Image Similarity, Personalisierung, E-Commerce, prudsys, GK Software

The fashion industry presents online retailers with special challenges and opportunities. The industry is characterized by a high level of dynamism due to constantly changing seasonal fashions and the need to offer clothing that is “in fashion.” In addition, customers have their own tastes when it comes to pattern, cut and style.

For fashion retailers, it is therefore important to offer customers the most suitable garments from the abundance of choices available. Relevance for customers in the form of personalized offers is one of the key factors for successful sales. The increase in mobile shopping only reinforces this requirement. Whether via the browser or in the app, the limited space on the smartphone demands relevant offers that convinces customers to add to cart.

Even minor adjustments within the shopping process can lead to a better shopping experience and help to increase sales. In this blog post, you will learn how to score points with your customers through image similarity, or product recommendations calculated on the basis of image similarities, on the item detail page, in the shopping cart and on the wish list.

1. Image similarity on the article detail page for the best purchase option

After clicking on an interesting product, customers are taken to the item detail page. Here they find more information about the garment and a detailed view, usually through various product images, videos or 3-D images. Many retailers place product recommendations under the garment, based on different recommendation types. A popular recommendation type – especially in the fashion industry – is sending recommendations calculated on the basis of similarity analyses. The customer receives suggestions for other pants that are similar to the pants they selected in terms of shape, material and/or color. Many customers use this “service” because:

  •  They like to browse when choosing an outfit.
  •  They’re disappointed if their desired clothing size is sold out, so this is mitigated by finding visually similar garments to choose from nearby.

Retailers themselves can determine which products are included in the calculation of recommendations, e.g. only from a certain category or a certain brand world. Our AI solution keeps an integrated test drive ready. This allows retailers to easily compare different configurations.

Image Similarity, Artikeldetailseite, prudsys, GK Software

2. Image similarity in the shopping cart to compensate for lost sales due to sold-out items

Once customers have decided to buy an item of clothing, they place it in the shopping cart. Now, if possible, customers should not be distracted by anything else to keep them from abandoning the purchase. Special promotions, surveys or product recommendations are no longer necessary at this point.

However, as we all know, exceptions prove the rule: if customers do not complete the purchase and return to the shopping cart sometime later, the frustration can be great if the item is sold out in the meantime.

To counteract the threat of lost sales, it has proven practical to display alternative products that are visually similar to the sold-out item of clothing. Experience shows that a considerable proportion of customers choose to add one of these recommendations to their shopping cart.

Retailers define the filter criteria for calculating the recommendations. For instance, they can ensure that only items displayed are available in the clothing size of the sold-out product.

In practice, retailers vary in how prominently the recommendations based on Image Similarity are displayed in the shopping cart. Some show the recommendations directly as product images under the sold-out garment, while others allow the customer to activate the display of the recommendations by clicking on a button under the sold-out item. Then, retailers usually choose a headline for the recommendations in such a way that customers can see at a glance that these are similar products that might interest them.

Personalisierung, Image Similarity, Warenkorb, E-Commerce, prudsys, GK Software


3. Visually similar recommendations for alternatives on the wish list

A wish list makes it easier for customers to remember interesting items of clothing they may want to buy later. Fashion retailers see this service as a preliminary stage to the actual buying process. Unlike in the shopping cart, where the customer should no longer be distracted, inspiration is still possible here, for example, in the form of product recommendations or special promotions. But even in the wish list, garments are sold out after a certain time.

As in the shopping cart, product recommendations for visually similar items of clothing in the wishlist also give retailers the opportunity to reclaim sales potential they thought they had lost. Some customers will decide to bookmark a displayed alternative for a planned purchase.

And here, too, retailers use the various filter options for displaying recommendations. For example, some online store operators want to display visually similar items of clothing only in a certain price range, based on the price of the original product.

The users of our AI solution like to monitor which recommendations are best accepted by their own customers in the test drive. This provides the option of visualizing the image similarities in a 3D view, which gives retailers a better overview in terms of analyzing results. Within the 3D view, users can interactively move around and apply a search function.

Test Drive Image Similarity, Personalisierung, prudsys, GK Software

Are you ready for Image Similarity?

Fashion retailers are not the only ones who benefit from product recommendations based on similarity analysis. In addition to the use cases presented, retailers are implementing a variety of other Image Similarity strategies.

Are you considering using Image Similarity in your online store? I would be happy to advise you! Make an appointment with me or take part in one of our next webinars to learn more.

Mr Personalization