The Genius Link service uses a patented five-step link translation algorithm for iTunes and Amazon, that ensures the best user experience possible. Similar services only translate a link for an international click when the same product ID ("ASIN" for Amazon, or "ADAM ID" for iTunes) exists in both storefronts. We have found, however, that oftentimes, the same item in a different store has a completely different product ID.
That is where our five-step translation algorithm comes into play. The Genius Link service matches on more than just the product ID, digging deeper to compare the metadata available from each product to match across storefronts. By comparing the metadata, we do not need to rely on the Product ID to be the same, and can ensure that your visitors are being sent to the best results possible in their local storefronts.
- The first step we take, is to "Geolocate" the user to find the appropriate Amazon or iTunes storefront (country) to send them to.
- If the user is coming from the same country as the original link, no translation is needed, and they are simply sent to the original link. In the example above, the link was built using the US Amazon storefront, amazon.com, so a user from the US would simply be directed to the original item in the US storefront.
- If the user is coming from another country, we then check to see if the product ID (ASIN or ADAM ID) is the same across storefronts. If the product ID matches in the destination storefront, we simply "localize" the link and send the user to that exact product within their own local storefront. In the example above, the ASIN for that product in Canada matched the original ASIN, so we simply send the user to that same ASIN but within the Canadian storefront.
- Note: To ensure the best possible user experience we even take things a step further and won't send users to an item if it's an "Import" (generally is a higher price than the non-import version of the same product), or if the item is out of stock.
- In many cases, however, the unique product ID does not match across countries. When this occurs, we then search within the Amazon or iTunes APIs to access the metadata for that product (title, brand, manufacturer, etc). We then compare this metadata from the original product to the metadata of products in the destination storefront. If it matches perfectly, we send the user on to that product in their local storefront. In the example above, the ASIN does not match in Germany, but the title as well as the manufacturer are the same as the original link.
Best Match (& Near Match):
- Sometimes, neither the product ID nor the metadata match, which is where we have to get creative. To address this, we work through the meta data to identify the most important information and strip out the remaining. In the example above, we take the title of that product within the destination country, and strip out the "(Retail packaging)" information, which can then be compared against the original metadata.
Decay (not shown):
- For some product types, such as music, books, apps, TV, movies and other "media", an additional step is done. During this step our algorithm will attempt to locate the same content but in a different medium. For example if a streaming movie wasn't available but the DVD version was this result would be chosen.
- Further, if we can't find that product in a related medium, we'll "decay" up a level. This means that a link for a track may decay to a link for an album (if the track couldn't be found) or to an artist page (if the album couldn't be found). For a book that couldn't be found in a local storefront the author page be returned, etc.
- If we are unable to match on anything above, we simply do a generic search within the destination storefront using metadata from the original link. If there are meaningful results, we send the user those search results within their local storefront.
- If no results are returned within the search, however, we simply send the user back to the original link to ensure we never degrade the user experience with an empty search results page.
If you have any questions about our Translation Algorithm, or would like to know more, please feel free to contact us.