Customer Journey Personalization of Azerty.nl: 15% Sales Uplift

2540 Views

Azerty.nl is a computers & electronics webshop that has been on the market for more than 10 years. The company has numerous rewards for excellence in customer service and is featured in the top 100 webshops list of The Netherlands.

Being passionate about the technologies and innovation, Azerty team together with Retail Rocket launched website personalization project that resulted in significant sales uplift. It included personalization of all significant pages of the website.

Home page personalization

Home page of any webshop is an entry point for the loyal customers – they come to the website directly (because they remember the URL or have it in their bookmarks) or search for the company name in Google. That means that they are very likely to have the website interaction history that could be used for personalization.

The Retail Rocket platform performs real-time analysis of the behavior of all website visitors and adjusts the website so that each person gets their own version of the webshop based on their browsing & purchase history. Here’s how it looked like for Azerty:

azerty-2 azerty-1
Non-personalized version of the home page Personalized version of the home page
for a user who was recently interested
in Samsung Galaxy Tab 4 tablet

Home page personalization efficiency analysis was performed using the A/B testing approach: all visitors of Azerty.nl home page were randomly split into two segments in real time. The first visitor segment was shown the non-personalized version of the main page, and the second was shown the home page with the real-time personalization applied.

The goal of the experiment was to identify the statistically significant difference in the conversion rate between the visitor segments that interact with different versions of the website.

During the A/B test period, Azerty.nl was visited by 113851 unique visitors. Website conversion rate uplift due to home page personalization was +6.6% with the statistical significance over 99%! Conversion rate distribution data:
azerty-3

Category page personalization

Category / catalogue pages serve as the main navigation hub for one of the most important customer journey steps – making a purchase decision. Webshop visitors that are not 100% sure what exactly they want come back to it over and over again. No wonder on average this pagetype is visited at least once in more than 85% of website visits.

The Retail Rocket platform identifies the key features of the products that a visitor is using for his product search and provides personalized product recommendations in real-time. Live example:

azerty-4 azerty-5
Non-personalized version of the category page Personalized version of the category page for a user who was recently interested in
Grafische kaart MSI N750Ti 2GD5/OCV1
for a user who was recently interested in
Samsung Galaxy Tab 4 tablet

As in the home page case, we have performed the efficiency analysis using the A/B testing approach: all visitors of Azerty.nl category page were randomly split into two segments in real time. The first visitor segment was shown the non-personalized version of the category page, and the second was shown the category page with the real-time personalization applied.

This test included more than 180 000 visitors and resulted in 5.7% conversion rate uplift with statistical significance more than 95%!
azerty-6

Overall numbers

As of August 2016, more than 10% of the website visitors interact with 7 personalized product recommendations blocks across all major steps of the customer journey home page, category pages, internal search, product page and the cart page.

Some more examples:

Product page personalization

azerty-7Related products
— Based on visitor’s behavior.
— Products database analysis (categories, properties, etc)

Similar items
— Content filtering (brand, price, color, etc.).
— Collaborative filtering (e.g. items viewed together or bought after viewing this product).
— Real-time personalization based on user interests.

Cart page personalization

Related products to boost AOV
— Based on visitor’s behavior.
— Products database analysis (categories, properties, etc).

As a result — more than 15% of all orders placed through the website have at least one product sold through Retail Rocket’s personalization algorithms.