YouTube chief product officer Neal Mohan joined Decoder this week to discuss YouTube’s new $100 million fund to begin paying creators who use YouTube Shorts, which is its competitor to TikTok. Mohan is bullish on Shorts, which he connects to YouTube’s earliest days as a way to quickly share personal videos.
But YouTube is now much more than that; it remains the default video hosting platform for the entire internet, in a way that can feel almost invisible. It operates a linear TV service that competes with cable providers and a music streaming service. And, of course, there are YouTubers, the influencers at the center of the creator economy — the people who have turned YouTube into not only careers, but multimillion dollar businesses that extend into everything from merch drops to cheeseburger restaurants. When people talk about creators and the creator economy, they’re often just talking about YouTube.
Two notes: first, I encourage you to listen to this episode if you can. We love providing these transcripts, but for this episode I think there’s a lot that gets lost in the back and forth.
Second, at the end of this interview I asked Neal to let me watch YouTube videos at 1.5 speed on my TV. He said he’d heard the request before and he’d look into it. Turns out, that feature was shipped a month ago, and neither of us knew about it.
Because YouTube is so big. And so is this conversation.
This transcript has been lightly edited for clarity.
There’s a new monetization system for YouTube Shorts. It seems very interesting — it’s a big program, and you’re a full participant in the creator economy. But I want to start at the beginning with some very simple questions.
What does the chief product officer at YouTube do all day?
My responsibility is really to help run YouTube. I’m responsible for all of our products, everything that you use as a viewer of YouTube, everything our creators use. I work very closely with our partners across Google in terms of solutions for our advertising partners.
So: all of our products, from the main app to the Kids app to the Music app. And I’m also responsible for our trust and safety organization, so everything that has to do with our content policies, what governs the type of content that we allow on our platform, the content that we take down. We call them our community guidelines, and so my teams are responsible for that, as well as the design of our products, how they look and feel, how they work, [the] critical user journeys.
That’s really what my team and I do all day, every day.
It’s unique, I think, among all the social platforms that the trust and safety team reports to the product team. I do want to dive into that, but just give me a sense of your day: what kind of meetings are you in?
I’m in a lot of meetings. Everything from one-on-ones to very large meetings.
But maybe I could give you a flavor of a couple of canonical meetings which, in my view, are the critical decision-making meetings at YouTube as they relate to product. A lot of the types of meetings that I’m in are what we call “product reviews.” They are meetings that could be about any of our products or features across all of YouTube, oftentimes involving multiple surfaces, multiple teams. There are venues where we are able to get a status of how a particular product is doing in terms of its roadmap, development, readiness to launch, etc.
But they are, critically, decision-making meetings. By the time those types of decisions come to me, they’ve of course gone through multiple levels within my organization, and haven’t been able to get resolved for one reason or the other. Oftentimes, my job is to help tie-break across them, and make what is essentially a trade-off between whether we should do a feature this way, or whether we should do it that way. That’s typically a 30- to 45-minute discussion, and we have lots of those throughout the course of a week.
Give me an example. What’s a tie you recently broke?
The interesting thing about YouTube — which will sound obvious but is actually something that’s pretty fundamental to all of our product decisions and in fact many of our policy decisions — is that it’s an app, but it’s really an ecosystem. That term gets used a lot, but in the YouTube context, it means a balance between viewers, creators, advertisers, and partners that are all participants in our ecosystem. Oftentimes, we are making decisions in our products that have a pretty big impact on all of those or some of those constituents.
And so, an example of some of those types of decisions; if we design a feature this way, is it going to be something that is easy for our creators to use and understand? Is it something that we need to put front and center? Is it going to be cluttered that way?
For example, if you’re a creator, one of the tools that you use a lot is called YouTube Studio. It’s the place where you go to upload your videos and where you understand the stats around your videos. An example of a trade-off there is: if we have a new metric that we’re bringing to creators to measure the efficacy of how their videos are doing — are they getting the audience they’re looking for? — we can put that front and center, or we can put that in a flow that might be behind a couple of clicks in a way where it might not be as prominent, but it’s alongside other reports and metrics that make sense to the user. That’s an example of a relatively straightforward type of trade-off, but my teams are making hundreds of those types of decisions on a regular basis.
I think, just pulling it back up from that very specific example, what I try to do in terms of all of these meetings is really to establish principles. What is the framework or the set of principles by which we would make those types of decisions? And one of those north stars is how effective, how useful it is for our creators, based on their input, but also just based on the data that we see in terms of how they use it, are they getting value out of it.
I always ask this question: what’s your decision-making framework? You lead right into it; you’ve got some principles. What are the two or three core principles that come up in every decision for you?
I think actually there’s three critical things that every company or organization needs to think about when it comes to decisions. The first are our principles. And so, I mentioned putting our creators first, putting our viewers first, just like a uber-north-star-type principle. Another principle that comes up in the context of YouTube is: YouTube is an open platform, but it also has community guidelines. How do you balance those seemingly two competing principles?
That’s an example of principles that go into decision-making, but I don’t think it’s just about principles in terms of good decisions.
I think that there’s a couple other things that are just as important. The second is obvious: the people that make decisions. There are thousands of decisions being made across the YouTube organization every single day on behalf of our viewers and our creators. Of course, I’m not making all of them. What’s important there, for me, is to make sure that I have the right people throughout my organization to be able to make those decisions, because I can’t be there all the time.
And the third thing, which I think often gets overlooked, [are] the processes by which you actually make those decisions. And so, you and I just talked about one of those processes in the context of meetings — product reviews. How are they set up, what’s the cadence, what’s the criteria by which decisions should come to those types of meetings? How do you measure the impact of those decisions? We have a process at YouTube called our OKR [objectives and key results] process. That’s a quarterly, yearly process.
And so, I think those three things, principles, people, and then the processes by which you manage all of it, go into any effective decision-making framework. That’s how I think about things as it relates to my organization at YouTube.
You mentioned people. You have a huge remit. How many direct reports do you have?
I’d say I have about eight or nine direct reports.
How is that structured?
I have various leaders for the various product areas, many of which you would be able to guess from the outside. We have teams of people, including a leader, focused on tools for our creators. We have leaders that are focused on trust and safety, both on the product side and on the operation side. I have a leader that’s focused on overall design and the UX and user research framework. Leaders focused on our core experiences, Music, Premium, YouTube TV...
