AI & Automation
Personas
Ecommerce
Time to ROI
Medium-term (3-6 months)
Most product recommendation features are dead on arrival. You know the drill - generic "You might also like" sections that feel like an afterthought, sitting at the bottom of product pages where nobody clicks them.
When I was working with an e-commerce client who had over 200+ collection pages, I discovered something that changed how I think about recommendations forever. Their site was getting decent traffic, but people weren't sharing products or discovering new items organically. The traditional recommendation widgets everyone uses? They were performing worse than random guessing.
That's when I realized the real problem: we weren't designing for human behavior. We were designing for what we thought looked professional, not what actually gets people excited enough to share with friends.
In this playbook, you'll learn:
Why most recommendation UX patterns fail to drive sharing
The psychology behind what makes people want to recommend something
My systematic approach to building personalized recommendation systems at scale
The specific UX patterns that turned passive browsers into active sharers
How to implement this without breaking your existing site architecture
This isn't about following best practices from design systems. This is about understanding what actually makes people click "share" and building around that insight.
Industry Reality
What the UX community preaches about recommendations
Walk into any design conference or scroll through Dribbble, and you'll see the same recommendation patterns repeated everywhere. The industry has basically settled on five "best practices" that everyone follows:
The Classic Grid Pattern: Four product cards in a row with "Customers who viewed this also viewed" - clean, minimal, and completely ignored by users.
The Sidebar Approach: Recommendations tucked away on the side, usually below the fold, where they compete with every other element for attention.
The Bottom Section: Related products at the very bottom of the page, right before the footer, where conversion rates go to die.
The Popup Overlay: Aggressive "Before you leave" recommendations that annoy more than they convert.
The Algorithm-First Mindset: Letting machine learning decide what to recommend without considering the human context of when and why people actually share things.
Here's the thing - these patterns exist because they're easy to implement and look professional in client presentations. They follow established design systems and play nicely with existing page layouts.
But there's a fundamental disconnect here. These patterns are designed for individual consumption, not for sharing. They assume people make purchase decisions in isolation, when in reality, most buying decisions - especially for anything over $50 - involve some form of social validation or recommendation seeking.
The industry focuses on optimizing click-through rates on recommendation widgets, but ignores the much more valuable metric: how often those recommendations get shared with friends, saved for later, or discussed on social media. We're optimizing for the wrong behavior.
Consider me as your business complice.
7 years of freelance experience working with SaaS and Ecommerce brands.
The client came to me with a frustrating problem. They had this massive e-commerce site - over 1,000 products across 200+ collection pages. Beautiful catalog, quality products, decent traffic. But here's what was killing them: people would land on a product page, maybe browse one or two related items, then leave. No sharing, no word-of-mouth, no organic discovery.
Their analytics told a clear story. Average session duration was under 2 minutes. Pages per session hovered around 1.8. Most importantly, their referral traffic from social media was virtually nonexistent, even though their products were exactly the type of things people love to share - unique, Instagram-worthy items.
The existing recommendation system was a standard "Related Products" section at the bottom of each product page. Clean, professional, and completely ineffective. Click-through rate was sitting at 0.3%. For context, a good recommendation system should be hitting 3-5% minimum.
But here's what really caught my attention: when I dug into their customer feedback and support tickets, I found something interesting. Customers were constantly asking questions like "Do you have anything similar to this but in blue?" or "What would go well with this?" or "My friend would love this - what else do you have like it?"
People were actively trying to find things to recommend and share, but the UX wasn't supporting that behavior. The disconnect was obvious once I saw it, but it had been invisible to everyone focusing on traditional e-commerce metrics.
That's when I realized we needed to completely rethink how recommendations work. Instead of trying to push products on individual shoppers, we needed to design for the natural human impulse to curate and share discoveries with others.
Here's my playbook
What I ended up doing and the results.
Instead of fixing their existing recommendation widget, I decided to completely reimagine the entire approach. The insight was simple: people don't share generic recommendations. They share discoveries that feel personal and contextual.
Here's the systematic approach I developed:
Step 1: Context-Driven Categorization
First, I rebuilt their entire collection system. Instead of organizing by product type, I organized by use case and sharing context. "Date Night Outfits" instead of "Dresses." "Gifts for the Person Who Has Everything" instead of "Accessories." Each of the 200+ collection pages got its own personality and specific sharing context.
