The selling power of social proof

My role: UX Lead // Project timeline: 4 months // Company: Ocado Technology (e-commerce platform)

Building the trust that turns browsing into buying.

Social proof is a powerful behavioural principle where people rely on the actions and opinions of others to guide their decisions, particularly in uncertain situations.

In grocery shopping, this often influences decisions like: choosing a wine for dinner, trying new ingredient or selecting unfamiliar brands.

When it comes to food shopping, around 70% of products in baskets are previously bought items. While this makes basket building efficient, it significantly limits new product discovery.

Problem to solve

In online grocery shopping, around 70% of products in baskets consist of previously bought items. Users tend to stick to familiar products, which limits new product discovery. This is a challenge for both Ocado and its supermaket partners, who are eager to showcase new products to a wider audience.

Outcome

The introduction of social proof features helped surface new products in moments where customers were making decisions. By highlighting what other shoppers were buying, rating highly, or searching for, we were able to increase visibility and trust around unfamiliar products without disrupting the existing shopping flow.

Despite the e-commerce platform already offering some social proof features, new product discovery and purchase remained low.

What the data told us?

The rate of new items added to baskets declines over time and eventually becomes stagnant.

Ocado’s platform is highly optimised for efficient basket building, making it very easy for customers to find and add previously purchased products. While this improves convenience, it also reinforces habitual purchasing behaviour and limits product discovery.

Several existing features contribute to this pattern: Repeat Last Order, Favourites pages and Recurring Orders

These capabilities help users quickly rebuild their usual shop, but they also reduce opportunities for customers to explore and try new products.

Define opportunities

The PM and I consolidated product data, past research, and behavioural insights to frame the opportunity.

We then organised a kickoff workshop with a multidisciplinary team including Product, Design, Data Science, and Engineering.

Research & best practices

I analysed:

past customer research
ecommerce best practices
social proof patterns across retail platform

Ideation workshop

Once we collected insights and use cases from user interviews, I organised a workshop session with the team, the multidisciplinary team involved the Product Manager, Product Designer, Data science and Developers. During the workshop, we aligned our understanding of the project goal, discussed social proof mechanisms I collected in a benchmarking presentation and engaged in ideation exercises to generate multiple solutions. We then discussed the feasibility of these solutions and the scope of the project.

What was implemented?

Reviews summarised by AI

Introduced concise AI-generated summaries that synthesize key insights from large volumes of customer reviews. Instead of requiring shoppers to read dozens of individual comments, the system performs sentiment analysis to highlight the most common positive and negative themes about a product.

This works as a powerful form of social proof because it surfaces the collective opinion of other shoppers in a clear and digestible format.

These helped customers quickly evaluate new products products with greater confidence.

Dynamic product badges

Dynamic badges are used to highlight items based on social attributes that enhance product discovery and trust to try new items.

They draw attention to special attributes, making it easy for users to scan the product grid and select an item.

This works as a powerful form of social proof because it surfaces the collective opinion of other shoppers in a clear and digestible format. These helped customers quickly evaluate new products products with greater confidence.

Attributes for dynamic badges defined with the Data Science team

💰Bestseller: based on sales volume and velocity per category. We are limited ‘bestseller’ to the top 3 products with highest sales velocity per category  

📈 Trending: applied to products with the highest conversion rate from trending search terms.
Note: trending search terms are defined as search terms with the largest increase in search volume within a 7-day window comparison.

⭐ Top rated: applied to products with a high customer rating (e.g: min 4.5+ stars) and a substantial quantity of reviews (e.g: a min of 100 reviews)

🔥 Selling fast: applied to products with the highest uplift in sales within a 24 hour period.

🆕 New: Products new to the catalogue that receive the badge automatically – after 6 weeks the badge is automatically removed

Testing

Multiple design variations were tested using Optimizely across search and product listing pages.

Key metrics included:

add-to-basket conversion
new product adoption
click-through rates on discovery surfaces

Results and impact

Following rollout:

+12% increase in new product additions to baskets
+9% higher click-through on product cards featuring social proof badges
+2.8% uplift in add-to-basket for products with AI review summaries
Trending searches drove a 11% increase in product discovery interactions

The capabilities were released as a scalable platform feature, enabling retailers to control which social proof mechanisms appear across their stores through feature flags.