You’re way over nine.
Some of these are lumped together. For example, YouTube Music and YouTube Premium are under one product leader on my team. We have a leader who’s focused on what we call our vertical experiences, community experiences. I know you alluded to this at the top around YouTube Shorts. That’s a new product. I have a leader that’s focused on those types of community experiences. So it’s roughly about eight or nine leaders.
Let’s talk about Shorts. Shorts is a new product that launched in India, but it’s now in the US, and now it is at the top level of the YouTube app on the home screen. It’s one click away, and there’s a new monetization system for it. That’s a pretty fast waterfall of emphasis, right? It was in one market as a test, now it’s on the top level of the app, now there’s a monetization fund. Walk me through that process.
The way I would give you some insight into how all of this came about is actually going back 15 years to the very first video that was uploaded to YouTube, which is kind of a canonical, famous video now: “Me at the Zoo.”
That was an 18-second video that was uploaded in the San Diego Zoo, and it was the genesis of YouTube. Back then, “user-generated content” was a new term. 15 years later, the world has evolved dramatically.
One of the biggest differences has been the mobile phone and creation on mobile devices; you have this incredible camera and all of these incredibly effective editing tools right there on your phone. If you take that 18-second video that happened 15 years ago and fast-forward it to today, if something like that was being created today, how would it be done? Well, it’d be done on a mobile device. All the editing horsepower would be right there on the device. You might be shooting it vertically instead of horizontally.
Shorts tries to address how video is made now
So my team and I put ourselves in the shoes of today’s creators. [That’s] really where Shorts came from. Unlike 15 years ago, if you were a creator just getting started on YouTube today, how would you go about doing that? And that was the key product insight or genesis, if you will, of YouTube Shorts.
You asked a couple of other questions in there in terms of how it got from that idea to where it is today. I’d say a few things; first of all, as you know, YouTube is a global platform. Some of our largest and fastest-growing markets are outside of the US and North America. You mentioned one of them, India. Increasingly, we have to think about things in terms of really trying to launch globally as quickly as possible. That’s always been a mantra at Google and YouTube, but we really want to live up to that with all of our products. And Shorts was an example of that, where, as you pointed out, we did launch in India first, then we came to the US, and now, of course, it’s global as of the last couple of weeks. That was a little bit of the insight in terms of why we decided to go about that route.
We’ve seen growth globally, across all of our markets, as we’ve rolled out this product. I think on the earnings call a couple of days ago, Sundar [Pichai] mentioned 15 billion views globally across Shorts, and that number continues to grow. So that’s how we got to where we are in terms of the core product. It is critically important, from a viewer standpoint and a creator standpoint, to the overall YouTube experience. I’ve talked about that and Susan [Wojcicki], our CEO, has talked about that in the past.
That’s why you see its prominence in the app. That’s why you see it as a tab. I would argue that there are other surfaces that, from a user standpoint, are even more important. You see it right below a video that you watch, right? You can tap that “create” icon right below a video. And that’s actually what gives that some of its YouTubey flavor, because it ties this new format, Shorts, to existing videos in our platform.
Then the third part of your question, which is monetization. I think this also ties back to where YouTube came from, the genesis of YouTube. There’s a lot of conversation today about the new creator economy. That’s almost become a buzzword. I think sometimes people lose sight of the fact that YouTube has been in the creator economy business for over a decade. [It’s been] 14 years since we launched the YouTube Partner Program. In the last three years, we’ve paid out over $30 billion to our creators, and the Shorts fund, this $100 million fund, is really just another step in that journey, and is really the first step in terms of figuring out what the long-term monetization program could look like for Shorts creators.
Sorry for the long answer, but I thought I’d give you the full color.
I’m here for it. I have follow-up questions for everything you said. Let’s start at the start. Your connection to the early days of YouTube and how you’d make “My Day at the Zoo” now, I buy it. I have a more cynical version of the story though, and I want you to clarify it for me.
The more cynical version is: Snapchat launched Stories, and then Instagram launched Stories, and then WhatsApp launched Stories, and then YouTube launched Stories, and then LinkedIn launched Stories, and now there’s Stories everywhere.
And then TikTok came out, and TikTok is a cultural phenomenon. And now there’s something that looks exactly like TikTok in Instagram, and there is Shorts, which looks exactly like TikTok in YouTube.
That feels like maybe an unfairly cynical reading, but it’s also definitely the correct timeline. Do you think of Shorts as a direct competitor to TikTok?
I’ll put it in context from my perspective. Thinking about things from a creator standpoint — you’re a video creator or you’re a creator that’s looking to build an audience — personally, I believe that it’s really great that there’s lots of platforms, lots of choices, lots of different ways that you can build an audience. I would argue that all of these platforms, while they might seem similar in many ways, are fundamentally very, very different, but I actually think that’s great for creators because it gives them a diversity of options. That’s what I would say first and foremost.
I would just say that we look at the Shorts product through the lens of “simple, fast, easy, but powerful mobile creation.” Ten years ago, you would have a camera, you’d have a tripod, you’d set it up in your family room or in your backyard or in your bedroom, and you would start vlogging. I really think the world is very different [now]. And, as you know, many parts of the world are leapfrogging that generation completely with the prevalence of mobile phones and the power on those devices. I really do look at what we’re doing with Shorts through that lens, and I think the roadmap that we have will also prove out that we’re thinking about those pieces.
Shorts can pull audio from YouTube videos or link out to them
When you think about it through a mobile creation lens, there’s lots of pieces that might, at the highest level, seem similar. But then, when you scratch the surface, I think hopefully our users and our creators will start to see many of the things that make YouTube Shorts unique to YouTube. I’ll give you a couple of examples of those. And by the way, the Shorts product — I love it, obviously, I watch lots of Shorts every day — but it is by no means feature-complete. There’s a lot more work that we have to do, and I’ll be the first to say that upfront.
But there are a couple of features that I’ll call out that you can see already, that I think are leaning into that unique YouTube aspect. The first is the ability to take any audio sample from an existing video on YouTube and mix that into a Short. That is the glue between this new product, Shorts, and our existing, vast corpus of videos. That audio of course can be music, [but] it can [also] be audio from another iconic, canonical video that exists on YouTube. So that’s a unique-to-YouTube feature in terms of going from videos to Shorts.
