Online recommendation – can the algorithm replace the seller?

As we know, personalization is one of the major challenges of e‑commerce, in order to boost customer satisfaction, to improve loyalty and to generate additional sales. Regardless of their industry or the size of their structure, 75% of marketers questioned as part of the survey conducted in 2014 by Adobe on digital marketing optimization are convinced that personalization plays a key role in the achievement of long-term business objectives. In this context, the recommendation is one of the most powerful marketing tools and is becoming more and more essential on retail websites, putting back the customer at the center of the sales process. Discover in this article the objectives of the recommendations and their benefits when they are implemented on a retail website.

The recommendation in marketing, yesterday and today

The aim of the recommendation is to help the client to choose, by offering products matching their needs or their tastes. Traditionally, this recommendation function is provided by the sellers in the stores. Let’s look at a simple example and let’s assume that I need a tent for my next camping holiday. I go to a specialized supermarket, in the tents section: if no vendor is available to direct me, I will choose a tent by myself, and then will quickly be heading to the cash desk. On the contrary, in the presence of a seller, I would be able to receive personalized advice, a recommendation adapted to my needs, and I would also probably be suggested additional purchases such as a sleeping bag or a floor mat. The benefits of the recommendation are clear: it increases the average basket while increasing the perception of a successful customer experience.

At the moment, in-store sales are still much higher than those conducted online, partly because of the possibility of personal advice. The introduction of dynamic, automated recommendations on your retail website is therefore a real need, in order to accelerate sales and profitability. I have already said it, providing quality customer experience is paramount. And in a world where supply is more overabundant than ever, where the slightest research can take hours and where immediacy has become essential, making user’s purchases easier online by immersing them within a specific product universe is a great way to enhance their customer experience. Today, technology allows marketers to, through the segmentation of audiences, the tracking and the analysis of behavioral data, offer always more personalized recommendations, therefore always more relevant. This is where the algorithm can, if not replace, at least become the equivalent of the seller, mixing customer knowledge, test and dynamic recommendations of product, content, and services.

The different types of recommendations

Recommendations can achieve different objectives: discover products, offer alternative products (such as a different model of tent if we use again the example of the camper client), offer products better suited (such as a family tent I have children), or lead towards complementary products. The aim is to improve the customer experience by putting him in contact with products that he wouldn’t have thought of on his own.

To achieve this, one of the most effective solutions is to rely on behavioral analysis, using an algorithm that will analyze visitor behavior and will either create audience segments based on this analysis, or will enrich pre-existing audience segments (created via Adobe Analytics, for example). With a tool like Adobe Target and its Recommendations module, you will thus be able to link the behavior of audience segments with the various existing products and services, in order to dynamically recommend the product best adapted to a specific audience.

Adobe Target is used by Sony to escort its customers during their journey on the PlayStation Network (using a PS4 or PS3), but also on his mobile and website. Elliot Dumville, Director of Product Management E‑Commerce at the Sony Entertainment Network, discussed at the Adobe 2015 Summit of Salt Lake City how the Adobe Target recommendations feature allows them to considerably increase the average basket of their visitors, and therefore their global revenue. The customer visiting one of its spaces is then offered each time a complementary product, an extra service, or a product that they are likely to help given their initial choice (People Who Bought This Also Bought). If you are interested to learn more about how PlayStation is using Adobe Target to recommend relevant content throughout the customer journey, I can only advise you to have a look at this video:

Video: Rethinking recommendations: A look at the NEW recs in Adobe Target Premium

In conclusion, the recommendation should be based on a solid understanding of customers and a careful data analysis, hence the need for an extremely powerful algorithm, a reliable technology and efficient tools such as Adobe Experience Manager or Adobe Target. However, when all these elements are combined, the results are generally significant, and allow the retail site to increase its average basket, to boost its conversion rate, and to improve customer satisfaction. Again, it comes back to the personalization of the customer relationship, which is fundamental for success in e‑commerce today: companies using visitors targeting multiply on average their conversion rate by 2 (White Paper Adobe — Results of the 2014 study on the optimization of digital marketing).

What about you, what do you think of online recommendations? Have you had the opportunity to test their performance on your website? Feel free to share your experiences in the comments section.