E-book – The Comprehensive Recommendation Engine Guide
Learn more about recommendation strategies, areas of application, personalization, and success factors for product recommendations, as well as successful use cases in this guide.
A recommendation engine provides inspiration in your online shop and increases the value of shopping carts
Did you know that...
35%
of Amazon's sales are generated by product recommendations – clear evidence of the effectiveness of recommendation engines (McKinsey via ameo-agentur.de, accessed in 2024).
299%
ROI achieved by companies in 3 years that invest in advanced personalization technologies such as recommendation engines (Forrester, 2022)
26%
of consumers in online shops would like to use purchase recommendations in online shops. (Source: Bitkom Research, 2020)
47%
Users find it useful to receive suggestions for similar products in online shops. (Source: Statista, 2017)
A recommendation engine triggers buying impulses in the purchase phase.
A recommendation engine is a software service that displays relevant recommendations in an online shop, providing inspiration during the purchase phase. To do this, the click and purchase behavior of each shop visitor is recorded in an AI-supported knowledge base and processed into a history. Self-learning technology is used to display product or content recommendations to shop visitors that match their interests. Various types of recommendation systems are used for this purpose, ranging from behavior-based to context-based approaches. This creates purchasing impulses that increase the value of the shopping cart. This is exactly what epoq Inspire offers.
Personalization is the key to relevant recommendations on every store page
Our recommendation engine generates an AI-powered knowledge base to calculate individual recommendations for each customer – tailored precisely to their behavior and preferences.
This inspires your customers in a targeted manner during the purchase decision phase and creates effective buying impulses. For example, a customer looking for "Tom Tailor" pants will be recommended other matching products from this brand. If they add pants to their shopping cart, we will show them matching T-shirts from their favorite brands.
This turns every touchpoint in the store into a personalized experience, which increases the shopping cart value and boosts the conversion rate.
Watch our short explanatory video about the Recommendation Engine
You are currently viewing a placeholder content from Default. To access the actual content, click the button below. Please note that doing so will share data with third-party providers.
More Information5.04% increase in revenue per session thanks to new personalization strategy of the recommendation engine
"Our premium and luxury customers expect personalized offers and recommendations—especially from us as a closed online store—based on brand preference, their own size, taste, etc. With Epoq as our partner, we are constantly finding new approaches that we implement agilely and optimize together."
Dr. Stefan Hoffmann | Managing Director | Outletcity Metzingen
Merchandising control: Our recommendation engine follows your business logic
Our recommendation engine is based not only on your customers' interests, but also on your specifications. With flexible merchandising options, you can set rules and priorities to tailor recommendations to your business goals. Whether you want to promote specific products, take stock levels into account, or make strategic placements, the combination of intelligent personalization and manual control makes our recommendation engine a powerful tool – perfectly tailored to your product range, your goals, and your KPIs.
Prioritize
Determine which products, brands, or categories should be displayed preferentially and highlighted specifically.
Business Objectives
Define which business objectives – such as inventory clearance, margin strength, or new products – should be taken into account when displaying products.
Rules
Define which rules should apply to specific use cases or target groups.
Customer targeting: Our recommendation engine also thinks in terms of target groups
Our recommendation engine doesn't just think in terms of products – it also thinks in terms of target groups. You decide who gets to see what: personalized, segmented, or according to strategic priorities.
1:1 Personalization
Our core competence is 1:1 personalization – we offer every user a tailor-made shopping experience in real time. From the very first click, we analyze behavior and immediately deliver relevant recommendations, without the need for a profile. For returning customers, we use previous purchases and interactions to tailor personalization even more precisely to individual needs.
Segment-based Targeting
Segment-based targeting specifically addresses groups with similar interests, such as "outdoor enthusiasts" with hiking equipment. Thanks to 1:1 personalization, segments are automatically created from individual purchase and behavioral data, allowing campaigns to be tailored even more precisely to customer needs.
Product-based Targeting
For a single product – such as a newly released book – as well as for a category or brand, customers who are identified as potential buyers for that specific book based on their purchasing behavior are targeted. Thanks to 1:1 personalization, these target groups can be automatically derived from individual data, allowing campaigns to be tailored precisely and effectively to exclusive customer needs.
The most popular use cases for a recommendation engine
Inspire your customers with a click on the shopping cart button
Once an online shopper has decided on a product, you can inspire them with additional products when they click on the shopping cart button. The relevant recommendations are personalized and displayed in milliseconds. This increases your shopping cart value and sales. Ankerkraut was able to rapidly increase its shopping cart value by 6% as a result.
- Additional products appear when you click on the shopping cart button
- Relevant recommendations are calculated and displayed in milliseconds
- Shopping cart value and sales are increased
Enable your customers to purchase a complete package
Offer your online shoppers the opportunity to quickly put together complete product sets. This allows them to create a stylish outfit or compatible snowboarding equipment, for example. The set can be purchased immediately or saved for later. This gives online shoppers inspiration for compatible product sets directly on the product detail page and also saves them time. This increases shopping cart value and sales.
