In this episode of McKinsey on Building Products, McKinsey partner Rikki Singh sits down with Yuhki Yamashita, chief product officer at Figma, a company that provides collaborative online design tools. With a career in product and design at companies of all sizes, Yamashita has an approach to product development that centers on generating “user love.” During their conversation, Singh and Yamashita discuss building and tapping into communities to make better products, how each team along the product development life cycle can work together for better results, and how to use generative AI (gen AI) to enhance customer-centric product design. An edited version of their conversation follows.
Where product and design meet
Rikki Singh: Tell us a bit about your background and career journey.
Yuhki Yamashita: I currently work at Figma, leading the product and design teams. Throughout my professional career, I have oscillated between product management and design, and I believe those boundaries should be more blurred than they can appear to be. When I was at Uber, we were experimenting with a variety of tools, and we experimentally brought Figma into the company. That’s how I was introduced to Figma and saw how it changed the way we work. I also noticed how it, too, was blurring the boundary between design and other functions—that’s what drew me to the mission.
Rikki Singh: How has that experience shaped how you think about product-led growth?
Yuhki Yamashita: We believe that our tool is for everyone, and everyone has a different way to contribute to the design process. We are an open tool that is accessible via URL, which has contributed so much to our growth because someone can share a design, and other people can use the link to open it and leave a comment. Those are the ways in which you find yourself in the tool, understanding how it works, and wanting to emulate the process. Reducing any boundary to get people to engage with the product is important—you don’t have to download a file or application. That URL feature allowed Figma to go viral within communities and companies.
Product-led growth inspired by connection
Rikki Singh: You touched upon two aspects: the community aspect of product-led growth [PLG] and the importance of lowering the barriers to adoption. Do you have examples of PLG initiatives that worked effectively?
Yuhki Yamashita: The biggest thing that contributed to PLG was taking advantage of the technology that was capable on the browser, making it multiplayer, and creating a new standard for product design. When I joined, we were working on an initiative to make Figma more extensible—to make it a platform that anyone can hook into and customize for their needs. One of the things that enabled us to do that was connecting to the internal tools that people were already using to create bridges and workflows for developers or other folks trying to consume the designs or participate via other tools. That connectivity helped create new scenarios and workflows that brought more folks to the product.
As a company, we want our users to authentically love what we built and feel like we are listening to them and evolving the product in a direction that they’re excited about. In the early days, when something wasn’t working well, we reached out to some of the most influential thinkers in the design community to get their feedback, followed up with our users when we changed things, and built loyalty.
Rikki Singh: How does every product manager [PM] or designer at Figma get that same feedback? What’s the most typical way to collect that kind of intel?
Yuhki Yamashita: We encourage people to spend time with our users and our communities. There are many ways to do that. The most accessible way is for people within the company to use the product. When I arrived at Figma, we were a document-heavy culture, especially when it came to running meetings. I changed it to a deck-based culture because that forced more folks, not just the designers, to use the product. PMs got involved, and it was a different view. We could uncover a new level of nuance and understanding of the product.
The second way we get feedback is by reading what customers are saying on support tickets or online. We share feedback with the rest of the company and see if something is going on that we can fix. We start to get drawn into those environments naturally to engage in dialogue with customers. You start to build relationships that way. A lot of members on my product team, for example, have relationships with customers that allow them to ping them and get casual feedback without making it feel like a formal survey.
Rikki Singh: How does that strategy translate to larger enterprises where the buyer may be different from the user? For example, the users may still be the designers, but the buyer could be either the IT team or a procurement team. How does Figma operate in that context?
Yuhki Yamashita: The way Figma spreads in those companies is often bottom-up: a small contingent of people try out Figma, and it starts to spread. Then someone from IT or procurement will help scale it in a secure way. Even in this environment, my teams think about the different personas and build relationships. There is usually a champion, such as an integrated-circuit designer or a design-systems person, who mediates some of the conversations with folks in the company who may not have as much visibility into Figma. We work hard to find a person who’s passionate about Figma, help elevate them, and help them make an internal case for why there should be a more strategic investment in Figma. The sales team shares the same perspective.
How product, sales, and marketing teams can work together
Rikki Singh: Which team owns the outcomes on product-led growth: product, sales, or marketing?
Yuhki Yamashita: It’s co-owned, but I think of it as a progression. To even be in the conversation, you need a product that’s amazing and that users love. Marketing has the superpower to expose the product to more people and amplify the message that gets people to connect to the product more. In our case, for example, there’s a team in marketing that builds a community for our product. Sales can then help people in the community who use Figma for self-service deploy the product in a more strategic way or on a larger scale.
Rikki Singh: As these groups collaborate, what do you view as some of the key enablers that help set up the PLG flywheel?
Yuhki Yamashita: For us, it’s the community. The design community is tight-knit, and designers in the community often learn from each other and share tips with each other. Then there’s the community of people inside the companies that use Figma For some of our large organizations, there are often advocates for Figma who share tips on the things that work well or who build internal plug-ins that make workflows more efficient. When those groups of people get together, a lot of amazing things happen.
Rikki Singh: Are there data aspects that inform your ability to assess that?
Yuhki Yamashita: We look at how people are using our product. Over the years, we’ve noticed that there are a lot of nondesigners that use Figma. In fact, two-thirds of our collective users are nondesigners. Knowing that, we knew we had an opportunity to figure out how we can serve those users better—there’s potential for us to deliver value to developers and product managers. In the past couple of years, we’ve been focused on developers, for example, so we can understand that persona better and make sure we can get to a place where developers love Figma as much as designers do.
How gen AI is revolutionizing design
Rikki Singh: How do you think gen AI will impact the product development life cycle?
