Pinterest is a visual discovery engine for finding ideas like recipes, home and style inspiration, and more.
Katie Hurley, Vice President of Marketing at unitQ, recently spoke with Rodion Gusev to learn more about his role as Manager II Technical Program Management at Pinterest, and how he uses unitQ on the job.
A transcription of their discussion, edited for clarity and length, follows:
Katie: Hello. It's great to have you here, Rodion. I'm very excited to speak with you. I know that you work at Pinterest, and before we get started, mind telling everyone what you do at Pinterest and a little bit about the company?
Rodion: Hi, Katie. Thanks for having me. Yeah, absolutely. I'm Rodion. I lead up the Strategy and Operations Team for User Services at Pinterest. Generally speaking, we're the team that handles anything involving data, tooling, insights, and infrastructure for Pinterest support provisions. And Pinterest, in general, is a visual discovery platform that's meant to help people discover and do the things that they love.
Katie: So tell me about how you are using unitQ to ingest user feedback and learning what your users care about?
Rodion: Our primary use case for unitQ is essentially making sure that we're able to identify larger, aggregate trends involving things that users are encountering in a way that we might be missing if we were just dealing with a single channel or with just human-level aggregation.
So if something isn't working for Pinterest users, they might flag it on Reddit. They might flag it in an app store review. They might flag it as direct Support tickets. But in a large organization, all those are separate teams with separate people, and they might not necessarily have the wherewithal to realize, hey, a 1% spike in mentions of a particular thing here, plus a 10% spike in a Support channel, is actually something that we need to immediately flag to Product. So unitQ helps us understand user pain points and in general, what users are talking about or writing about across multiple channels, which generally increases our ability to quickly identify specific issues or points of interest. And then we can take action on them.
Katie: What was the state of gathering user feedback before unitQ? Was it manual?
Rodion: Well, we definitely have always had a lot of automation and NLP-based stuff happening at our Support verticals. So thankfully, it wasn't that manual to start. But with unitQ, we've been able to iterate a lot faster than we could with self management. So previously, we would look at things like keyword searches combined with user-form selection on structured submissions, combined with somebody actually manually going through things.
So we were definitely less proactive, and a lot more reactive pre-unitQ.
Katie: So maybe we could walk through how unitQ has improved Support, Engineering and Product roles?
Rodion: unitQ in general was implemented on our end as a way to increase our partnership with Quality Engineering. The Engineering Team is specifically responsible for rudimentary app stability. They're not the ones necessarily trying to push out completely new features. But once those features are developed, they're the ones who are making sure that they function within the broader ecosystem of a big engineering org where lots of people are pushing various code out into an end-user product.
So the main reason that we started using unitQ as an individual use case was just like, how do we make sure that it's the easiest possible turnaround from us seeing things come into Support and an Engineering Team understanding that there's something happening that needs to be worked on? So that allowed us to really quickly show what impact was happening on a particular case. Because unitQ also tags stuff, we would then be able to really extrapolate that to an existing Jira filing, or something like that, to make it actionable for teams. Anything that writes tags into our Support content management system creates a lot of flexibility around what we could do with it.
So in addition to having a really easy point-to-click view for Engineering and Product to look at what's going on across all surfaces, we have been able to use unitQ tags to automate Support responses. So if we know that a big spike of tickets that are coming in are tied to a specific unitQ assessed bug, we're able to bulk respond to those items.