- Customers can quickly click together product sets
- The set can be purchased directly or saved
- Time savings for online shoppers
Meet the individual tastes of your shop customers right from the start
Online shoppers want to be inspired right from the start page – it is important to cater to the individual tastes of visitors as soon as they enter the online shop in order to inspire them. Visitors are not always known – in this case, top sellers or auto-generated recommendations based on the behavior of all visitors as a whole can help.
- Online shoppers are inspired right from the start page
- If the visitor is unknown, top sellers can be used
Let your shop customers make spontaneous purchases
What works in brick-and-mortar retail can also be used online. The checkout area and spontaneous access to additional items can be represented via relevant recommendations in the shopping cart. This fills the shopping cart and increases sales.
- Spontaneous purchases can be tracked via the shopping cart
- Products that are cheaper should be displayed
Offer your customers related products
"Often bought together" – connect what belongs together. Identify thematically matching products in the range and actively offer them. This creates needs and inspires visitors. For example, a Harry Potter fan doesn't just buy one book, but also a second one or a Blu-ray, etc. Based on visitor behavior and customer history, targeted recommendations can be displayed on the product detail page.
- Thematically related products are offered together
The AI engine of our personalization platform makes it possible
Before the recommendation engine is deployed, our AI engine first creates a comprehensive knowledge base for your online store. It uses your product catalog, your customers' click and purchase behavior, and our and your expert knowledge. This data is intelligently linked to calculate relevant product recommendations and deliver them in a targeted manner.
Since the recommendation engine is part of our modular personalization platform, you can use the knowledge base it generates for other touchpoints in your store, allowing you to comprehensively personalize the customer journey. This makes it possible to flexibly expand and continuously improve the recommendation system.
Want to learn more about the technology behind the AI engine? Then visit our E-commerce Technology page.
Discover our insights into the recommendation engine
01 / 5
Customization is a must-have for recommendation engines
Data Models
Depending on which phase of the customer journey a customer is in and which shop page they are on, different data models can be used. A hybrid model would be suitable on the product detail page, e.g., these pants are purchased with this T-shirt. Here, the click and purchase behavior of customers as well as human insights play a role.
Hypotheses
Hypotheses about customer purchasing behavior are formulated and tested using appropriate test procedures, such as A/B testing. The findings are incorporated into the recommendations.
Rules
With use case-oriented and highly customized rules, recommendations from the system are narrowed down according to the needs and industry knowledge of the shop operators.
This is how the integration of the recommendation engine works
Generation of a knowledge base
To generate the knowledge base, we need your product catalog in XML or CSV format and the click and purchase behavior of your shop customers. The latter is recorded via a tracking code in your online shop.
01
Setup in the test system
We discuss your visual ideas together and send you the div element as a placeholder for the front end of the relevant recommendations. You integrate this into your test system. Now the search algorithm is fine-tuned and the necessary configuration is carried out.
02
Activation in the live system
After setup in the test system and successful functional testing, the recommendation engine is activated in your live system and is available to your shop customers for displaying relevant recommendations. If necessary, we will perform fine-tuning during live operation.
03
Introduction to our backend
We will introduce you to our backend, the Control Desk, and train you on the various analysis and configuration options as well as feature usage. This will enable you to monitor and control the recommendation engine.
04
Display of relevant recommendations
Your shop customers receive relevant recommendations on various shop pages, which lead to impulse purchases. This provides inspiration during the purchase decision phase and increases the shopping cart value, which has a positive effect on your conversion rate and sales.
05
Monitoring and control of the recommendation engine
The epoq Control Desk provides you with extensive options for monitoring and controlling the recommendation engine.
Configuration
The "Configuration" tab shows you immediately how many recommendation widgets are integrated into your online shop and where they are placed. You can also see the various rules that are used as a basis. Clicking on the "Edit" button takes you to the Rule Editor, where you can create, edit, prioritize, or delete rules.
Key Figures
The "Key Figures" tab provides you with information about various KPIs for your recommendations, such as click performance, revenue rate, and conversion funnel.
Analysis
Using the "Analysis" tab, you can analyze your recommendation widgets with the help of key figures and display the most popular recommendations.
Features
The "Features" section offers you various options, such as Whitelisting, Blacklisting, Combi Creator, and Theme Worlds.
When is it worthwhile for you to use a recommendation engine?
Basic Functionality
You already run an online shop or similar and have all the basic features covered, and now you want to add incentives to buy on various shop pages.
Orders
You have more than 1,000 orders per month.
Product Range
You have more than 1,000 products in your range.
Want to inspire online shoppers to buy with relevant product recommendations?
Our Reco Engine fills shopping carts along the customer journey.
Popular blog articles about the recommendation engine
01 / 8