Client
Craigslist LTD
Role
Lead Product Designer
Direct Reports
1
Total redesign of $5b+ ecommerce platform
No, this isn't a fake portfolio piece - I was actually contracted by Craigslist to create a total redesign for this site.
No, this isn't a fake portfolio piece - I was actually contracted by Craigslist to create a total redesign for this site.
No, this isn't a fake portfolio piece - I was actually contracted by Craigslist to create a total redesign for this site.
▷ Overview
Massive $5bi global ecommerce and classified site
Revenue dropping at Craigslist due to usability and aesthetic issues
Contracted as the first designer to work on Craigslist in its 25-year history
▷ Overview
Massive $5bi global ecommerce and classified site
Revenue dropping at Craigslist due to usability and aesthetic issues
Contracted as the first designer to work on Craigslist in its 25-year history
▷ My responsibilities
100% of UX and UI
Brainstormed new features with founder and prototyped them
Responsible for validation and iteration with users
▷ My responsibilities
100% of UX and UI
Brainstormed new features with founder and prototyped them
Responsible for validation and iteration with users
▷ Team
Me, Head of Engineering, Head of Partnerships, CEO
Worked with other developers and tested with users
▷ Team
Me, Head of Engineering, Head of Partnerships, CEO
Worked with other developers and tested with users
▷ Outcome
Simplified hundreds of flows in over 100 categories
Currently rolled out in Hawaii for beta users with:
35% less usability issues , 15% less time on task , 15% increase in post creation, 18% increase in scroll depth
Estimated $50 million YoY savings
▷ Outcome
Simplified hundreds of flows in over 100 categories
Currently rolled out in Hawaii for beta users with:
35% less usability issues , 15% less time on task , 15% increase in post creation, 18% increase in scroll depth
Estimated $50 million YoY savings
▷ The Challenges
Craigslist is a site like no other. The design is immediately recognisable. No fluff, no upselling, just to the point bare-bones styling. This gives it a strong brand identity, but it also means that it is not very attractive nor very easy to use. However, over 25+ years, millions of users have become accustomed to its quirks.
But craigslist was hemorrhaging users to Facebook marketplace and other classified sites, with revenue dropping by 10% year on year.
Any redesign had to:
incorporate for the 'feral' quality of the site - it has a charm that couldn't be lost in any redesign.
integrate and create seamless workflows between highly disparate categories.
correct hundreds of usability issues while ensuring user's muscle memory was not affected
How could I introduce modern UX standards to a massive, multi-category ecosystem without breaking the brutal simplicity that defines the Craigslist brand?
▷ The Challenges
Craigslist is a site like no other. The design is immediately recognisable. No fluff, no upselling, just to the point bare-bones styling. This gives it a strong brand identity, but it also means that it is not very attractive nor very easy to use. However, over 25+ years, millions of users have become accustomed to its quirks.
But craigslist was hemorrhaging users to Facebook marketplace and other classified sites, with revenue dropping by 10% year on year.
Any redesign had to:
incorporate for the 'feral' quality of the site - it has a charm that couldn't be lost in any redesign.
integrate and create seamless workflows between highly disparate categories.
correct hundreds of usability issues while ensuring user's muscle memory was not affected
How could I introduce modern UX standards to a massive, multi-category ecosystem without breaking the brutal simplicity that defines the Craigslist brand?
▷ Defining success
To help clarify our goals, I established proto-personas that reflected our goals of new vs existing users: (anti-persona) and our "Sacred Cows” (things we couldn't change). I focussed our markers of success for the business in terms of revenue: retention (D7, D30), CTR of categories, scroll depth and post creation. Then, the criteria of these personas were used in a UX review I conducted to establish key pain points to address. Speed of change was vital: I developed a process to tackle high priority issues each week, address them with new designs, and get quick feedback from users each week.
▷ Defining success
To help clarify our goals, I established proto-personas that reflected our goals of new vs existing users: (anti-persona) and our "Sacred Cows” (things we couldn't change). I focussed our markers of success for the business in terms of revenue: retention (D7, D30), CTR of categories, scroll depth and post creation. Then, the criteria of these personas were used in a UX review I conducted to establish key pain points to address. Speed of change was vital: I developed a process to tackle high priority issues each week, address them with new designs, and get quick feedback from users each week.