And then, one of the other things that I think is cool is it can go in the other direction too, which is Shorts are oftentimes about music, right? We just did this big rollout with BTS and “Permission to Dance,” you might’ve seen over the last few days. Well, it should be really easy from that Short to be able to click on the song that’s sampled there, to get to the music video that exists on YouTube, where that video is getting hundreds of millions, oftentimes over a billion views.
The conductivity between those two things can go in either direction, and I think that is something you can expect to see more and more of. That’s what I would say in terms of context: mobile creation first, and then what are the unique-to-YouTube features we can continue to build for something like Shorts.
I think one of the underappreciated aspects of TikTok is that it is a very powerful video editor. That’s the thing it presents to the user; they roll out a bunch of AI filters. Those filters take on a life of their own, they’re easy to use and repurpose. You see [their names] on the top level of the videos. Are you heavily invested in that as well?
That’s an area where you can expect to continue to see us invest [and] where I also think we’re going to take our cues from our creators. One of the analogies that I like to use with my product team is our job vis-à-vis creators is to “set the stage.” Our job is to actually create the world’s best stage from an innovation standpoint, but it’s our creators who basically perform on that stage, and that’s where the magic comes from.
One of the things that we’re already starting to see, in terms of feedback from our creators, is they really like the product, but they would like to see XYZ. Some of that might be around effects. Some of that might be around things like filters, editing, all those types of capabilities that you described. Like I said, I think that we have the makings of a really fun and powerful product there, but we’re by no means done. So the answer in short is yes, we should be doing a lot more there.
Let me play with that metaphor for a minute. You say stage and players on the stage, but a much more boring way of breaking that up is that YouTube is a great distribution platform. Ninety-nine percent of the people who make YouTube videos make their videos in Adobe Premiere Pro, or Final Cut, or LumaFusion, or whatever, and they use YouTube for distribution and monetization.
TikTok has a distribution component, but it also has a powerful creation engine inside it. YouTube has historically not spent a lot of time on creation tools. Now you’re saying you’ve got to build up that creation skillset. Do you have that in-house? Are you ready to go? Is there a product roadmap there? Are you waiting to figure it out? What does that look like?
The first thing I would say is that it is true that the predominant use case — at least when it comes to videos that have wide distribution, lots of views on YouTube — are videos that were shot [and] edited elsewhere, and then uploaded to YouTube. And YouTube Studio is about managing those videos on YouTube as opposed to having a suite of editing tools. So that’s correct.
But I think I’ll go back to what we were talking about earlier, which is the genesis of Shorts and things like that. And if you zoom back out from just the Shorts product and the Shorts roadmap as it is, and if you think about one of the core product insights — I’ll keep coming back to it — which is creation through mobile phones and the power that these phones bring to that process.
And it’s not just power in terms of tools, in terms of what you described. It’s actually also the flip side of power: ease of use. Because you can do things so powerfully on the phones, oftentimes on the client itself, you are enabling a whole suite of tools that can be really easy for creators to use. Yes, that’s an area that we want to continue to invest in. It’s also an area where, as we grow, our creators that are producing predominantly on their mobile phones, we get feedback where, it’s awesome that you’re doing this, but here’s the next set of things that you could be doing.
So I really do look at it through the lens of, if a creator is getting started today, what are their expectations? How are they going to build their audience on YouTube? And part of that is enabling a set of tools on mobile phones.
TikTok is a merging of distribution and creation. Instagram, I think historically, is the first app that really merged a creative tool with distribution.
YouTube is a big app. There’s a lot going on in the YouTube app. You’re adding more to it. Are you going to add the creation tools to the YouTube app or are they going to be in a different app?
I don’t think I can say specifically today. What I will say though is, bringing it back to Shorts, a lot of the ways that you can enhance the video that you shot for your Shorts should live in that overall Shorts workflow. Now, does that mean that there shouldn’t be another set of capabilities for videos outside of Shorts? I can’t comment on that specifically today, but our goal is to continue to make the Shorts experience more feature-rich for video creators.
I see a lot of repurposed TikTok videos on Shorts. I see a lot of repurposed TikTok videos on Reels. I know repurposed TikTok videos on Instagram Reels bothers Instagram. Does it bother you?
I would tie this back to where we started the conversation. One of our core north star principles in terms of product design at YouTube is doing what really works for our users and our viewers. That is the lens through which we look at it. Obviously we have lots of signals and lots of data that we get from our users there. And our goal is to make Shorts content of whatever form easily and readily available to our users.
That’s why you see the Shorts shelf right on the home feed. It’s also why you see the tab that’s there, the permanence of that Shorts tab. So that’s how we look at it from a user standpoint. From a creative standpoint, I think what I’ll tell you is that those trends are changing on a regular basis. The amount of, first of all, new creators, but also creators that are coming back and producing original Shorts on our platform, that number is going up and to the right on a week-on-week basis.
And so that is another set of metrics that we look at, not just the viewer metrics, but also how easy, fun, powerful the toolset is for our creators. We’ve begun the journey there. We launched globally just a couple of weeks ago. So there’s still a lot more to be done, especially on the creator side, to make it so that it’s easier and more powerful and even more fun for them to produce original content for YouTube. But even with where our tools are today, that number continues to grow.
When I say, “Instagram is not happy about it,” what I mean specifically is they openly announced that videos with watermarks in them would get lower engagement than videos without watermarks. Which is a very roundabout way of saying, “Hey, when you export a video from TikTok, it has a watermark. And if you repurpose it to Instagram, we’re going to lower its engagement.”
Have you thought about similar moves?
The high-level answer I’ll give you there is that first of all, we make decisions and tweaks and updates to our recommendation and ranking algorithms all the time, multiple times a month, just to make it work better and better for our users. But we have a set of core north star principles and metrics by which we govern that. And they are driven by things like viewer satisfaction, engagement from our viewers, the feedback that we get directly from our viewers in those little surveys that you see in the app on a regular basis. That is the lens through which we will make decisions on things like ranking of Shorts, just like we do for quote unquote traditional videos on YouTube.
