Recommendation relevance is a data problem before it’s an algorithm problem. To improve it, work through five inputs in order: identify more of your visitors so their behavior actually gets recorded, clean up your product catalog (categories, tags, attributes — that’s the vocabulary the engine thinks in), capture browsing and cart activity rather than just […]
Category Archives: Personalization and Product Recommendations
The working split: automate the decisions that repeat per customer — which product to show whom, when a behavior should trigger a message, which segment someone belongs in — and keep human control over the decisions that define the business: what may never be recommended, how far personalization is allowed to go before it feels […]
A personalized journey across email and SMS means one sequence, driven by one customer profile, where each message goes out on the channel suited to its job — not two parallel channels sending the same things twice. Email carries the content: product blocks, education, comparisons. SMS carries the moments: the back-in-stock alert, the “your cart […]
Personalized recommendations convert better than bestseller blocks — but only for contacts you actually have data on. For a subscriber with browsing history and past orders, a block built around their behavior will almost always outperform a generic “our most popular products” grid, because it answers the question they’re already asking. For a brand-new subscriber […]
The fix is a suppression rule: your recommendation blocks should automatically exclude any product the customer has already purchased, with one deliberate exception — consumables and other products people buy again on purpose. Most email tools with a product recommender can do this, but many stores never switch it on, so the follow-up email proudly […]