In our post “Adaption of Personalization Rules – Pros and Cons” we‘ve mentioned the general pros and cons of adapting personalization rules. Our general recommendation is to let your personalization solution like Recolize do its job and use the personalized algorithms to suggest your products. But we totally understand, that there are occasions when an adaption is absolutely reasonable and needed. Here are 7 reasons to adapt your personalization rules.
Reason 1: Sellouts
If you have to quickly push a sellout to your customers, then you probably want to suggest these specific products to them. In this case it does not matter, if these products are slow-moving products, because you have to sell them anyway.
Reason 2: Seasonal Changes
If you have seasonal collections or switch your stocks regularly, the personalization solution will be confused for some time. The reason is, that huge amounts of products are completely new. The personalization solution will need some time to learn and determine the best matching personal products.
Reason 3: Boost Specific Products
If you want to boost specific products, then you can change the underlying logic of your personalization rules. Sometimes this is necessary because of e.g. great margins of these specific products. If you earn more through the sales of these products than you would normally earn through the recommendations, then boosting specific products can result in more revenue.
Reason 4: Push New Products
New products very seldom directly appear in your recommendation widgets. As already mentioned in reason 2, it takes some time until new products get recognized by the personalization logics. Depending on how successful the new products are, this time can be short, but in some cases new products have to appear immediately in recommendations to push them even faster to the top of all lists. Therefore it can make sense to adapt your personalization rules in this case.
Reason 5: Blog Post Usage
If you are blogging about specific products, categories or product series, then in these blog posts you probably also want to link to the according shop pages. You can use recommendations for this use case easily, but normally these recommendations don’t have to do anything with the contents of your post. To change this, it can make sense to adapt the recommendation rule to just display products from a specific category or to even limit it to specific SKUs. Of course, the limitation of SKUs avoids nearly all personalization logics behind the algorithm. But for this specific use case it makes totally sense.
Reason 6: Category-based Recommendations
A quite common use case is to display recommendations from the same category. Many online shop owners want to display products from the same or similar categories. This is especially reasonable in online shops with lots of different kinds of products like e.g. market places. The customers often decide to browse in a specific category and displaying other personalized products from the same categories can direct them to the right products.
Also possible is to display products from a chosen other category. This is often the case in clothing online shops where the shop owners display fitting accessories after e.g. a woman’s dress is added to the cart. Therefore they filter for the category “accessories” and only recommended products from this category are displayed.
Reason 7: A/B Testing
A/B tests are a great way to test your recommendation rules. Compare your manually adapted rule against the automated rules of the recommendation engine. After some time you will get your answer: is your manual rule converting better than the automated one? Afterwards you can decide which rule you want to use and what the goal with this rule is. Of course, if you have to fulfill one of the use cases mentioned above, you are not always free to decide to use the automated rule.
Think about this: if the automated rule is converting better, then does it really make sense to overrule it? In most cases the answer is no. Even in very hard cases like blog posts, where you really want to feature specific products, it can make more sense to display personalized products instead of the few featured ones.
The reason is, that the visitors of pages like the blog and also all the other ones don’t think like you do as a shop owner. Most of the traffic didn’t even want to visit these specific pages. They are landing there from search engine results, performance marketing channels or paid search ads. Therefore personal recommendations make much more sense, because in lots of these cases the visitor is not interested in the content of the current page. And the personalized recommendations then guide him to the really interesting content for him.
For a starting point: offer more than one recommendation widget on these pages. Your manual adaption and a really personalized one. And after some time you can compare the results. If the automated logics win, then you know, that you don’t have to adapt your personalization rules in most of the cases.