How is your listing performing?
Claudia, 24 is currently working as a Sales Assistant. She declutters her home every now and then and lists the items she doesn't want as for sale items in Carousell. She managed to sell some of her old items she never thought she could sell. Her secret? She tracks how her listing is performing, and tweaks the photos, description, and drops the price.
How did we build listing insights for Claudia?
Carousell is a C2C Marketplace app expanding from South East Asia. The product focuses on casual sellers, like people with a day job or school to de-clutter their space by start selling.
Who I worked with?
I closely worked with Alla, The Product Manager, Andrius, The VP of Product, and amazing Seller Services Scrum team to make this possible. This project wouldn't be successful without help from Community, Legal and business team at Carousell.
Web and Mobile Web.
How Long did it take?
One and a half sprint(3weeks) for Listing insights We work in Sprints, and as a designer, I work ahead of the engineering sprints.
Understanding the problem statement from business/users
Breaking down the problem statement is the best way to understand the core problem and it helps us, as a team to solve what's necessary.
- Sellers are unable to make a sale after they list an item, immediately
- Sellers measure likes, chats and offer sent for the listing as exposure and they feel they aren't getting any after initial few after posting.
- Sellers are worried they can't sell in Carousell
As Seller services team, our goal is to identify problems that affect our seller's success and how to build additional tools that addresses the problems in exchange for a small price. We are the team that focuses on both user and business problems and we measure our success by seeing whether we solve user problem and reach business goals.
We broke up the above 3 problems in two How might we statements,
- How might we show our sellers their item's real exposure
- How might we give our sellers a tool, that would get them more exposure, so that they can sell fast?
Currently, sellers were measuring the exposure through a number of chats, offers, and likes they get and evaluate whether their listing is performing well or not. While they are correct, these numbers just mean the buyers actively engaged with their listings, and doesn't correspond to the passive interactions that shows the buying intent.
We wanted to show a value that is also passive and less user intrusive. We decided to go with views. Views are traditionally characterised to show exposure and users don’t need extra learning to understand the benefits of views. More views = more eyes on the listing.
The initial idea involved around abstracting the views into simpler forms like meter, temperature, bubble. While all these ideas are quirky and fun, they didn’t come across as transparent as showing the actual views. Based on the competitor research, my product manager did, we realised that each view count matters and since it only takes one buyer to view it and offer the seller, we felt it's best we show the real views, upfront and clear
We know that a seller, after they list the item, land on the profile page, because of our listing journey. We also know that sellers, after they list their item tend to revisit the profile page to see how can they improve the listing. We added two call to actions, one on their listing card and another on their listing details page.
Listing insights Design
Our main goal of this problem is to give our sellers an insightful page that shows their real exposure, and not just views. We ideally saw this as a place where all the future analytics will sit on. We named this page Listing insights, and like it says, it shows insights about your listing and how can you boost your exposure.
We wanted to show the views upfront and transparent. Our team ideation session helped us in understanding the engineering feasibility. With that in my mind, we isolated views into one screen and came up with a structure to show the content. We want this page to be accessible from different places in the journey and we want to show the most liquid information in the top, then show the trend of views for their listing for the entire week. We felt for casual users, views alone doesn’t make any sense, so we built the tips section which helps them to understand how to get more views.
Tips for how can they sell better. We wanted to show our best tips to get more views. This is from buyer’s point of view and how would they find a listing good. These tips would be for first-time sellers too if they are looking for ways to boost their views.
When the item goes reserved, the item no longer exists in the marketplace, so will likely not receive any more views. But there is a chance of item getting unreserved, and goes back to the marketplace. So we decided to have listing insights available when the item is in Reserved state, but not when it's Sold
Edge cases and issues
Issue 1: Listing insights involved a launch of new service to count the views better from the iOS app. This means that the app will be unable to show the number of views prior to the release or from the old app. This is largely due to the timeline and in terms of design, the product team decided to support this one time issue.
Issue 2: There are a lot of reasons why a page in an app won’t load and poor internet connection is one of the common reasons for why the page fails. We wanted the listing details to fail gracefully and not disrupt the entire page. We decided to have a retry, upon failure when our server calls fail, consecutively for X times.
Even though the best experience is the one where users don't have to encounter this experience, the compromise we might make would help us ship faster. I learnt that as product designers, we constantly need to thrive for a good user experience that fights the unknowns from the users. These are not use-cases we normally think while ideating and it’s extremely challenging to come up with a solution to the latter part in design process, which is both easy to implement and clear to the users.
Validation and Takeaways
Internal User testing We wanted to move to the next problem faster, so we settled for an internal user testing as a way to validate this solution. I recruited few users inside office who are non-tech to get their feedback and how they feel about a Marvel prototype I showed. The main goal for me is to understand if the users would miss out / spot the new listing insights page. Unfortunately, I didn’t find any good insights to validate that because most of the internal users spotted the difference, but may be because they were also pro Carousellers. We even discovered a lot more use-cases for listing insights that we never thought before. But the discoverability problem remained unsolved for now.
General Feedback We decided to launch the app into all markets to see how users feel about the new listing insights.
The overall feedback was positive and users wanted more out of listing insights, ranging from impressions (in search/browsing) to whether they saw all the photos in the listing. There was, of course, negative reaction and users quickly realised that their listing views go down after the first day. This also pushed users to re-list their items in the marketplace so they can get a first day visibility. Some pro sellers felt the impressions needs leads/source so they can identify what drives their sales since they run external marketing with carousell listing links.
The solution was declared success after few weeks and it built the foundation for Seller services team to work on additional problems later. We were pretty proud of what we achieved in short time and we continued to iterate on this problem.
Read the next case study
Case studies from my work & fun Projects