Empowering users to manage recommended content on Shop
I designed the content for the Shop app’s hide and report flows, which collect negative feedback on products and shops recommended to users. I also designed the new user education experience that introduced users to a multitude of new actions they could take with recommended content, giving them the ability to more carefully curate their in-app experience. The result: an unanticipated 2.61% increase in GMV for our merchants, a 0.65% increase in high-quality clicks on shops, and more than 230,000 potentially NSWF products hidden from user Home feeds.
The problem
As the Shop app began driving more in-app sign-ups, it became clear that we needed to resolve a prickly problem across the experience: a fairly nascent data model served too many users with recommendations for shops and products they considered lewd, obscene, or offensive across their Home feeds. We recognized that we inadvertently exposed users to a negative experience, which ultimately would lead to dropoff, attrition, and general bad feelings about the app.
The proposal
The team and I developed a UX strategy to offer users more control over what they see and don’t see in the app. We developed a long press UI for hiding and reporting undesirable content. I designed the content for the hide and report flows, which collect feedback on products and shops we recommend. Users can hide any shop or product they deem inappropriate, and if the offense is particularly egregious, we give them an opportunity to report it.
Content was the connective tissue in this project, as the language we used needed to clearly inform users of their options while reflecting their preferences depending on whether they were hiding or reporting a shop or product. For instance, we ask a user who decides to hide a product their reason for hiding it — data we use to inform what kinds of recommendations we should and shouldn’t show them in the future. For hiding, we ask about the nature of the content they’re concealing to inform our data model about what kinds of shops and products are unsuitable for specific users. The outcome is that users can tune their shopping feeds to their tastes without haplessly running across content that may sour their experience or cause unintentional harm.
Within a month of shipping these new options, we saw more than 230,000 products hidden and more than 30,000 products reported. That nets to about 500 products reported a day, exceeding our expectations. At first blush, users hiding and reporting this many products may feel staggering, even deflating, for a shopping app. But our team’s perspective was different: this piece of altruistic UX helps users clean their feed, improving their product experience and motivating them to regularly use the app without fear of being spammed by violent, lewd, or useless. content. Simply: it was the right thing to do.
Maximizing the long press
Our work wasn’t done at hide and report. We didn’t simply want this new UI to help users obfuscate products and merchants that didn’t hold their interest. We wanted to utilize the same interaction to elevate their overall shopping experience in a more overtly positive way. So, we folded new actions such as adding items to the cart, sharing products with friends and family, and viewing similar products into the long press UI.
This next iteration of the project was primarily content-led as I crafted the appropriate messaging necessary to educate users about their new shopping options. I explored a tooltip and bottom sheet as the primary UI for delivering this message, undergoing several iterations and permutations. After a few days of exploration, I determined the bottom sheet was the most effective NUX pattern in this scenario as it allowed for more breadth in the message without tying the content to one particular UI element on the Home feed like a tooltip would. I worked with my engineering and product management partners to determine the ideal logic behind rendering an intrusive pattern like this on the Home feed. I presented the work in stakeholder and Design reviews and drove alignment with Product and UX leadership, even showing off the UI during a team-wide demo. Finally, I codified the terminology we’d use to characterize the long press in perpetuity (i.e., calling it “press and hold” instead of the jargon-heavy “long press”).
The outcome
We shipped the long press NUX as an experiment to determine whether adding it to the Home feed would produce any negative outcome. Fortunately, the bottom sheet didn’t create any adverse impact on the Home feed’s overall UX. In fact, we were able to attribute the NUX to an unexpected 2.61% increase in gross merchandising value for our merchants and a 0.65% increase in high-quality clicks — a metric that helps us gauge whether users are intentionally viewing shops and products.