Item-item recommendations on non item pages

Question

Theoretical question here. But first, let me say that I have already read https://experienceleague.adobe.com/en/docs/target/using/recommendations/criteria/recommendations-algorithms and still haven't found the clarity I'm looking for.

Imagine that I want to recommend articles to users who visit a non-article page, such as my homepage. My understanding is that I would not be able to use the "content-similarity" nor the "people who viewed this, also viewed" algorithm in this scenario, because the page needs pass in an entity attribute for the algorithm to anchor on. Am I off base here?

However, on a homepage, it would be entirely plausible to use a popularity based algorithm or even a user-based algorithm (such as "recently viewed" or "recommended for you") because these algorithms are not anchored to a specific entity.id. Correct?

So my question is this. What happens if I still try to use the algorithm "People who viewed this, also viewed" on my homepage to display recommendations? What would Target anchor on? Would it gracefully degrade? Would it silently fail?

kandersen-1 · Answer

Hi @maccg12 ,1. No, you’re not off base. You’re spot on. 2. and you’re spot on with the homepage too. 3. as for if you were to use a people who viewed this also viewed that, on the homepage? Nothing. Recs wouldn’t show anything. If there is no entity.id returned in the request nothing will be shown. there are ways to get around this. Example, recs running on an internal search result page typically won’t have an entity.id and you would have to use a popularity algorithm. However, some would identify the search term with a product and set the entity.id based on this product allowing for product-specific recommendations. You could do something similar on the homepage. If you ‘store’ the most recent article id (or category) in local storage, you could set this on the homepage when the user returns/navigate to it. Allowing you to run article or category specific criterias.hope this helps. Happy to specify the exact details for how I would do it, if needed.

Sign up

Login with SSO

Login to the community

Login with SSO

Scanning file for viruses.

This file cannot be downloaded