The other thing is, with Instagram particularly, they’re very explicit with their creators that using their additional products boosts their overall engagement. So when they had Stories, using Stories boosted your engagement. Now they have Reels; using Reels boosts your engagement.
You have a similar set of products. You have a similar, if not nearly overlapping base of creators on your platform who are used to that dynamic. Are you saying to them, “Use Shorts to boost your overall engagement, and if you don’t, your engagement will go down”? Because I hear that dynamic from creators all the time, that they feel like they have to expand into every surface of the app to keep that core engagement high.
We don’t talk about that in the explicit way that you’re describing at all.
I like to keep tying these back to what the core principles are — and I gave you examples of that on the user side. On the creator side, our goal is to give every creator a voice. If the creator wants to do that through a two-hour documentary about a particular topic they’re passionate about, then YouTube should be the place for that. If they want to do that through a 15-second Short that mixes in their favorite hit from their favorite music artist, they should be able to do that.
That’s why I work at YouTube; the diversity of creators that we support. I think it’s the core mission of YouTube. When we have conversations with creators, it’s about — what are the sets of products and features and policies that we should put in place that allows them to do that.
“We have lots of creators that are never going to try Shorts, and that’s fine.”
We have lots of creators that are never going to try Shorts, and that’s fine. We have lots of creators who have been video creators and vloggers on our platform for 10 years and are super, super excited about Shorts. And that’s awesome too. The way I look at it is [that] all of these products can be successful, and creators know what really works for them and their audiences. Our job is to address all of those lanes for our creators.
We’re talking here about Shorts, which I’m incredibly excited about. It’s great to see the growth of it. But we’ve also had other products and other modes that have existed on YouTube for a long time, and not all creators take advantage of all of them. For example, we’ve had live products on YouTube for many, many years. Not all creators are comfortable or want to do live content, but lots and lots of creators do.
I have a team that’s focused on live products for creators. They want to make those best in class for creators all over the world. It’s probably going to be a fraction of overall creators on YouTube that use those live products, but when they do, we want them to be the best.
I think it’s important to be explicit though, because a lot of the creators we’re talking about, the professionals — this is their living, their livelihoods. The pressure to win the algorithm to make more money, or to make more stuff to feed the algorithm, is very high, and the opacity of the algorithm drives people crazy.
I always joke that the life cycle of every YouTuber hits a point where they all make a video about how they’re mad at YouTube. It’s the end of one road and then that road branches off. And I think the heart of that is the opacity of how the product works. How do you balance needing to change, to be flexible, with needing to be explicit and transparent about how your product affects people’s money?
Yeah, I think it’s a really, really interesting question.
I would say a few things, and there’s a lot to your question because you talked about money, there’s a distribution component obviously to that as well. I think that our job is to be as clear as we possibly can about what works on YouTube [and] what doesn’t work on YouTube. We have entire teams — business-facing teams, creator-facing teams — that are focused on helping everything from creators getting started on our platform, to how do they grow their audience, [to] how do they think about monetization — not just ads, but a lot of the new monetization products that we have developed. So we do have teams that are working with our creators. We have ways of doing that, not just for our top creators, but we also have ways of doing that in a scaled fashion for our up-and-coming creators as well.
Those are programs that we’ve invested in. Could we do more there? Yeah, of course we could continue to do a lot more there. Every year I feel like we get a little bit better in terms of explaining all of this and laying this out for our creators so that they can make informed choices.
I also recognize that there’s lots of things that creators have to think about: distribution, monetization, our community guidelines, etc. So I do think it is my team’s responsibility, YouTube’s responsibility overall. Susan, our CEO, has talked about this too, of making it easier for our creators to navigate all of this. And so that’s a journey that we’re on. We need to continue to get better and better at it.
But I will also say that we have creators that have been on the platform for three, five, 10 years who continue to build a living, [who] continue to build an audience. They are some of the most inspiring, amazing people in the world. They are attuned to their audience. They adjust with their audience, they get feedback from their audience on a regular basis and they continue to produce incredible, awesome content.
One of the other areas though — how do we give creators the tools that they need to be able to take a break, something as simple as just literally being able to take a vacation? How do we give them the tools and the insight in terms of whether that actually has any impact in terms of how their videos show up in recommendations, etc? Those are always ongoing conversations with our creators. I personally get input from creators along those dimensions, and we have a team on my team that’s focused on building out those sets of tools for our creators.
We’re talking a lot about distribution and monetization, [but] one of the big areas of feedback that I get from our creators that factors into their overall experience with YouTube is how they interact with their fans, the comments below videos. That is an area that is oftentimes incredibly rich. You can see some of the most amazing conversations there around picking apart a physics video or something like that. But it’s also an area where I feel like sometimes our creators in the past have been getting overwhelmed, just with the sheer volume, or even sometimes the tone of the comments.
So we’ve built a whole slew of tools to help creators manage that both in an automated fashion, like moderated comments, a much better ranking of those types of comments, etc. That’s an example of how we’ve taken feedback from our creators and tried to turn it into tools that allow creators to have a much better experience managing their channels.
One of the theses I have on the show is that distribution platforms have a direct effect on the kind of media that is made. This is historically true: 12-inch singles made for longer disco songs.
I look at YouTube, and I look at all the things you’re describing, and all the pressures and the ad pressures, and I talk to YouTubers, and it comes up every time: the combination of rules around monetization and the [pressure] to win the recommendation algorithm means that YouTube titles are always hyperbolic, the thumbnails are over the top, and videos always hit 10 minutes so you can get one of the midroll [advertising] slots in. Those pressures, they just create that product.
I think I could abstractly describe that kind of YouTube video to anyone, and they could immediately think of a video they’ve seen that fits that exact mold. Do you think you need to ease the pressure on the product so it stops producing that kind of thing? Because I’m not sure everyone loves that kind of video.
The way that I think about it is, ultimately the way the YouTube system works, so the way our product works is a reflection of, again, some of these core north star metrics — how satisfied are our viewers? Are they getting what they want out of their connection with their creators?
Similarly, for creators, we have similar ways of looking at it. And so two things that I’ll call out that are, even up-leveling, the way that you described it, that I think about are, do we feel like the core milestones for our creators are correct, and are they healthy? So, for example, are there lots of creators that are getting started on YouTube? Is it dramatically harder to get started? Is it easier? How do we measure that? What’s the velocity of creators getting started? I can tell you that that’s something that not just my creator teams, but all of the YouTube teams look at carefully.
