Product recommendations. Squeeze more from your online store


    Recommender systems (hereinafter referred to as RS) analyze the interests of users and try to predict what will be most interesting for a particular user at a given time.

    Recommender system identifies the needs of visitors to your online store and at the right time makes interesting is they offer on the site, increasing the income of the online store by increasing conversions, average ticket and the frequency of repeat purchases. According to the results of A / B tests, we can expect revenue growth of up to 50%.

    Such services analyze all available information:
    1. user behavior on the site,
    2. viewed products
    3. order history
    4. information about him from social networks.

    In this article, we consider the main application of recommendation systems - personal product recommendations , as well as two cases of connecting a PC to an online store.

    There are also trigger e-mails, search personalization, and other technologies, which we will cover in future articles.

    How it works

    So how does the PC widget define “who is who”?

    The fact is that the PC is connected to a large data system called BigData, which stores preferences, interests, and other various metrics of specific people. Identification takes place according to the tags recorded in the browser cookies.

    A new visitor on your site may have already visited other online stores before you, made purchases, looked through goods, thus leaving a “mark” behind you. All his actions were saved in BigData under a special tag, and this tag was saved in browser cookies. Even if the user once changed the browser or lost cookies, it is possible to find it in BigData by some parameters of its equipment.

    How it looks on the site

    In the product card in the Similar products block, a block of recommendations is placed, which the system automatically considers suitable for the current product. Thus, we do not have to independently indicate the relationship between the goods, the intellectual system itself forms the relationship of goods among themselves.

    On the main page there is a block of popular goods that are most often ordered in your online store.

    After placing an order and moving to the basket, the system also generates products suitable for the buyer (you may also be interested), which are also available for adding to the basket. Thereby increasing the average check, by adding related products.

    Product recommendations increase conversions in three main areas

    1. Product recommendations make it easier to navigate your site and increase conversions.
    2. Product recommendations form cross-references to product cards, as well as increase site viewing depth and average session duration , which has a positive effect on search engine positions.
    3. Product recommendations allow you to sell more expensive products or related products using the cross-sell and up-sell mechanics.


    Shell Engine Oil Online Store

    What has been done: integration of product recommendations on the main page, in the product card, in the basket.

    Result: during the observation period (55 orders, 7 of which contained recommended products), about 6.6% of the money was brought precisely by the product recommendations service.

    Adult Online Store 18iposle

    What has been done: a block of popular products has been introduced on the main page, a block of recommended products in the product card.

    Result: exactly 50% of orders contained recommended products, 26.3% of the money was brought by the recommendation service. Consumer Goods Online Store

    What has been done: a block of popular goods is connected on the main page, a block of similar goods is displayed in the product card, and a block of goods that may interest a potential buyer is integrated in the basket.

    Result: during the month of work, the store’s revenue growth ranged from 30% to 50% (compared with the previous month). An increase in the average bill was also recorded.

    To summarize

    For PC, the future is unique. This is the WIN-WIN model, when we give the visitor exactly what he is looking for, of that color and size, while at the same time increasing profits from the business.

    For our part, we are ready to discuss with the owners of online stores the connection to the recommendation system, consultation and information support. Interested in? Write to us!

    Some materials are taken from the article.

    Only registered users can participate in the survey. Please come in.

    Do you use product recommendations in your online store?

    • 28% Use on the main page (popular products) 7
    • 36% I use in the product card (they also buy with this product) 9
    • 20% Use in a basket (you may be interested) 5
    • 12% Do not use product recommendations 3
    • 48% I do not own an online store 12

    Also popular now: