In one of the recent articles we’ve covered what does Self-learning Search means and uncovered how
Search optimizes order and suggests more relevant items for a visitor. It's a very efficient
technique, and brings interested items to a visitor first. But Kea Labs search make step further -
it’s a profit-oriented search, which orders products based on the chance of the purchase and a
potential profit of transaction.
Let’s unveil some of the algorithms behind this!
Kea Labs combines various calculated metrics to keep balance between the profit and the chance of the purchase. It’s not a straightforward task, and the importance of each coefficient is getting tuned based on the historical sales performance.
How search evaluates potential profit
Shortly speaking, profit is the difference between earning and the costs. On the level of
individual product we may look on the margin which equals to product price minus product farm
But it will be incorrect to pull only high-margin products on top. As search will be unrepresentative and prices may be higher than competitors have.
Furthermore it doesn’t keep relations between product margins. It’s very often when accessories have huge margin, but the general item has very low(and sometimes negative) margin. For example, a mobile phone and a suitcase or screen protection glass.
Having low margin on the general product helps in keeping price competitive, and profit is coming from the accessories. Kea Labs may detect relations between group of products and smooth the way it pushes high or low-margin products.
Unfortunately, farm price is not available by default in many platforms, including shopify, so please contact us if you need any help with exporting it.
Popularity and trends
Thanks to a product comparison algorithm used by Kea Labs Advanced Recommendations, Kea Labs
Search detects popularity not only for individual items, but also for a group of similar
As a result, Kea Labs Search detects trends and promotes even new, or not so popular items of the group. This is really useful and solves issues when you add new product and visitors can’t find it as nobody viewed or purchased it yet.
Trends automatically detects which group of products needs to have higher priority. For example, if a visitor searches for ‘jacket’ on winter, search detects the rising interest for warmer jackets and pulls them up.
Quality of content
It’s simple, but many stores are forgetting that search need to know the quality of product content. One of the key goals of the search is to take user’s attention and show attractive products. And quality of content, like presence of a picture or video, detailed description, product attributes, affects the impression. Search needs to evaluate the quality of the product content and pull up the most attractive ones to keep visitors attention.
Number of available products
Based on the historical transaction search evaluates average number of bought items per product,
and checks how many items are left. Based on this information search may either push product to be
sold out(and release free space on your warehouse) or promote other alternatives, having enough
Of course, if the product is not available, it’ll have higher priority that alternatives available for purchase or for preordering.
Your marketing goals
Algorithms are good, but no one better than you may know goals of your business. Having control over the algorithms and ability to implement your marketing strategy is vital. Kea Labs Search is totally controllable and it’s a powerful tool for your marketing team. We’ll cover the marketing features in one of the nearest articles.