Yuhki Yamashita: I think about it in a few different ways. First, AI has the capacity to both lower the floor and raise the ceiling. In other words, with AI capabilities, more people can participate in the design process in a meaningful way. So how can we equip people who otherwise might not have known how to get started with a starting point to express their ideas in a visual way? At the same time, how do we enable professional designers, who are intimately familiar with our tool, to work even faster or to be more expressive than they are today? Both goals are important for a company trying to make design more accessible and be a caretaker of the designer community.
AI also has the capacity to help us move from an idea to a final product in a rapid and condensed way. It allows us to ideate quickly, too, so we can visualize an idea, scrap it, and go back to the drawing board without the process being as time-consuming. The design process today can be tedious, so if we can find ways to save designers minutes or hours, then they can have the capacity to generate more ideas or think more about the user problem and strategy.
Rikki Singh: How do you see the roles of individuals evolving along the product development life cycle?
Yuhki Yamashita: I believe that the roles between disciplines have been blurring more over the past couple of decades. More product managers can design, more designers can do PM work, more engineers can design, and more designers can code. Before, design happened on someone’s computer offline and then was shown to other people in a very static way. Now, you can see what’s happening in real time and can inspect it and emulate it easily. When you’re iterating on work, you can get more feedback. With these changes, less focus goes into the execution details. Very rarely is a designer inside a company starting from scratch: designers assemble the building blocks that have already been created in a scaled way and operate at one level higher. More and more, we’ve been pushing designers further up the stack. With AI, this will happen more often.
Rikki Singh: Increasingly, folks are realizing they can give individual tools gen AI capabilities, but they don’t see instantaneous results. Figma has some great gen AI capabilities. What are you seeing as the behaviors that are emerging right now?
Yuhki Yamashita: We’re finding a lot of potential in things like similarity search. For example, you might be working on a design in a big company and not know another designer has worked on something similar. There are a lot of common patterns that people have figured out, so designers don’t have to reinvent the wheel. Similarity search allows you to borrow some ideas or inspirations.
Rikki Singh: Are there any internal experiments you’re running with your PMs, designers, and engineers to adopt some of these tools? What do you learn from these experiments?
Yuhki Yamashita: The similarity search feature was actually part of a path toward a different feature. We stumbled upon it, and this building block that we thought of as a means to an end was actually useful in and of itself. This is a very experimental age in which people have a lot of ideas, and the only way to prototype things is to play with them.
Overcoming potential challenges in an AI-native world
Rikki Singh: What are common challenges or pitfalls that companies or organizations could run into as they’re adopting some of these tools?
Yuhki Yamashita: We’re all still learning. Gen AI tools create a lot of efficiency, but they don’t replace the thinking. Even if we go from prompt to design in a high-fidelity way, someone still has to think critically about users and their needs. The biggest takeaway is that gen AI enables us to accelerate something that’s already there, but it’s not going to displace creativity or problem-solving.
Rikki Singh: What will be the next set of challenges that designers and PMs are going to confront in an AI-native world?
Yuhki Yamashita: AI-native interfaces are interesting. Some people ask us, for example, “Is the future all about chatbots?” The reality is that natural language isn’t always the best way to express yourself or manipulate something. Chat happens to be a convenient medium because it’s easy to place on top of your existing experience. Chatbots also have some familiar characteristics, such as a natural latency to make you feel like it is thinking and typing. As those features become less of a consideration, you can think more about how to find the right contextual ways for AI to be native inside certain experiences.
We’ll also see more-dynamic interfaces. As people build products, they will now have the capacity to rate interfaces in real time to allow for more flexibility and creativity. Before, each interface looked different across different types of users, and by trading off features, you can now make features more efficient for everybody because you can create great experiences on the fly.
How to cultivate ‘user love’
Rikki Singh: What are some tactics you use to build design communities at scale?
Yuhki Yamashita: We’ve been lucky to be part of a community that values community in the first place. Designers are excited to share their craft and make each other better. More broadly, we’re considering how to build similar kinds of communities among product managers and developers. In many cases, we’re learning as we go. I think it’s important to create the space and invest in any effort that builds communities, rather than assuming they already exist.
Rikki Singh: What is the most counterintuitive thing you’ve learned or insight you’ve gained over the course of your experiences across design and product teams?
Yuhki Yamashita: I think a lot about the contrast between starting my career in a larger company, learning the “right” way to be a product manager or the “right” way to do product thinking, and then moving to much smaller, scrappier environments where a lot of those fundamental assumptions were overturned. For example, in a larger company, you’re encouraged to focus on defining problems, setting up goals, and making sure those goals are measurable. Whereas at Figma, we care more about what we’re building and how happy our customers are with it. We can define a problem or metrics, but they don’t matter because we don’t get credit from users for having defined them. There have been many times when we’ve launched or announced a feature, and we had to iterate on it because we got a lot of strong feedback.
We recently redesigned our Figma Editor. We had a vision to make the editor feel more immersive and, as part of that, the design had floating panels that users could push away. But people commented on how distracting the panels were because they took the focus away from the work. So we reversed that feature. I think our community was appreciative that we did that instead of sticking to the decision because this showed conviction.
Rikki Singh: What advice would you have for other organizations that are embarking on their product-led growth journeys and aspiring to build their communities?
Yuhki Yamashita: You hear about all these frameworks that you can follow to find product–market fit or the right metrics, but at the end of the day, “user love” is very hard to measure precisely or optimize. You have to spend time building relationships in a very human way so users can give you honest feedback. That way, you won’t have an illusion about how much people love your product. The biggest piece of advice I have is to spend time with your users and not focus too much on a particular goal or metric. A lot of great things can happen.