Creator feedback often focuses on building connectivity with an audience
So that’s sort of at the beginning of the journey, if you will. We also — to the second part of your question around monetization — one of the key milestones we look at is how many creators on our platform, new creators, are able to have a sustainable living on our platform? And we actually measure that, not just here in the US, but around the world, [adjusted] according to purchasing power parity and the like. Those are not perfect because they’re measuring snapshots in time. But again, just to kind of peel the onion back a little bit, those are things that I look at pretty closely, because I think they are fundamental ways of answering the questions that you’re describing.
Are we actually still creating a place where new people with new ideas can come, get started, find an audience? And once they find that audience, are we building products and capabilities that allow them to stop whatever else they were doing, and actually earn a living on YouTube if they want to become full-time YouTubers?
And those are metrics we look at literally on a weekly, monthly, quarterly basis to make sure that we’re doing the right thing by our creators. When I talk to creators, the feedback that I get is around, “What are the new types of tools and capabilities you’re building to help me continue to connect better with my audience?” Every single question or feature request I get ultimately boils back down to that in some flavor. And there might be a monetization objective behind it, there might be just a “growing an audience” objective behind it, but it really is about connectivity with the audience.
And so I gave you the genesis of Shorts, that’s one example. Everything I talked to you about with respect to comments and how creators can manage that, it’s all about connectivity to an audience, even a lot of the monetization products we’ve developed. There’s ads, but we have eight or nine other products that allow creators to monetize on our platform. And the fundamental thread to all of them is actually connecting with the audience. And so whether it’s Super Chat or Super Stickers or channel memberships. And so that’s always, for me, the fundamental question that I get when I speak with our creators.
We had Jack Conte from Patreon on the show. He said, “I started Patreon because I put music videos on YouTube and I was getting no return. You need an independent relationship with your audience that’s outside of the platforms.” I hear this argument a lot.
You’ve built a suite of products where the entire relationship a person has [with their audience], could be through YouTube. Everything they do comes through YouTube and one of its products.
You run a business. Do you think it’s smart for a creator to have one supplier, one platform vendor, or should creators be distributed?
I will say that I think it’s great that creators have multiple ways that they can generate revenue. If you talk to YouTube employees, I think many of them will tell you this, what’s really cool and inspiring about YouTube is that it creates this platform for YouTubers where — of course they can monetize on YouTube, and I gave you super detailed metrics in terms of how we actually measure the health of that for our creators, but lots of creators start on YouTube, but then they diversify in terms of their revenue streams.
I know you talk to YouTubers all the time: many of them have written books. Many of them have shows on other platforms or on television, or what have you. Many of them have built audiences elsewhere. All of that is awesome, from our perspective on YouTube.
What I also hear from creators, often, is that their core, most authentic, most deeply engaged, leaned-forward audience, is on YouTube. I hear that over and over, from our largest creators, to up-and-coming creators, to literally creators who’ve only been on the platform for a few months. That is the reason why we also want to give them the choice to be able to monetize beyond ads on YouTube itself. That doesn’t mean that they shouldn’t or can’t use other products or capabilities to earn a living on our platform, but what I hear is that seamless integration within YouTube with their core audience is what’s powerful.
I’ll connect it to what I said previously, which is the reason why these monetization tools have worked so well for our creators is because they’re not just about helping the creators make money, they also help them enhance their connection with their fans on YouTube. Look at a product like Super Chat or Super Stickers. Yes, it’s a way for a creator to generate revenue from a live chat that’s happening on a stream, but it’s also a way for them to connect with specific users, specific fans. They give them shout-outs, and they build a community that way. Those two things are reinforcing. I think that that is what’s really cool about all of these monetization products that we’ve built.
YouTube is also a search engine, right? You can show up on YouTube, you’re not part of anyone’s audience. You’re just bouncing between home repair videos. How do you think about that audience? Does it belong to the creators? Does it belong to you? Is it a mix?
I feel like that audience wouldn’t be there if it wasn’t for the connection they have with those creators. I have three kids. They have their favorite sets of creators on our platform. Back to my stage analogy, we want to build the world’s best stage, but it really is about the people in the audience connecting with the people that are on the stage. That is the magic of YouTube.
That is our creators’ fanbase, the audience that they’ve built. Our creators say that all the time. We enable it through our platform. Our platform allows them to get scale. One of the cool things that I like about a creator’s experience on a platform like YouTube is they can share an idea, and it can be received by somebody on the other side of the world in a way that’s incredibly powerful and inspiring. That happens on YouTube all the time. That’s really the way I think about it, and the tools that we’ve built are along those lines too.
You gave an example I thought was interesting, around people searching on YouTube, “how to fix my garage door,” or what have you. I would say, there, too, there’s a means to actually create a connection between me as a viewer and that particular creator.
I had to make an adjustment to the water heater in my house. There was an awesome video on YouTube that enabled me to do it. It probably saved me like two hours of time and probably several hundred dollars to do that. We just launched a product called Super Thanks, because I think it was certainly worth $10 from me to that creator, because that creator delivered enormous value for me. Am I going to watch that creator’s videos on heater repair every week? Probably not, but I still connected with that creator. That’s an example of another monetization product that’s about connectivity, but also allowing that creator to make money. That’s how I think about it.
The advertising products on traditional YouTube are very direct, right? You watch a video, there’s some breaks, the ads play. Some percentage of that revenue goes to the creator. With Shorts, there’s not a one-to-one. You watch Shorts, you scroll, an interstitial plays, and then this new Shorts program is going to deliver bonus payments somehow, from $100 to $10,000. How does that program actually work? How do you determine the value that creators get?
I’ll say a few things about it. First, it is really the first step. That’s why we’ve described it as a finite Shorts fund, I think it’s $100 million through the end of next year, because it is a means by which my team’s goal is to develop what we hope will be the long-term, scalable monetization program for Shorts.
I think you should put it in that context, which is a way to get going and to actually really start to figure out a way to describe how monetization can work for Shorts creators. Because you’re right. The experience is very different. You’re essentially consuming a feed of Shorts, and so the model has to work differently.
