Meteor Bikes - Conversion Boost
OVERVIEW
Meteor Bikes is a hypothetical eCommerce project I did through my Springboard training. Analytics revealed 50% of users open on average seven item pages and then abandon our site without moving items into the cart. Additionally, 70% of users who place items in the cart abandon their cart at the registration page. Currently users must have an account to purchase.
PROBLEM STATEMENT
Users need a more engaging way to search for bikes online. The site needs to allow users to checkout as guests to increase conversion. I aim to find a solution and build a workable prototype.
USERS AND AUDIENCE
This site is "geared" for high income earning bike enthusiasts who research bikes/bike products thoroughly. They expect to spend a lot of money on their investment.
RESPONSIBILITIES
As the lead on this project my responsibilities included competitive analysis, expert interviews, low and high fidelity prototyping, and usability testing.
RESEARCH
My first step was to conduct a competitive analysis to see what others in the field are doing on their sites while making note of my likes, dislikes, as well any items they left out.
The first sites I hit were Amazon, Target, and Trek. Target and Amazon had essentially identical flows for a bike search. They have an overwhelming amount of options, but with very clear ratings and reviews. However, a user may not purchase from their sites without a member login.
Trek had a unique user experience of the three. They had action-filled, engaging product photos which highlighted the quality of their products. They had a comparison feature to allow users an easier selection process. Most intriguing to me however, they had a Bike Finder which helps users find their perfect bike based on the type of riding they prefer.
METHODOLOGY
Since I needed to build wireframes before I could conduct user tests, I utilized a Lean UX approach and started with assumptions. With bike enthusiasts as the audience, my assumption is that users would have a good idea of the types of products they needed; the Bike Finder feature could streamline the process of matching users to their desired products.
A user flow enabled me to visualize how this would play out on the mobile web version.
LOW FIDELITY PROTOTYPE
The low fidelity prototype allowed me to test my assumptions on actual users without dedicating too much time on an untested design.
This prototype included a comparison feature, assuming that would aid the user in making a selection between suggested products.
It also included and a guest checkout option assuming that would boost conversion.
I conducted five remote usability tests with bike enthusiasts. I gained some insights into the mind of the enthusiast that would be crucial for designing a user experience catered to that specific audience.
“A cool photo would almost interest me more than all the other data.”
Users want to imagine themselves out riding, having their own adventures. Aesthetically vibrant imagery is crucial for engaging users to stay on your site.
"The Bike Finder thing makes me think this site is for beginners."
Bike enthusiasts have a great deal of knowledge about what type of riding they do, the products required for that type of riding, etc. The Bike Finder, although useful for novice riders (such as myself), is not something real enthusiasts would have a need for.
This sparked an idea for the next iteration to include a Bike Builder feature, allowing users to select components to make a whole bike to their specific needs. I also removed the comparison feature, feeling it would not be necessary since the user would be building their ideal bike from scratch.
HIGH FIDELITY PROTOTYPE
I added some engaging imagery and reformatted the flow for this iteration.
I added some engaging imagery and reformatted the flow for this iteration.
I included a segment that shows the bike as it is being built, with the assumption being users would be excited about seeing their almost finished product as they select the final components.
After building the High Fidelity prototype I conducted another round of usability tests in which I was able to test and interview five professional/semi-professional cyclists.
MORE INSIGHTS
In the second round of usability tests, the pros offered some crucial insights about the Bike Builder feature.
"You would want an option for buying a complete groupset because even if you know your stuff it's hard to know what is compatible with what."
Parts like the rear cassette, chain ring, front and rear derailleurs, full brakes setups, chain, and cranks should be sold in groupsets to eliminate the risk of incompatibility. This would also streamline the funnel and likely lead to higher conversion rates.
"Star ratings always seem like bullsh*t to me. I usually check blog posts, Reddit threads, customer reviews before I buy anything."
Users have come to distrust the star rating system; it's too easy to flood products with bogus five star ratings. The products should have links to reviews/blogs/subreddits etc. to gain that trust true enthusiasts require to make a purchase.
NEXT ITERATION
I made some adjustments to the prototype based on this user feedback, replacing the star rating with links to reviews, and making groupsets for the components.
TAKEAWAYS
My competitive analysis led me to initial assumptions that proved to be a good starting point for a workable prototype. Usability tests were crucial in attaining familiarity with the target audience's POV so that we may keep them engaged and gain their trust which is ultimately necessary to attain conversion. Also, the addition of a guest checkout option brings inclusivity to the site and will increase conversion.
NEXT STEPS
The iterative process is never truly done. With changing trends and tastes there is almost always something to improve upon for user centered design. One of my initial assumptions for Meteor Bikes was to include a comparison feature to allow an easier selection process for the user. When the Bike Builder replaced the Bike Finder I had assumed this would no longer be necessary. Testing showed the comparison feature would be helpful in the Bike Builder feature to help compare the individual components and groupsets the users are selecting to build their bike, so the next iteration of this prototype would be built to include that feature.