Across Retailers with and Without GTIN?](https://hashnode.com/posts/how-web-scraping-helps-businesses-match-products-across-retailers-with-and-without-gtin/6a3296c033faa5b044b1ecda)

[OnlineProxy](https://hashnode.com/@onlineproxyio)

[Jun 17](https://hashnode.com/posts/how-web-scraping-helps-businesses-match-products-across-retailers-with-and-without-gtin/6a3296c033faa5b044b1ecda/comment/6a32b00a08fc44270f72d51d)

GTIN data is a total mess depending on what you're selling. Fashion and apparel are basically ghost towns with 70-95% of GTINs just missing, while food and CPG stuff is a bit better but still losing 30-50% of the data. Now, if you actually want to build something that works, you're looking at systems that pull off an F1 score around 0.92-0.96 when you throw multimodal AI and human eyes at it together, though you'll probably still get 4-7% false positives that slip through. But if you go the SaaS route, you're dropping 50k-500k a year and waiting 2-8 weeks to go live, whereas building your own pipeline will cost you 300k-1M upfront but pays for itself in 18-30 months if you've got over a million SKUs to manage. To actually make this useful, you'd want a breakdown of GTIN coverage by category, workflows that route matches based on confidence scores, and a simple build-vs-buy comparison that teams can actually use when they're trying to figure out if this investment makes sense for their operation