Payment for Shorts take into account engagement at the account level, not just an individual video
You obviously have the standard eligibility criteria, to be in good standing in terms of our community guidelines, [be] over 13 years old. We’re starting in 10 countries, although we want to expand beyond that very quickly. [You’re eligible for the Shorts payment] every single month based on viewership metrics, engagement metrics at the channel level, and if you obviously have created a Short in the last, I think, 180 days, is the cutoff we have right now. You will then qualify for a payment which, as you described, is between $100 and $10,000. That criteria could change, but it’s roughly on the order of viewership and engagement.
That’s one of the things we want to learn, is, how do we actually measure [engagement] well? How do we measure that in a way that can be transparent with our creators? I think a lot of that learning will come about through the Shorts fund. It might vary regionally because audiences are different in different regions. The addressable viewership is different. Those are all the types of things that we’re looking at, but that’s roughly how the program is designed to work.
I want to make sure we spend some time at the end here on trust and safety. As far as I can tell, it is unique that the trust and safety organization reports to you as the head of product. Why is that?
First, it’s hard to imagine a YouTube product experience without taking into account how the community guidelines impact our viewers and our creators. We spent the last hour talking about all of these products and features, and the decisions that go behind creating them. I would argue that a core part of that overall product experience for our entire ecosystem — viewers, creators, advertisers — are our community guidelines and how they work. That’s the first thing that I would say.
The second thing is a little bit more subtle, but I think in some ways is even more critical in terms of how YouTube works, which is [that] when we talk about content moderation or community guidelines, we tend to focus on [a] very, very small portion of what happens there, which is videos that stay up or come down on the platform. That is not by any means the be-all and end-all of how content manifests itself for our users.
What I mean by that is it’s not just about videos that come down or stay up, it’s also about how our recommendation systems work. It’s about how the ranking system works. It’s about how our search systems work. It’s also about how creators, 99.9 percent of whom are looking to do the right thing, are rewarded.
In terms of our responsibility to our ecosystem, to all of our viewers as a global platform, the way I think about it is this four Rs framework that we have — again, going back to first principles and frameworks. Only the first R is about removing content that’s violative of our community guidelines.
We have three other Rs: Raising up authoritative content. You’ve seen this in the context of the pandemic that we’re living through. We have information panels. We have a COVID news shelf that runs on your home feed every single day that has videos from authoritative news sources, from health authorities, etc. We raise up authoritative content when it comes to people looking for news or medical misinformation, critical information like that.
We also reduce content in our recommendations and ranking algorithms when they are not clearly community-guideline violative, but could be borderline for other reasons, harmful misinformation, etc, where oftentimes the lines are very blurry and not clear. What’s the line between political speech and misinformation? We might reduce the recommendations of that type of content in users’ feeds. That’s the third R.
We bring it all together with the fourth R, which we call reward, which is orienting the monetization resources towards those creators that are looking to do the right thing, build an audience, build a business on YouTube. I just would say that all four of those Rs are actually what govern the overall experience for our users. We tend to always focus the conversation on the first R, but don’t realize that those other three Rs oftentimes play an order or two orders of magnitude more impactful role in terms of number of videos on YouTube than just what we remove.
All of that is seamlessly integrated into how the product works. Everything I described to you is around our ranking algorithms, in terms of how the UX works. Where do these all show up? How do they show up? What’s the language of them? Can they click through? How do our monetization programs work? Those are all products that we built for our creators.
I think viewing them as two separate things is actually doing a disservice to the entire ecosystem. It is creating a situation where you can’t actually develop the right set of products and services you need to address this challenge in a holistic fashion. You’re just focused myopically on one sliver of it, that I would argue is in some ways maybe the most high-profile because it’s very clear to see, but in other ways, the least impactful in terms of what’s actually happening on the platform. It’s a bit of a long answer to your question, but I wanted to give you that full color.
Let’s focus on two of the Rs, raise and reduce. That’s the recommendation algorithm. That’s, “We’re going to show more people good videos, authoritative videos. We’re going to show fewer people the videos that are fuzzy.” How big is the team that runs the recommendation algorithm?
I can’t get into the specific numbers of the team, but there’s dozens of engineers, and product managers, and UX designers who are making tweaks and improvements to our algorithms on a regular basis.
By the way, one thing that I would just clarify is it’s not just about recommendations and ranking in our home feed, on search, etc. It’s also about products and features that we bring to bear, like information panels or search result panels. For example, if you search for “COVID information” or “coronavirus information” on YouTube, you will get a full panel that has information, including text, from health authorities, whether it’s the CDC, or, in other parts of the world, the World Health Organization, or whoever their national health authority is. It’s not just about how our recommendation algorithms work, it’s about how the entire product works.
In fact, the COVID panel, that you’ve probably seen underneath videos or in feeds in many places, I think has received on the order of 600 billion impressions since March of last year. We’re going to continue to run it, because we believe it’s a good source of information as users are making decisions with respect to the health of their family, whether they should get vaccinated or not. ... Raise and reduce are, yes, about recommendations, but they’re also about all of these other products that we give to our users.
I would put all this in the category of discovery. You’ve got a lot of discovery features, but I want to stay focused on recommendations. I’m assuming there’s a single main PM for recommendations, a product manager for recommendations. What are her [key performance indicators]? When you do OKRs at Google, what are her KPIs for the recommendation algorithm?
What I’ll say is our recommendations team is a fairly large team across YouTube, so it’s not just one product manager, there’s a number of them, but I’ll give you a few things. Maybe to give you even more insight, I’ll actually give you kind of the full picture.
You’ve been, obviously, very familiar with YouTube since the early days. You’ll remember that, in the very early days, one of the things that we looked at were things like views of a video as a signal of whether that video was interesting, or not, or what have you. That was obviously not sufficient, because it would lead to some of the things that you mentioned earlier, like, clickbaity thumbnails, or what have you.
Then we evolved to measuring not just the views, but the viewership of the videos. Are people actually watching the entire video? That factored into the overall KPIs of how that team thought about it. We’ve evolved since then, recognizing that it’s not just about that type of engagement, but it’s about long-term satisfaction of our viewers. That’s why you see some of those questions that you get directly in the feed that ask you how you felt about that video, that channel. That was an evolution. KPIs then evolve around satisfaction of our users.
