01 · Homepage
Recommendations on entry
The algorithm personalizes products on the homepage in real time based on preferences - the customer immediately sees a relevant assortment.
Case study · e-commerce / home furnishings
Courage in innovation - transforming the e-commerce marketing of homla.com.pl
For Homla (N2H) - a home furnishings brand with 62 stores and an online shop - Persooa implemented personalized real-time recommendations on the Synerise platform to increase revenue from existing traffic. This was confirmed by a 7-day A/B test.
+14%revenue-per-user growth (vs. comparison group)
+8%increase in the number of transactions
~10%improvement in average order value
62stores in Poland + online sales
Client
Homla (N2H)
Area
Personalization, real-time AI recommendations
Segment
B2C · e-commerce (home & decor / home furnishings)
Technology
Synerise + Persooa
Starting point
A rich assortment and high traffic, not fully leveraged
The brand has strong recognition and thousands of visits a day. The challenge was to increase revenue from existing traffic, not to acquire new users - within a range so broad that it's easy to get lost in it.
Thousands of products - hard to find what the customer is looking for
Goal: more revenue from existing traffic, not new users
Manual recommendation strategies consumed time and resources
Suboptimal purchase path within a broad assortment
Low number of products and basket value
No real-time personalization
Persooa's solution
Real-time AI recommendations across the entire journey
We launched a set of personalized, real-time recommendation scenarios - from the homepage, through the product page, to the cart - taking manual work off the team's plate.
02 · Product page
Similar products
The module shows items with similar features to those previously bought or viewed - making it easier to discover the range.
03 · Returning visitors
Recently viewed & cart widget
For both recognized and anonymous users: recently viewed products or a reminder of items left in the cart, throughout the session.
04 · Automation
Algorithms instead of manual work
Recommendations adapt automatically in real time - no manual strategies, with resource savings.
Phased implementation
From implementation plan to verified results
1
Plan and platform
Phase I
- Implementation of the Synerise platform
- Delivery plan for the objectives with timelines
- Priority: personalization of recommendations
- Map of scenarios across the purchase journey
2
Recommendation scenarios
Phase II
- Recommendation module on the homepage
- Similar products on the product page
- Recently viewed / abandoned cart widget
- Unique, personalized module in the cart
3
Testing and optimization
Phase III
- 7-day A/B test (variant with personalization vs. without)
- Measurement of revenue per user, transactions and AOV
- Insights into the role of recommendations in site performance
- Foundation for further campaigns and scaling
Tactics implemented
Recommendation scenarios across the entire purchase journey
A set of dynamic recommendation modules whose effectiveness was confirmed by a 7-day A/B test on existing traffic.
Recommendations · home
Homepage module
Products personalized in real time right after landing on the site - matched to user preferences.
real-time from the first view
Recommendations · PDP
Similar products
On the product page, a module with items similar to those previously bought or viewed.
similar easier discovery
Return · widget
Recently viewed
For users with no products in the cart, a widget shows recently viewed items - even for anonymous users.
anon+known quick return
Cart · reco
Items left in the cart
When there are products in the cart, a widget reminds the user of them and adds a unique, tailored offer.
1:1 unique offer
Verification · A/B
7-day A/B test
The variant with three recommendation modules and the returning-visitor widget was compared against a variant without personalization.
50/50 A vs B, 7 days
Efficiency · resources
Automation of recommendations
Synerise's algorithms replaced manual strategies - efficiently and in real time, with resource savings.
auto less manual work
Results
More revenue from the same traffic
+14%
increase in average revenue per existing user compared to the comparison group - alongside +8% in the number of transactions and a nearly 10% higher average order value, in a 7-day A/B test.
Results of the 7-day A/B test
Revenue per user+14%
Average order value~10%
Number of transactions+8%
Real-time recommendations
Real-time product personalization based on preferences.
Similar products
On the product page, based on prior purchases and browsing.
Returning-visitor widget
Recently viewed items and abandoned cart (recognized and anonymous users).
Cart personalization
A unique, tailored offer in every customer's cart.
A/B test
A 7-day measurement of the real impact of personalization on results.
Automation
Algorithms instead of manual strategies - resource savings.
We squeeze more revenue out of existing traffic - turning AI recommendations into a bigger basket, more transactions and a better experience, in a Human + AI model.
- Persooa · Revenue Growth Agent
Frequently asked questions
FAQ - recommendation personalization at Homla
What results did recommendation personalization bring at Homla?
In a 7-day A/B test, the personalized variant increased average revenue per user by +14% compared to the comparison group, the number of transactions by +8%, and average order value by nearly 10% - all on existing traffic, without acquiring new users.
What recommendation scenarios were implemented?
A real-time personalized recommendation module on the homepage, a similar-products module on the product page, a returning-visitor widget (recently viewed items or a reminder of an abandoned cart - for recognized and anonymous users), and a unique, personalized module in the cart.
How was the effectiveness of the actions measured?
We ran a 7-day A/B test: variant A, with three recommendation modules and the returning-visitor widget, was compared against variant B without those elements. This allowed us to clearly assess the impact of personalization on engagement and sales.
Why did the focus stay on existing traffic rather than new customers?
The brand has strong recognition and thousands of visits a day, so the greatest potential lay in making better use of existing traffic. Personalizing recommendations shortened the purchase path within the rich assortment and increased the purchase value of existing customers.
What technology was used?
The Synerise platform combined with Persooa's framework and components - with recommendation algorithms running in real time, which replaced manual strategies and brought resource savings for the Homla team.
What does Persooa do?
Persooa implements and activates AI in e-commerce - mainly on the Synerise platform - with an emphasis on a fast start and the first measurable results within 90 days. We combine strategy, data and execution in a Human + AI model: personalization, marketing automation, recommendations and AI agents that translate into revenue growth.
Get started
Want a similar result?
Book a free call - we will analyze your situation and show where we can deliver measurable growth fastest.