Product Recommendation Engine Development for E-Commerce
Kandy scales e-commerce brands, and it wanted its partner stores to recommend products based on real shopper behavior instead of static catalog rules. Through custom software development, Acquaint Softtech built a machine learning recommendation engine with behavioral data pipelines and real-time recommendation APIs. Once deployed across several stores, it brought clear improvements in customer engagement, average order value, and repeat purchases.
5+
E-COMMERCE BRANDS BUILT3+
STOREFRONT TOUCHPOINTS PERSONALIZED5+
STRUCTURED DELIVERY PHASES
CLIENT STORY
Kandy For Scale is a boutique, full-stack growth partner for DTC e-commerce brands looking to scale to eight and nine figures. Its team handles paid ads, creatives, email marketing, landing pages, and strategy under one roof. Kandy has also launched more than five e-commerce brands of its own, which it uses to test ideas in real markets.
Kandy wanted to give its partner brands smarter personalization. Rather than showing products through static catalog rules, it needed a system that learns from browsing and purchase behavior. Kandy partnered with Acquaint Softtech for custom software development: a recommendation engine and behavioral analytics backend that brands could use without building separate analytics infrastructure.
Country
Lithuania
Industry
E-Commerce Growth & MarketinginTech
Team Size
6-10 Professionals
Hours Invested
1,720 Hours
Technology
Shopify, Python
Project Background
Kandy’s partner brands generate browsing, product view, and purchase data. Acquaint Softtech was brought in to turn that behavior into personalized product suggestions that appear at the right moment in the shopping journey.
Behavior-Based Recommendations: A smart recommendation system that personalizes product suggestions from user behavior rather than static catalog rules.
Scalable Behavioral Analytics: A backend platform that analyzes browsing patterns, product views, and purchase activity across partner brands.
Personalized Shopping Journeys: Helping brands present relevant products at the right moment during each browsing session.
Their engineers focused heavily on performance and scalability, ensuring that recommendations appeared instantly.
Research & Ideation
The project moved through clear phases: discovery, system design, development, testing, and rollout. Acquaint Softtech kept regular contact with Kandy’s product and technical teams, and built a backend designed to support machine learning models that analyze user behavior and product interactions.
Behavior Over Static Rules
Designing models around product categories, browsing sequences, previous purchases, and engagement signals.
Built for Every Catalog
Making the system flexible enough for each store to implement recommendations based on its own catalog and customers.
Controlled Staged Rollout
Delivering in stages so features could be tested in a controlled environment before reaching additional stores.
Creative Design
Design here meant putting the right product in front of the shopper at the right moment. Recommendations appear on store pages, product detail pages, and inside checkout flows, and they load instantly without slowing the page. For Kandy’s teams, the aim was practical tools e-commerce marketers can use without deep technical expertise, supported by dashboards that show how recommendations perform.
Project Challenges
Limits of Static Catalog Rules
Product suggestions based on fixed catalog rules could not respond to how individual shoppers actually browse and buy. Kandy needed personalization driven by behavior.
Behavioral Data Across Many Stores
Browsing activity, product views, and purchase history from multiple partner brand stores had to be collected and organized before any meaningful insight was possible.
Speed on the Storefront
Recommendations had to appear instantly. Any delay in page load time would hurt the shopping experience the system was meant to improve.
Different Stores, Different Catalogs
Each partner brand has its own catalog structure and customer behavior. One rigid setup would not fit every store.
Our Solutions
Acquaint Softtech designed and built a custom recommendation engine that processes e-commerce behavioral data and generates intelligent product suggestions in real time.
Custom Recommendation Backend
A custom backend that powers machine learning models capable of analyzing patterns in user behavior and product interactions.
Behavioral Data Processing Pipeline
A pipeline that collects and organizes browsing activity, product views, and purchase history from partner brand stores, turning real interactions into usable insight.
Machine Learning Recommendation Engine
An engine that evaluates product similarity and customer interest patterns across categories, browsing sequences, previous purchases, and engagement signals.
Dynamic Recommendation APIs
APIs that display personalized product suggestions within store pages, checkout flows, and product detail pages in real time.
Recommendation Monitoring Dashboards
Dashboards that let Kandy’s teams evaluate how recommendations perform across stores.
Custom Functionalities
Beyond the core engine, Acquaint Softtech built features that make personalization practical for every partner brand.
Product Similarity Scoring
Scoring models that measure how closely products relate, so shoppers see relevant and complementary items.
Real-Time Storefront Recommendations
Personalized suggestions delivered instantly on product pages, in carts, and during checkout without slowing page loads.
Returning Customer Personalization
Recommendations for returning customers that align with their previous buying patterns.
Store-Level Flexibility
Each store can implement recommendations differently, depending on its catalog structure and customer behavior.
On Going Support
Throughout the engagement, Acquaint Softtech stayed closely involved as the recommendation system moved from controlled testing to live stores.
Recommendation Logic Adjustments
When Kandy requested changes to recommendation logic, the team responded and refined the models.
API Performance Improvements
Requests to improve API performance were handled promptly, keeping recommendations fast on the storefront.
Store-by-Store Expansion
Staged deliveries let features be proven in a controlled environment before rolling out to additional stores.
Project Outcome
Once the recommendation system was deployed across several stores, Kandy saw clear improvements in customer engagement. Shoppers interacted more often with suggested products, especially on product pages and in their carts, and discovered items they might not have found by browsing alone.
Average order value also improved, as shoppers explored complementary products before completing a purchase. Repeat purchases grew more personal too, with returning customers receiving suggestions based on their past buying patterns. Kandy’s partner brands can now offer personalization without building separate analytics infrastructure.
India (Head Office)
203/204, Shapath-II, Near Silver Leaf Hotel, Opp. Rajpath Club, SG Highway, Ahmedabad-380054, Gujarat
USA
7838 Camino Cielo St, Highland, CA 92346
UK
The Powerhouse, 21 Woodthorpe Road, Ashford, England, TW15 2RP
New Zealand
42 Exler Place, Avondale, Auckland 0600, New Zealand
Canada
141 Skyview Bay NE , Calgary, Alberta, T3N 2K6
Your Project. Our Expertise. Let’s Connect.
Get in touch with our team to discuss your goals and start your journey with vetted developers in 48 hours.