Discogs is an online database and marketplace for music collectors. It started as a database for electronic music and today it facilitates the sales of physical music. Vinyl records make up 75% of those sales. This case study represents a conceputual product design.
80 Hours
Adobe XD
Illustrator
Photoshop
Miro
Sole Product Designer and Researcher
Identify the problems with the exisiting platform
Clean up the UI within the framework of the current platform
Implement a feature to allow users to save preferred filters
Clarify wording around wantlist items within the seller's inventory
Build, test, and iterate interactive prototypes.
Inquiries into to the Discogs community revealed two common pain points:
Searching the archives of the Discog message board helped confirm this and suggested other problems.
A usability test was conducted with six participants, a blend of casual Discogs users and others who were not familiar with the specific platform, but had other ecommerce experience.
Users were asked to log in to an existing account for the purposes of testing the “Items I Want” page in the Discogs Marketplace. The results confirmed the need to set and save filters, especially the ability to set some default settings (eg. Show only items shipping from the United States).
"I don't want it to be some redundant listings."
-Joe on original "More Items I Want" feature
"My first thought is I kind of hate it. Everything is so compact."
"It reminds me of websites from 2005. That's the vibe I get from it."
-Laurel on look and feel
"A lot of information. A lot of small information."
-Jared, on look and feel
"There is so much information. I want to, like, get rid of some stuff."
-Elise, on look and feel
"It would be more helpful than [that number], because [that number] is a little bit...not correct"
-Emily on original "More Items I Want" feature
"I don't love the blue hyperlinks."
-Stephanie on look and feel
"It's always going to be VG+. It's always going to be USD."
-Jared on "Saved Filters" preferences
Results confirmed the initial hypothesis and brought other issues into the the light.
• The UI was outdated and cluttered
• The existing filters were not effective
• No ability to save filters
• The “Has Items I Want” feature was confusing in the wording and execution
This was found to be the most confusing and inconsistent. If it reads "Has 10 More Items I Want" it might mean that the seller has ten unique items on the user's wantlist.
However, if the user has The Beatles' album Abbey Road on their want list, it might mean that the seller has 10 copies of Abbey Road for sale.
Since the user likely only wants a single copy of the item on their wantlist, this use can be very misleading.
The feature is implemented this way because each listing of the same release could be unique. There may be copies in various conditions (eg. Mint, Near Mint, Very Good) or defining attributes, such as being signed by the artist.
Filter styles from three other eCommerce styles were compared: Amazon, CycleTrader, and eBay
All made use of a fixed column filter model, with the last two also having a C2C business model like Discogs.
Both CycleTrader and eBay had the most desirable traits with eBay's ability to save a search being somewhat limited, and CycleTrader's filter layout being too large.
Discogs feature set for filters did not match any of the competing designs that were evaluated.
Due to constraints of the project, the scope was narrowed to the following features:
Allow a user to establish default settings for subsequent visits. The current default is to show most recently listed items on the user’s wantlist from around the world.
The existing UI was cluttered with information that is pertinent only to the most die-hard collectors (eg. catalog number of a release and record label), which is already available for inspection within their wantlist.
Users across the board were overwhelmed with the amount of information shown.
There was a lot of confusion over what the numbers indicated.
Having established that the original structure would remain intact, the primary element tested in mid-level wireframes was the individual listing.
In each of the examples below, the album size is larger than the original version. Increasing the size of the album image lends to more space vertically, but less space horizontally. Different versions were shared with some peers and the third version became what was used for the final layout before moving to the prototype.
I opted to remove the record label, catalog numbers, and seller's description.
This lead to a much cleaner and easier to read listing.
Six participants were recruited for testing the prototype and carrying out the tasks above. Two of those participants were from the first round of testing the existing product as well as casual users of the site.
Testing had a 100% success rate. All testers made it through in short order and zero red flags.
Testing results were color coded, with successes in green and items to consider for iterations in yellow.
Those Under Consideration items were plotted in a quick Priority Matrix measuring the user benefit against the perceived difficulty of implementing those changes either in the prototype or down the road for development.
© 2023 John Everett Morton. Built with Semplice and Wordpress.