Every shop should know its “Top Seller”. I’m pretty sure that if you work in a shop, you probably know the best selling products. Often you don’t even have to work in the shop to know them, because in the online store they are often advertised directly as “top sellers”. But what circumstances can help certain products to achieve this status?
Similar rules apply in the online store to search engines. In both cases something is searched for and either the searcher quickly finds what he is looking for, changes the search or is gone. Filters in shops can be very helpful here, but products that appear further up will always have an advantage over products that appear further down. A product on page 1 will sell better than one on page 5 and a product on the last page will not be a top seller if it is not advertised separately. Logical, isn’t it? But isn’t this a self-fulfilling prophecy? What do I do if I show 50 dresses per page, but 300 new dresses are put online at the same time in a seasonal change? Which dress should I prefer and show to the customer front and back? By chance, gut feeling or data? If data, which data?
In some cases the sorting of products is also decided strategically and not automatically on the basis of numbers, statistics or algorithms. For example, when running a certain campaign (TV, radio, direct marketing, etc.), the advertised products should be at the top of the list. If you want to make a product attractive to me, I don’t want to have to search for it first, I want to find it directly.
Ok, this point is usually simply skipped because hardly anyone would admit that their products are sorted incorrectly. An example, which I hope does not concern any of the readers, is if you combine the conversion rate, margin etc. to a “business value” and use it for sorting, but overlook or forget the return rate in the calculation. Then you boosted products that sell, but come back like a boomerang. A few years ago we recognized this case at a shop, but never saw it again. Nevertheless, there are often cases where the sorting in the shop is influenced by errors and therefore does not deliver the optimum.
Many things can be identified on the basis of the best-selling articles, the so-called top sellers. A very interesting way to use our system is the visual top-seller analysis. In this case we put all products from a range in relation to each other and see if there are top-seller clusters. For this we use our image recognition, but also other tools such as normalizing the ranking in order to reduce the influence of the placement. For normalization, we use statistical algorithms to compensate for the difference in the number of products between the categories of different shops. For example, if Shop A has 1,000 and Shop B only 100 dresses in its assortment. This allows us to see the influence of the ranking on the products without being disturbed by the different number of products.
With this analysis, overlaps of clusters can be identified that could lead to top sellers. Looking at the clusters in detail, attributes or features can be identified that have the greatest influence on a top-seller cluster. Attributes can be e.g. frills, dots, V-neckline or a certain sleeve shape.
With the Top-Selling Attributes some questions can be answered, e.g:
- What could be the reasons why certain products are top sellers?
- In which categories should which attributes be preferred?
- Which products should I boost in my sort order?
- Which features should be considered for the next collection?
There are certainly a lot of other questions that could be answered with it. If you have an idea, let’s talk about it! We will be happy to help you set up a visually driven Business Intelligence.