Creating opportunities for Sysco customers to discover more products.
Recommendations in Sysco Shop
Feature improvement for recommending personalized products to Sysco customers.
Company | Sysco LABS
Time frame | 3 months, FY 2021
Platform | Responsive Web, Mobile App
The “did you forget?” recommendation in the checkout experience was only one recommendation occurrence in the product initially. It served as the launchpad to add more recommendation types in Sysco Shop.
Our technical constraints included having a limited recommendation algorithm. The “did you forget?” recommendation type used the customer’s order history and frequency to suggest these products.
We decided to build an algorithm that would populate relevant products specific to the recommendation type.
The challenge
How can we help Sysco customers find more products
to provide a more personalized shopping experience?
I was the lead product designer throughout the project cycle from defining initial requirements and collect user research to delivering concept ideations, user flows, wireframes, prototypes, and visual designs.
I collaborated within the product team to create a robust product recommendation engine through various touch-points in Sysco Shop.
We measured engagement with these recommendations, add to cart conversions, and defined a roadmap for feature improvements including other recommendation types.
The team & my role
In collaboration with a UX researcher, we sent a survey to 100 customers to determine what sort of recommendations they would want to see at various points in the shopping experience.
Our Questions:
How customers did customers get product recommendations?
Where do customers expect to see recommendations?
Where do our customers want to see product recommendations were they would be more helpful rather than annoying and intrusive?
The research
By mapping out a high-level overview of what our customers were doing in the product helped to identify opportunities to add recommendations and what types of recommendations would be helpful for that specific page.
Customer flow
I began my design research by evaluating e-Commerce best practices and analyzing other competitors to determine the most optimal design strategy for our product.
I will create a space in my design project to collect screenshots, links, and design inspiration to use as a reference checkpoint.
I start my projects by checking out how the leading B2C e-Commerce companies solve a similar problem.
Competitive analysis
From the findings from the survey and reviewing other e-Commerce patterns, it was a clear decision to create a scrollable carousel with multiple products on product detail pages, in cart, and on the homepage.
To create a personalized experience, the products in these carousels needed to be tailored to the customer’s relevant tasks throughout their journey.
Synthesis
From my analysis, I began brainstorming with the product manager and UX research for different types of recommendations and where they might be seen in the Shop experience.
For example, “recently viewed” recommendations could be helpful to returning customers when they arrive to the homepage. Others types of recommendations include “frequently purchased next” on the product detail page.
Research outcomes
I analyzed the current structure of the “did you forget” recommendation and identified opportunities for design alignments with our design language system. I gathered usage and engagement data from the current experience.
I collected many successful and unsuccessful examples of this experience from other apps to perform a competitive analysis. I was primarily interested in the interaction design for web and mobile.
Design Iterations
Final Design
Product recommendations generated $91 million in sales and $71 million in revenue for Sysco in fiscal year 2021.
Currently, Sysco is testing +10 recommendation types across all United States regions and Sysco Canada.
Results
The launch of this feature proved to be the right step in creating a personalized customer experience in Sysco Shop. The robust design of these recommendations allowed for further recommendation types to be added and expand our A/B testing capabilities.