Then we’ve layered another on top of that, which is related to this overall responsibility framework that I described. Which is, how do we measure whether we are doing a good job in our recommendation algorithms to raise up authoritative content or reduce some of this borderline content, or however you want to call it? We have KPIs around that.
One of the ones that combines a number of these Rs is one that we actually released a couple months ago now. It’s called violative view rate, VVR for short. I think the latest number that we had was, I think, somewhere on the order of like 0.16 percent to 0.18 percent. It fluctuates, of course. That is a core north star metric that this team would be OKRed on. What it measures is, in a sample-based way, the number of videos, weighted by viewership, or views, that were deemed to be violative of our community guidelines. Of course, the goal of our teams, the KPI, is to drive that down as near as possible to zero.
That is a challenging task. We’ve driven that number down dramatically over the course of the last year and a half, two years since we’ve been measuring that in earnest. But that’s an example of something that that team would get OKRed on and treat as one of their core north star metrics for the year.
One of the big criticisms of recommendations — I’m sure you have an answer for it — is the radicalization funnel. We hear about it over and over again with YouTube. We’ve heard there have been big internal debates inside of YouTube about how the recommendation algorithm fueled radicalization.
Is that something you’re measuring? Is that something you have a handle on, and you can turn it off and show people more viewpoints, or keep them out of the radicalization funnel?
As you can imagine, there’s a lot to that question. I think what I will say is that we have looked at this internally. There have been external studies on this. I think there was one a few months ago that was a joint study at the University of Pennsylvania and Harvard that didn’t find any of this. They felt that YouTube viewership was really just a reflection of media consumption habits, just in the world in general. There’s been studies in both directions. It’s obviously a really hard thing to measure, because each individual user’s journey is different. In an aggregate level you could look at things, but that’s not to say that I’m dismissing any individual viewer’s own experience.
I’ll just point back to the fact that the best way I know how to address all of this in a holistic sense is to really focus on those two pillars of raise and reduce. That is a way that we can determine what our north star principles and metrics should be and really work towards those. Yes, I look at things like, “are we doing a good job of raising up authoritative content,” especially in those contexts that I described. It is a continuous journey. I will be the first to say that we are not perfect, and we can always get better.
We also do the same thing on the reduce side. I gave you a very detailed explanation of how that works. That’s what my team is focused on on a weekly basis, and we do measure ourselves on those in terms of whether we’re hitting those metrics. That is the best way I know how to make sure that our recommendations are doing what we think is the right thing by our viewers and our creators.
The other major YouTube content moderation system that is of particular interest to me is Content ID. When I talk to YouTubers, right at the top of the list [of frustrations] is the automated copyright system that exists in YouTube.
Some of it’s out of your hands, right? There’s the Digital Millennium Copyright Act, there’s notice and takedown. There’s a lot going on there. But it is the largest automated copyright enforcement system in the history of the world. How do you think it’s going?
Just from a product and technology standpoint, I do think that it is one of the most critical aspects of the YouTube experience for our creators, but also for our rights holders. In some sense, it’s what has enabled a lot of the incredible user experiences that happen on YouTube. Again, like I said, with our other product areas, all of these products are capabilities that we want to continue to make better year on year, month on month, I would say that when it comes to Content ID, we really do try to strike the balance as best as possible. You mentioned some of the constraints, of course, that this whole world exists under, but again, tying it back to principles, our core principle there is trying to balance the needs of rights holders with content creators on our platform. For example, if we’re having a product review discussion around something like Content ID, it is trying to balance that core set of principles.
Where do you think Content ID needs to improve right now?
I think that one area that we have been focused on is really around ease of use and transparency. You described DMCA. You described all of these constraints. That’s a really complicated world that exists out there. There’s lots and lots of lawyers that are very focused on it in terms of both protecting rights, creating fair-use use cases, etc. So I think one of the things that YouTube has done and I think we can continue to do is just offer more transparency on that in our tool set.
So how do we build those capabilities for our creators and rights holders in a seamless way? We’ve done that. I also think that accessibility of the tools, that’s an area that we’ve been very focused on. I think you’ve seen just over the last couple years how we’ve expanded access to those tools in a way that works for a much larger set of creators. So those are the types of things that I think you can see us continue to invest in going into next year and beyond.
I want to tie these two ideas together. Content ID is the largest automated content moderation system at scale. It has a bunch of constraints, but it runs. People can issue strikes. It is contentious. And so you probably know more about the pitfalls of automated content moderation than anybody, except for maybe the people on your team that actually run it.
Next to that, you have a recommendation algorithm, you have principles around it, you have community guidelines, you have AI content moderation. It’s all building towards another kind of automated moderation system.
What have you learned from Content ID that’s going to shape the future of your community guidelines moderation and what are you going to avoid?
I think that’s a really interesting lens to look at it.
I would say the common thread between those two pieces is what I believe technology is good at — what machine learning classifiers can do well — is around scale and speed. That sounds obvious, but that is where a lot of the focus here is. And in the realm of Content ID, I believe that it has served rights holders, as well as content creators well. There’s always more we can do, but in general, that has been a system that has enabled billions of dollars worth of economic value for both of those constituents over the course of many years on our platform.
I would say the difference between something like that and the broad realm of content moderation is [that] machine learning and algorithms can do that first part around scale and speed, but I have also learned that it is important to have a highly trained group of individuals all over the world to be able to make the actual decisions, the nuanced decisions around our content.
So what machine learning classifiers are really good at is identifying a set of videos, or comments that are candidates for potentially being violative of our community guidelines. But where they often fall short is in terms of the actual decisions about whether that candidate video is actually truly violative or not. My learning, over the course of the last three or four years being really focused on this — and again, this is one of the reasons why it’s such a core part of our product organization — is that those two things are fundamentally tied together. In order to achieve what we are looking to achieve and drive that violative view rate number down as low as possible, it really has to be a combination of machines because they bring us scale and speed, but also highly trained individuals that can take our enforcement guidelines and apply them to videos in a systematic way, thousands and thousands of times a day to achieve the best possible results.
Does that moderator workforce work for Google? Is that part of your team or is that contractors?
