I designed a product recommendation tool that helps users find the right console, games, and accessories based on their personal preferences and behavior.
What is it?
The Help Me Choose tool is a web experience that creates a personalized recommendation for consoles, games, and accessories, for each user based on their response to a series of questions.
The challenge
As the Xbox One platform expanded beyond a single console, our client’s research began to indicate that users were experiencing product confusion; the need for a guided purchase experience became apparent. Our clients asked my team to explore options to help users make the best purchase decision.
Audiences
Our audiences were primarily those new to console, and guardians looking to buy their child the right console, but we wanted to make sure that people familiar with the brand also got value out of the experience. We had to speak to the full range of potential visitors.
The process
Exploring guided decision making
I started by exploring the landscape of guided decision making on the web. I audited experiences like chat bots, on demand customer service reps, product selection wizards, and curated collections. I found that chat bots and on demand customer service representatives required the user to avtively reach out and that users would be required to ask the right questions to get the best answers. It proved too high a barrier to quick information. Curated collections was a good alternative but wasn’t personalized to individual users, just broad user groups. We wanted to recommend the best product for each user and the clients had a desire to explore personalization across the marketing site, so a guided product selection experience was chosen.
Understanding competitors
I audited competitive product selection marketing experiences and defined a list of best practices and applicable experience patterns. I explored storytelling, design patterns, feature-sets, results pages, and opportunities for CRM extensions.
Defining the algorithm
I began to explore options for laying out a product selection experience. I first looked at presenting the products and communicating the difference between them. But we found this to be too technical for new users. Looking to lower the barrier to usage, I explored asking for user preferences and behaviors emulating the experience you’d go through in store or with a friend who knew about the product. Quick and easy questions, that would effectively let us know what a user’s needs were.
I explored different options for algorithms like a branching experience that refined from a number of products to a single product recommendation. The branching experience would require dozens of nested questions and it simply was too linear and required too intense decision making for new users. We wanted users to move freely back and foward through the process and to be able to skip a question if they didn’t feel confident. The branching experience required confident decision making which wasn’t right for our audiences. I chose the weighted choice model for the algorithm that allowed us to ask a question, accept a range of responses, and assign a weight to each response leaning a user towards the best console for their needs.
Creating the questions
Based on our product offerings and our user research I simplified the objective into 3 questions:
- Which console is right for me?
- Which games are right for me?
- Which accessories and services are right for me?
I established the key differentiators between products and began generating questions that would get directly to the user needs. I came up with 9 behavior and preference questions that would get users to the best recommendation of products for them. Each question had a range of responses, and each response was assigned a weighted value that would drive them to one set of products or another.
To minimize barriers, I designed our algorithm so that each question was potentially skippable. Key responses like “I want to play games in 4K” had a high value assigned so that only certain products like the Xbox One X console (which is the only console that plays games in 4K) would be recommended.
Wireframing
Once the algorithm and questions were defined, I wireframed the experience. Each screen asked a single question with a range of responses. The user moves forward by making their selection and clicking NEXT.
I worked with the visual designers as they created mood boards, defined the visual styling, and created the illustrations. Our emotional direction leaned towards fun, friendly, and conversational and I think our visual designers did a great job communicating that.
The results page highlighted the console right for the user. Below the console, an algorithmically generated paragraph described why the console was right for each user based on their responses. Below that, console bundle options were shown based on the results. A list of recommended games based on age appropriateness and genre preference was recommended. Accessories and services based on behavior and preferences were recommended.
Usability testing
With our wireframes complete, I performed remote unmoderated usability tests using an inVision prototype through our internal testing service. With the test complete, we made minor adjustments to the questions and our algorithm, and began handoff to developers.
Handoff
During the handoff to developers, I created functional spec documentation plus additional accessibility specs documentation. I also performed QA on early versions of the tool.
Post launch
After launch, our marketing sciences team generated a performance analysis report which revealed extremely high completion rates (>75%). To increase performance, we made minor copy edits at our points of greatest bottlenecking. Overall, the tool accomplished our goals, it made it easier to find the correct set of products and helped our clients reach out to users new to the ecosystem. It still lives on xbox.com today as a key component in the discovery process.
It still lives on xbox.com today. See it live