CX may not be rocket science, but at Pinterest, Kitty Xu has found ample opportunity to make use of her background in neuroscience.
As senior quantitive user experience lead at the San Francisco-based image sharing service, Xu plays a key role in helping to gather knowledge that will help product designers and software engineers improve the way Pinterest works.
This involves both the kind of quantitive research that most CX leaders are used to — customer surveys — as well as the analysis of login data that shows how Pinterest users work with various features in the app.
“When I came to industry, I found the research methodology was similar (to academia),” Wu reflected in a panel discussion during the Customer Data, Insights & Analytics virtual summit hosted by CX Network this week. “The key difference is that in academia the results of research can take a long time. In industry no one is going to wait for five years.”
Xu suggested this becomes an important part of the discussion with stakeholders within an organization that want to have scientific-level rigour applied to the CX problems they’re tackling. She said it’s critical to agree on whether research can be conducted over a quarter, a month or even a few weeks. In some cases, the urgency to make changes to a product could affect whether it’s feasible to do research at all.
Much of what Xu researches and quantifies is sentiment — in other words, how do customers feel when they are using Pinterest to find a new recipe, ideas to redecorate their living room or an outfit to add to their wardrobe.
Though it was founded in 2009, Pinterest continues to evolve the way its users can find and share content on its platform. According to a recent report from Social Media Today, for example, the company has been testing a scrolling, vertical feed of content similar to TikTok, with a new ‘Watch’ mode being tested for Pin discovery. A number of new shopping features, meanwhile, such as product tagging, are being rolled out to more countries after Pinterest initially launched them.
Insights from multiple data sources
When she’s using surveys, Xu said she tries to approach users with in-app questions that assess their comprehension of the product, their key challenges, their motivation for using Pinterest and any ideas they might have for improvement. Log data, on the other hand, can show what happens when they are actually using the app.
“Customers” in this case represent not only app users or “Pinners” but content creators who specialize in designing images to be pinned, as well as brands who might work with Pinterest as an advertising partner.
Surveys and log data aren’t the only sources Wu studies, however. In some cases, she said it’s possible to create a fuller picture by working with colleagues doing more qualitative research by creating user stories based on in-depth interviews. She also looks at customer service and support tickets, which can help identify areas of friction.
“On Pinterest, one of the challenges could be comprehension of the product. Some people don’t really know what Pinterest is for, and think Pinterest is a social media service, which it’s not,” she said. “The research becomes an exercise in figuring out what are they mistaking it for — Instagram, Facebook or a Google search?”
Of course, some of the data gathered in the research can’t be taken at face value, Wu said.
“Frankly a lot of people, when they’re writing ticket, are extremely satisfied or extremely dissatisfied,” she said. “So you need more research to determine how representative is that?”
This is where behavioural data really becomes helpful. A Pinterest user may say they can’t find the inspiration they’re hoping for, for instance, but log data could show they’ve never actually used the app’s search function.
If that becomes a trend, it could mean Pinterest has to do more to surface its search function in the app experience, Wu said. She also said it’s important to think about what users aren’t saying about their experience, and whether that points to additional improvements to be made.
Acting on the data is obviously the next step, but Wu advised establishing how stakeholders intend to use the research before it is conducted. That makes it easier to ensure everyone is on the same page.
There is also room for creativity in how you act on the data. If people aren’t sure what Pinterest is, she gave the idea of pairing an experienced user with someone new to the platform and having the former explain Pinterest in their own words. This can be translated into messages that are shown in the app when people sign up, or which the company’s marketing team can use in other channels.
Wu also said that while researchers should tell stories with data, it shouldn’t necessarily be about the data. Instead, she tries to tailor it to her audience, particularly if they are designers rather than data scientists. In many organizations, a big part of the value is not in simply doing research but prioritizing what needs to be studied. After all, Wu pointed out, product and engineering resources in many firms are limited in what they can do in a given period.
“People can come into Pinterest with baggage from other apps they’e been using and expect to see the same features,” she said. “We have to take that and ask is this aligned with our product mission? Or, what is the customer really asking for?”
The CX Network session, which also featured experts from Cashapp and Coca Cola, its well worth viewing in its entirety and is available on demand.
Shane Schick tells stories that help people innovate, and to manage the change innovation brings. He is the former Editor-in-Chief of Marketing magazine and has also been Vice-President, Content & Community (Editor-in-Chief), at IT World Canada, a technology columnist with the Globe and Mail and Yahoo Canada and is the founding editor of ITBusiness.ca. Shane has been recognized for journalistic excellence by the Canadian Advanced Technology Alliance and the Canadian Online Publishing Awards.