Didn’t Google use to have an “inhtml:” operator or something that let you search for matches in the html of pages? Shopify powered websites have commonalities in their html that can give away very quickly the fact to a crawler that it is powered by Shopify, just by looking at the html of the main landing page of the shop.
Also, the DNS entry for a Shopify hosted site, even if using a custom domain, will be either a CNAME pointing to shops.myshopify.com, or an A record pointing to an IP address owned by Shopify, like 23.227.38.65. https://who.is/whois-ip/ip-address/23.227.38.65
We still have a long ways to go, but you can search for items and filter based on things like category and distance (if you are trying to shop locally). We're doing things on a very tight budget at the moment, so sometimes the site goes down, but we're working on it :)
If you have any feedback, we would love to hear it.
Store scams someone and Shopify takes the hit.
Upside of an aggregator is discovery/distribution. Downside is you rely on the aggregator for that discovery/distribution.
Rare for billion dollar businesses to be built on aggregators, not so rare for them to be built on platforms.
Probably hands-down the most irritating metadata issue with Amazon is when sellers don't include quantity information properly - or even include quantity in the title when selling just one size - which I'm sure they do on purpose to make price per unit shopping more difficult.
Sometimes the way the items are listed will be different between one product and another made by the same company, or between Amazon's listing and a third party's listing for the same general product...and they each have different types of the item available. Say, Amazon will sell Ken's Socks in purple and blue, all sizes in purple but only small in blue...while Blankenship E
Things seem to get especially messy when you have a matrix of two 'flavors' (such as a sub-model and color, or unit size), and not all combinations were manufactured or available.
Even the UI isn't consistent. Sometimes it's a drop-down list, sometimes it's the rectangles.
Oh, and often the model/part number isn't entered...