We have multiple components of this. But at the highest level we do have employees of YouTube that are part of that team. And then we also have an extended workforce in offices all over the world where we work with partners to really wrap that up in terms of scale and numbers. And they’re all incredibly highly trained, always going through training in terms of new policies and enforcement guidelines, also are provided resources in terms of helping address a lot of the challenges of their work because it is obviously very challenging work. And so it’s a combination of full-time YouTube employees, as well as our extended workforce.
One of the things I hear from people in the larger conversation around platform moderation is we spend all of our time looking at Twitter and all of our time yelling at Facebook. And YouTube slides along in the background. It’s a frustration I hear from academics and even other reporters that YouTube doesn’t seem to get the attention that Twitter — in particular Twitter — gets.
Are you still paying attention to the debates? Around [Section] 230? Around the guidelines for misinformation? Does that influence you or are you waiting for someone in Congress to wake up and haul you in front of a hearing?
I mean, everything I talked about with respect to our responsibility as a global platform is something that I pay attention to, [that] Susan [Wojcicki] pays attention to. It is literally my number one priority. It is the thing that I pay attention to more than anything else. And the reason I pay attention to it is less for the reasons you’re describing and more because I believe it is a fundamental part of our viewers and our creators and our partners’ experience on YouTube. It is core to how they experience the product. So it is my number one priority. It is the highest set of OKRs and goals that I have and my team has. And that’s the reason why we’re focused on it. It’s the reason why we have thousands of people working on it every single day.
I’m going to let you off easy with lightning round. You just have to tell me over and over again that you can’t talk about future features. Don’t worry. It’s going to be easy.
YouTube wasn’t in 4K on the Apple TV for the longest time and now it is. What changed? What was that fight with Apple like?
I don’t think I can get into the specifics of that. But I’ll just say that our goal is to try to bring all of these features to all of our surfaces as quickly as possible. And sometimes we’re just constrained by the reality of the various platforms. But the goal from a product standpoint is to try to bring it as quickly as we can to all of our surfaces.
Did you win that argument or did they?
I don’t think I can get into specifics of any of the ongoing conversations we have.
I’m going to note for the listener that Neal was grinning and you can just interpret that however you want.
Two, when can I watch YouTube at 1.5 speed on a television? Because I cannot do that on my Chromecast or my Apple TV.
It is a good question. So I can’t ... Again, I can’t give you specific timelines, but you should know that’s not the first time I’ve heard that request.
All right. Put it at the top of the pile.
YouTube viewership on TVs, smart TVs in particular, is skyrocketing. You also make YouTube TV. So now you have two apps with very different business models that are competitive. Do you think core YouTube will overtake YouTube TV?
Overtake in terms of viewership on living room devices? Is that...
Does YouTube on television just replace television?
That is not a lightning round question. [laughs] That is a very involved answer.
It’s a yes or a no — there’s a way to answer that that is a lightning round answer.
I don’t see that in any time in the near future. There’s two different use cases there as you know through your own experience. Linear broadcast or cable is still very different, primarily driven by things like news and sports, from the video-on-demand world of YouTube. Stay tuned for how we’re thinking about those things in the future, but today they’re very different experiences.
YouTube TV, another product that you are responsible for, recapitulates a cable bundle. The price of it is going back up towards the price of a standard cable bundle. What’s the future of that product? Does it remain a big linear TV bundle or are you going to disaggregate it in some way?
I shouldn’t talk about how we’re thinking about it in the future. What I would say is there’s the economic realities of how content pricing works and the like. A lot of what you’ve seen in YouTube TV in terms of the new channels that we’ve added have been driven by user demand. Users have asked for that content to be explicitly added to the bundle as it exists on YouTube TV. So that’s what you’ve seen there.
We are also very focused on making it easy for users to add add-ons, to give them the flexibility of what else they’re looking for in the package. And we’re also doing more around either content-vertical type bundles or even feature type bundles. You saw that with the 4K bundle on YouTube TV. So we’re trying to bring more choice to our users. And a lot of it is, frankly, just driven by the feedback that we hear from YouTube TV users. There’s not another any price hike or anything in the near term.
When can I watch YouTube TV picture-in-picture on my phone or my iPad?
That is not something that I think I have a timeline on. And there’s more to that than just some of the technology enablement, but I don’t think I have a timeline to give you on that today.
That’s a contract problem, not a technology problem?
I may not have all the details on that specifically, but there’s lots of pieces that go into that. But again, like I said, I know that’s something that our users have asked for. Are you looking for that around normal picture-in-picture use cases like sports?
That’s the core use case, yeah. Yeah. I’ve heard that before.
I run a team. There are things I want on the site and I can’t get them to the top of my pile. You’re the [product] boss of YouTube. What’s the feature you want that you can’t get to the top of the pile?
Oh, wow. That’s really interesting. In the realm of Shorts, for example, it’s not a feature-complete product. There’s lots of things that I would like to see and I am by no means a talented creator at all. But I play around with all of our products. You mentioned some of the ones that I would like to see, which are: better filters, better editing tools, etc., for that product. And of course, I’d like to see it sooner rather than later, but I also know our teams are working on it. So that’s probably the latest and greatest thing that I’ve called out. But I think that if you asked me that two weeks ago, I probably would have said a different set of features.
Well, we’ll have you back in two weeks and we’ll ask it again.
What’s next for YouTube? What should people be looking for from your team?
I think we touched on so much here. I think that I’ll just say at the highest level you should continue to see us invest both on the creator side and the viewer side around the concept of mobile creation, Shorts as one example of that. We talked a lot about that and you’ll continue to see that.
I think another big area of investment is continuing to develop streams for our monetization for our creators. I talked about Super Thanks — we literally just launched that a week or two ago. There’s more to be done there.
One area that we didn’t talk about that’s related to that, that I’m really excited about because it also helps connect viewers and creators is shopping and commerce. That is something that I think I’m particularly excited about. And then the thing that I know for sure on the overall roadmap, that will remain at the top of it going into next year, also is all the work that we do from a product standpoint, policy standpoint, around our responsibility efforts. So that’s one thing I know, even though we’re not in planning season for next year, that I think will be at the top of the list, will be that, because it always is.
Decoder with Nilay Patel /
A podcast from The Verge about big ideas and other problems.