Project Brief
Position
Product Designer
Team Count
3
Platform
Web + App
Background
Houzz.com offers 3 distinct verticals: Ideas, Marketplace, and Professionals. Ideas allows users to get home inspiration, Marketplace is where users can shop for home goods, and Professionals facilitates the hiring of professionals for various home projects. Because of these three distinct verticals, each vertical homepage is a crucial touchpoint for users coming to the site, especially for navigating a marketplace.
Problem + Research
Houzz.com provides a unique challenge with its multi-vertical offerings: once users navigate to marketplace we need to ensure their experience prioritizes what they are looking for because there are three times as many places for a user to navigate onsite.
Testing across the Houzz marketplace has shown that users tend to engage very well with personalized shopping modules. Despite this, there was no personalized content or ways for users to re-enter their shopping journeys from prior sessions on the control experience. As well, the experience lacked branding and a modern UI. Among this massive intent gap, there were other major pain points with the page:
Pain Points:
1. No personalized browsing continuity modules
2. Users have to leaf through many categories before arriving at browsing products
3. Recommended products are buried at the bottom of the page
The question then became, how could we improve users shopping journeys starting at the marketplace landing page?
Competitive Analysis
Taking a look at competitors, the most commonly included personalized modules were recently viewed products or products recently added to cart. There were also entry points for categories users recently browsed and suggestions based on prior interactions. As well, each landing page made sure to give users full context to the shopping experience on their platforms, like showcasing branding, recommendation algorithms, and clear navigation.
Hypothesis
If we create a dynamic personalized shop homepage then we can create browsing continuity from this page, reducing the amount of effort it takes users to get where they want to go.
Since the control experience is bulky and removes users from their previous shopping journeys and the treatment would get users closer to where they left off their shopping journey, we would see improved marketplace metrics (orders, GMV, profits).
V1 Design
For the first round of testing we did on this page, we started with a design that prominently featured recommend content based on prior user intent. The new design included: Personalized modules — “Recommended Categories” to browse and “Recently Viewed” products.
As well, to give the page a more dynamic feel I included a rotating sale banner. Then, to keep navigation streamlined, I prominently featured department tiles.
Further down on the page, I included entrypoints to “Trending Categories” which gets users right to browsing popular products. At the very bottom of the page I added Recommendation carousels to continue providing users product offerings.
Testing Phase 1
After extensive QA, the filter treatment was tested on Houzz.com at 50/50 for 2 months for mobile users.
Control
Treatment
Results
On the Houzz App, the treatment was ramped to 100%. Although we saw success on App, the web experiences did not win out. So we went back to the drawing board to figure out how we could replicate these wins on Web.
Phase 2 Hypothesis
Because we didn't see success on Web but did on app for this redesign, we had to really think. Why did it work on our app? Was it the difference between signed in and signed out users? Was it because we pushed on personalized content too hard? What are web users looking for?
Our analysis of why the new treatment didn't work revealed a few things — on control users can very easily see the departments and there isn't much other content on the page. In treatment, we condensed the departments and also included more content to the page. The intention of the page didn't feel as clear—which can definitely hurt performance for signed out users who are not as familiar with Houzz's product. In order for users to use the Houzz App they have to be signed in. This means these users are more familiar with the product and understand how to navigate.
For the next iteration , then, we decided to take a more aggressive approach to really communicate all that Houzz offers.
V2 Design
For this next round of designs we wanted to lean even further into making the page feel more dynamic which would be a strong contrast to control.
We also wanted to create more visual variation to make users feel like they had access to more types of content. This would make the page feel more curated, more elevated, and more personal.
The hope with the updated designs was to see better engagement with the different modules.
Testing Phase 2
This treatment was run at 50/50 for 1 month
Control
Treatment
Results
After testing, treatment was ramped to 100%. Clicks were up 4.52% and page views were up 4.19% for down funnel products, with a .24% lift in conversion rate. For mobile users of Houzz, which represents more than 50% of traffic to our site, this is what they will see when the arrive at the marketplace homepage. This test is still running on dWeb.
© 2024