"Discovery" is thrown around a lot as an end in itself for new entertainment and commerce platforms. But at Pinterest, which has been described as a search engine for people who don't know what they're looking for, it's a true core-value proposition.
Finding an inspiring new thing on Pinterest should, says Pinterest's head of product Tim Kendall, inspire "curiosity that haunts users to keep exploring the surface," and be "compelled to actually take action on that interest. If you see a hotel on the coast that you never even heard of, you’re compelled to maybe book a weekend there in a couple weeks. If you see a recipe, maybe you’re compelled to go to the store to buy those ingredients and then go home and cook that recipe that night."
This principle, says Kendall, is the starting point for any new feature introduced to the platform, and is what all product teams are charged with improving when coming up with concrete enhancements. For example, a key feature launched this year is guided search, which lets users start with a broad idea (say, chicken) and telescope down to something specific (say, cheesy chicken enchiladas) by offering easily navigable guides in different categories related to the original term.
Kendall says that to successfully turn an abstract concept like discovery into technically sophisticated, user-friendly functionalities, he employs a specific interdisciplinary approach to product development, with three key elements.
When a new feature concept goes into development at Pinterest, it doesn't get assigned to an existing group, but to a carefully selected team of two or three people maximum, usually an engineer and a designer, and if it's a particularly nebulous user-facing problem, a user researcher.
"We mix and match those teams very aggressively," says Kendall. "We think when we get the magic combo of the right two or three people, we just see an order of magnitude difference in terms of what they’re able to come up with and how fast it happens."
The small size is important because the team is better able to focus on what might start as a very broad problem, such as "search is too complex," that needs to be brought down to a very concrete technical feature. "They feel like a mini startup," says Kendall. "When people feel empowered, they come up with great things."
After the teams are formed, says Kendall, Pinterest encourages them to get out of the confines of the office. "When we built guided search, for instance, in February, the designer and the two engineers went off to the designer's house and camped in their basement and just iterated on the product for several days," Kendall said. "We also have a team that just recently started working on some messaging products and on their own they flew to Portland, rented an Airbnb for five days, and came back Friday night with an end-to-end completed prototype."
When you go to the same office every day, says Kendall, "it's very hard to clear your brain from all the legacy projects and legacy relationships that you have, especially from being somewhere for several years. If you go someplace totally different, that environment will embody the very project that you're working on and that creates a level of focus that gives us results and really great, great stuff."
"One of our primary core values is something we called knitting, which captures this idea of different disciplines coming together and 'knitting,' creating something more powerful than they can create on their own," says Kendall.
"We often have an interaction designer paired up with a core engineer. What often happens at companies I’ve been at previously is they’ll have a designer take a problem, go off and come up with mock-up. It's really important on day one that the designers don't go off in a corner and start drawing pictures, but that instead the designer gets in a room with the core engineers and they start to understand what is possible, from each of their disciplines, that can create a feature that really is the embodiment of their disciplines.
On guided search, for example, the team started planning the interface and then started looking at the potential technology to power that feature. "In fact, to enable the interface that we arrived at, we actually invented a derivative technology to power it. I don't think we would have broken through with that kind of innovation if we hadn't put those teams together and had them knitting from day one."