▷ Getting the user to point of purchase
Filtering, scanning posts and parsing the UI was key to getting users to find items to purchase. On the app, this was hugely important (roughly half of users used it) but the experience was ill-designed, with a bloated UI, hidden filters, un-scannable posts and an inability to select multiple categories. I completely redesigned the UI to help get users to easily filter their searches and immediately recognise what they want in post listings.
▷ Getting the user to point of purchase
Filtering, scanning posts and parsing the UI was key to getting users to find items to purchase. On the app, this was hugely important (roughly half of users used it) but the experience was ill-designed, with a bloated UI, hidden filters, un-scannable posts and an inability to select multiple categories. I completely redesigned the UI to help get users to easily filter their searches and immediately recognise what they want in post listings.
▷Preservation vs. Progress
The site is "immediately recognisable." Accordingly, I had to to identify which "quirks" were functional and which were actual blockers. On the homepage, I tested rearranging the groupings. However, this lost some of the density of information. I also tried a more ‘modern’ look of the old site while emphasising core features like search and location, however, this didn't feel "Craigslist" enough, with enough brutalist quirk. Finally, I found a design that encompassed what we needed. It has the same ‘core’ page layout so it was immediately recognisable. Including system emojis felt personal and fun to the user and modernised the quirk. Iconic and useful features, like recent searches and missed connections were added as tabs so they could be easily findable. Other useful features could be dynamically incorporated into the page as it got larger. In this case, the option to create a search feed, and trending posts were added.
These elements were core JTBD that we found while speaking to users.
▷Preservation vs. Progress
The site is "immediately recognisable." Accordingly, I had to to identify which "quirks" were functional and which were actual blockers. On the homepage, I tested rearranging the groupings. However, this lost some of the density of information. I also tried a more ‘modern’ look of the old site while emphasising core features like search and location, however, this didn't feel "Craigslist" enough, with enough brutalist quirk. Finally, I found a design that encompassed what we needed. It has the same ‘core’ page layout so it was immediately recognisable. Including system emojis felt personal and fun to the user and modernised the quirk. Iconic and useful features, like recent searches and missed connections were added as tabs so they could be easily findable. Other useful features could be dynamically incorporated into the page as it got larger. In this case, the option to create a search feed, and trending posts were added.
These elements were core JTBD that we found while speaking to users.
▷ Building a scalable design system
I built an entirely new design system from scratch that addressed key usability issues (such as inconsistent buttons and no clear CTA colour) while enabling the devs to spin up a modular system that built on individual 'atoms'. I worked with the devs to create minimal, lightweight animations that could scale well across the site. The emphasis was on supporting the UX - they could never be superfluous.
▷ Building a scalable design system
I built an entirely new design system from scratch that addressed key usability issues (such as inconsistent buttons and no clear CTA colour) while enabling the devs to spin up a modular system that built on individual 'atoms'. I worked with the devs to create minimal, lightweight animations that could scale well across the site. The emphasis was on supporting the UX - they could never be superfluous.
▷ The Outcome
I personally did the high fidelity page for every page on the site. Additional pages in the future could be spun up easily by the devs using my design system. In testing, each page I redesigned averaged 35% fewer usability problems (even for legacy site users) and approximately 15% less time on task. Additionally, there was a 15% increase in post creation and 18% increase in scroll depth. Though these changes are in beta, these signals point to the site ceasing to lose 10% users year-on-year - saving $50 million a year.
Note that these are not the final designs due to confidentiality requirements. They are several iterations in, and are close to the final design.
▷ The Outcome
I personally did the high fidelity page for every page on the site. Additional pages in the future could be spun up easily by the devs using my design system. In testing, each page I redesigned averaged 35% fewer usability problems (even for legacy site users) and approximately 15% less time on task. Additionally, there was a 15% increase in post creation and 18% increase in scroll depth. Though these changes are in beta, these signals point to the site ceasing to lose 10% users year-on-year - saving $50 million a year.
Note that these are not the final designs due to confidentiality requirements. They are several iterations in, and are close to the final design.
