Prototype
Building a product comparison prototype built around reliable matching.
The hard part of product comparison is proving two retailer records represent the same thing and being honest about how reliable each source is.
The problem behind the job
If you want to compare the same product across different retailers, matching the products and explaining the quality of each data source matters more than drawing a price table.
What I was working with
I have been developing and testing a product-comparison idea using information from several retailer sources. The useful problem is recognising the same product across those sources and making the confidence of each offer clear.
One product can look different in every source
Retailer titles, URLs and local codes can differ even when the underlying product is the same. The comparison therefore needs a reliable shared identity before the offers can be placed together.
Variants and kits still matter: a similar model name does not guarantee the same package.
Different sources need different confidence labels
The prototype keeps direct and third-party sources separate instead of making every offer look equally current or equally verified.
That distinction is important because a clean interface can otherwise make uncertain data look more definite than it is.
The wider system is still a prototype
The broader idea can accept several structured data sources and update routes. Not every part of that network is a completed production system.
I describe it as an experiment rather than an established comparison service with guaranteed market coverage.
Are you comparing information that doesn't share the same identifiers?
If your problem looks similar, send me the current setup and the outcome you need. The implementation may be different, but the same kind of thinking usually applies.
What I am testing
The useful areas are product matching, the comparison experience and the way different source types are labelled and checked.
Third-party offers are kept separate from confirmed live availability.
Why it is useful work
It brings together data extraction, product matching and automation. It also exposes how easily a tidy interface can hide uncertain inputs.
You can see the other things I build on my projects page.
HAVE SOMETHING AWKWARD?
Are you comparing information that doesn't share the same identifiers?
Send me the URL, what is happening, what you have already tried and what a useful result would look like. A small anonymised sample is enough to start - do not send passwords, credentials or sensitive customer data.