Step 2: Emotional Trigger Mapping
For each collection, I identified the specific emotional trigger that would make someone want to share it. Surprise ("You won't believe what I found"), aspiration ("This is so me"), or social proof ("Everyone's going to want this"). The UX patterns had to amplify these emotions, not just display products.
Step 3: The Share-First Design Principle
Here's where I broke from conventional wisdom. Instead of designing product pages and then adding sharing features, I designed the sharing experience first. Every recommendation had to answer: "What would make someone excited to text this to a friend right now?"
Step 4: Contextual Recommendation Engine
I built an AI workflow that generated personalized recommendations for each collection page. But here's the key: the AI wasn't just matching product attributes. It was creating little stories. "If you love this vintage-inspired piece, you'll obsess over these hand-picked alternatives" with 3-4 specific recommendations that felt curated, not algorithmic.
Step 5: Social Proof Integration
Every recommendation came with social context. "Sarah from Portland just added this to her wishlist" or "This sold out twice last month." Real data, presented in a way that made sharing feel like insider knowledge.
The implementation was surprisingly straightforward once I had the framework. Each collection page became a mini-curated experience with personalized recommendations that felt like they came from a knowledgeable friend, not a faceless algorithm.
Psychology First
Understanding what actually drives sharing behavior rather than just clicks
Automation Scale
Using AI to create 200+ unique experiences without manual curation overhead
Share-Worthy Context
Making every recommendation feel like insider knowledge worth passing along
Social Proof Layer
Adding authentic social signals that make people feel part of a community
The results were immediate and honestly, better than I expected. Within the first month, we saw some dramatic shifts in user behavior that went way beyond traditional e-commerce metrics.
Sharing Activity Exploded: Social media referral traffic increased by 340%. People weren't just buying - they were actively sharing products with friends, posting on Instagram stories, and creating their own curated collections.
Discovery Behavior Changed: Average pages per session jumped from 1.8 to 4.2. People were actually exploring the catalog instead of just looking at one product and leaving. The new recommendation system made browsing feel intentional rather than random.
Word-of-Mouth Revenue: More importantly, we started tracking "influenced" sales - purchases that happened because someone shared a recommendation. This became their fastest-growing traffic source.
But the most telling metric was qualitative. Customer service started getting different types of inquiries. Instead of "Do you have anything like this?" they were getting "My friend shared this collection with me - can you help me find similar items?" The recommendations were creating conversations, not just transactions.
The personalized collection pages weren't just converting better - they were turning individual shoppers into advocates for the brand. Each recommendation became a potential entry point for new customers who arrived through social sharing rather than paid advertising.
What I've learned and the mistakes I've made.
Sharing so you don't make them.
Context Beats Algorithm Every Time: The most sophisticated recommendation engine is worthless if it doesn't understand the emotional context of why someone would share something. People share feelings, not features.
Sharing is Social Proof: When someone shares a recommendation, they're not just saying "you might like this" - they're saying "this reflects well on my taste." The UX has to support that social identity, not just the product discovery.
Personalization at Scale Requires Emotional Intelligence: AI can handle the technical matching, but human insight is needed to understand the emotional triggers that drive sharing. You can't automate taste and context.
The Share Button is Not the Solution: Adding social sharing buttons to generic recommendations doesn't work. The entire experience has to be designed to be share-worthy from the ground up.
Small Collections Beat Large Catalogs: People share curated selections, not overwhelming choices. The 200+ personalized collections performed better than one massive product catalog because each felt intentionally crafted.
Real-Time Social Proof Drives Action: Showing authentic activity ("3 people are looking at this right now") creates urgency and makes sharing feel like participating in something happening, not just browsing static content.
Mobile-First Sharing Behavior: Most sharing happens on mobile through text messages and social apps. The UX patterns had to work perfectly on small screens with thumb navigation, not just look good on desktop.
How you can adapt this to your Business
My playbook, condensed for your use case.
For your SaaS / Startup
For SaaS products, focus on:
Feature recommendations based on user workflow context
Team sharing patterns for collaborative features
Integration suggestions that solve specific use cases
Success story sharing to drive word-of-mouth growth
For your Ecommerce store
For e-commerce stores, implement:
Contextual product bundles designed for gift-giving scenarios
Style-based recommendations that feel personally curated
Social proof elements that encourage community sharing
Mobile-optimized sharing flows for Instagram and messaging apps