Working at Storefront, we’re always excited about new ways we can keep our users engaged. Recommendations or suggestions are a fantastic way for a platform to encourage users to stick around and keep browsing. The problem is, recommendations can be tricky.
How do we know what to recommend to our customers? Is it similar item descriptions? Colors? Locations? It could be anything, and a linear combination of parameters grows significantly in complexity with every additional variable. If we use SQL queries, they can quickly become unmanageable tangled messes of JOINs. These can that take minutes to hours to generate unique recommendations for each and every user across your entire user base.
... read the whole story at medium.com.