Prabhakar Raghavan Isn’t CEO of Google

05.19.2021 07:00 AM

Prabhakar Raghavan Isn’t CEO of Google—He Just Runs the Place

In his first interview since taking a top job, Raghavan gets into the future of search, misinformation, employee ferment, and robots making phone calls.

Photograph: Google


Though Prabhakar Raghavan recoils at hearing himself described as “the CEO of Google,” the 60-year-old engineer turned executive is as close to being that person as one could be. He runs search, ads, commerce, maps, payments, and Google Assistant, businesses that bring in the lion’s share of the company’s revenue. And he’s paid like a CEO—last year the company paid him $55 million in salary and stock.

In some ways, Raghavan would seem in the mold of his boss, Alphabet CEO Sundar Pichai. Both were born in southeastern India and attended the prestigious Indian Institute of Technology, which has become a sort of finishing school for future US tech executives. (Raghavan entered the US pipeline with a PhD from UC Berkeley.) But while Pichai focused on a management career—he has an MBA and once worked at McKinsey—Raghavan is known as a world-class computer scientist who has authored definitive texts in the field.

He was publishing on information retrieval well before Larry Page and Sergey Brin thought of tackling the problem. Indeed, while teaching at Stanford in the 1990s, he had spirited discussions with the pair of grad students about the wisdom of starting a company based on web search. Instead of joining them, Raghavan wound up heading the research division at Yahoo.

When he finally did land at Google, in 2012, he moved from research to management, a pivot that some of his admirers assumed might be ill-fated. “We thought he’d be eaten alive,” says one. But he thrived at the top, first as the person in charge of the G-Suite of applications and later as the top executive for ads and commerce. In June, he assumed his huge current role, running an organization of more than 20,000 Googlers doing the company’s most critical work. During that time, his division became a center of controversy. Pichai took heat in Congress for Google’s alleged favoritism toward its own products in search. And when Google forced out an AI ethics researcher of color—and then a second researcher—for what looked like flimsy reasons, the company suffered blows to its reputation, and some employees left. This compounded existing rancor over Google’s lax responses to sexual harassment and unequal payment. During this year of turmoil, Raghavan’s profile has been lower than a business that doesn’t use SEO.

In the lead-up to this year’s virtual I/O developer conference, Raghavan sat (at home, of course) for his first interview since taking the job almost a year ago. He wanted to talk about the announcements there, of course, including improvements to the search engine that allowed it to handle more complex queries. But he was also unflappable on questions regarding misinformation, Google’s diversity and workforce woes, and whether a robot should identify itself when asking a restaurant what its hours are during Covid. And he also gave a rare report of Larry Page’s current role at the company, or lack of it. The interview has been edited for length and clarity.

WIRED: Your background is research. But at Google, you've chosen this management track, and now you've got over 20,000 people reporting to you. Why did you pursue that course and take this job?

Prabhakar Raghavan: I’ve had this exact question from well-meaning mentors who have said, “You do this thing called ‘scientist’ well, why would you try to kill yourself? Stick to what you know.” OK, as a scientist, you put forth an idea, it's your opinion. Here, you have a responsibility to serve 3 billion people on the planet. Having an impact on the scale of billions, like Google, is such a rush, and with the brightest minds on the planet.

Especially with the job you have now. You're sailing a ship into a hurricane.

Yeah, it's a large ship. It's a very important ship. And, yes, you're sailing into the proverbial hurricane. But that doesn't in any way diminish the responsibility or the need for creativity and innovation. In my view the problem of search is not solved. I remember one of the very first things that Larry Page said to me many, many years ago: Search is not going to be a solved problem. I've been pondering that since then.

That was at Stanford before there was a Google. What did you say to him?

Back then I was in algorithm design. I was obsessed with how to make algorithms go faster, less cycles, less memory, all of that. And then in the mid-’90s, when web search became a thing, there was this epiphany. For the first time in my life, I encountered a problem where, even if you were to give me all of the compute cycles in the world, with an infinite budget, I wouldn’t have a better answer. That's when it hit me that the problem is no longer just computer science in its traditional sense. It's one of cognitively representing the search problem. Even now, with all the resources of Alphabet, if you remove one of the biggest constraints, such as competition, the problem doesn't become easier.

You might be forgiven for saying, “What do you mean? I use Google every day; it seems to work fine.” But I claim there is so much more to do. Take a very simple example. Let's say you are planning to go hiking on Mount Fuji in the fall. One of the things is attire. Do my hiking boots suffice? That is a need; that's not a query. Today, what you do is you transcribe it into hours, and potentially days, of interaction with Google and other information sources. And then finally you say, “OK, I don't need to order new hiking boots.” Wouldn't it be a lot better if you could express yourself in the way you need to express yourself, and let Google or whoever else figure this out and address the need behind your query? This is a question I've been on. It requires really, really difficult problems to be solved. Number one, we have to let you express yourself. In the beginning, you had to type two keywords. Then we let you ask us, what's the height of the Eiffel Tower? I want to be able to get to a point where you can take a picture of those hiking boots and ask, “Can these be used to hike Mount Fuji?” We have to make enough sense of the physical world and the online world to be able to answer a question like that with fidelity. So you ask, why would I do this? The pursuit of those challenges is going to long outlast me.

Well, it’s telling that you say you want to reach that peak regardless of competition. Your critics note that you have no competition. You're not looking behind your shoulder and saying, “Boy, what's Bing doing? Or DuckDuckGo?” Maybe you'd work even harder if there were a comparable rival.

If you ask the question—Is there competition for the two-keyword query?—there's people like Bing who are good. But I think the real competition is among the people who are going to reimagine search, in the way we just described—who will let the user be far more expressive and who will do a much better job than today. Today everybody is analyzing the world's information and making sense out of it. My goal is to make sure that we are unparalleled in the understanding of the world. And that we give users the ability to express themselves, and that we bring the two together. It's that duality. I don't think this is a matter of competition in a conventional sense. There are plenty of people with amazing ideas. Some of them are going to figure it out.

What’s Google’s role in tackling misinformation? In early Google, people were fine with just getting the most relevant links. Now there seems to be a push for giant services like Google or a social network like Facebook to stop circulating misinformation. Are you going to be implementing product decisions to satisfy those demands?

We respond to what users are looking for. But if you are looking for what might be misinformation, I think it's our role to be as clear to the user as possible about what it is they're looking at. And then trust you to figure it out. So one of the things we're doubling down on is a feature you may have just started to notice called About This Result. Which is to say, you ask this query, here's a result, and by tapping here, you can find out more about the source: how long we have been indexing this site; what others are saying about it. We're going to highlight it and push it out globally.

Will this be a continuing effort where we’ll see more products to deal with misinformation?

Yes, it's absolutely something we continue to do. But I want to be careful to say we are not in the business of what should or shouldn't circulate. We don't think it's our role to adjudicate information versus misinformation.

One thing people in Congress and the agencies are trying to figure out is whether Google search favors its own products or forces companies to advertise to get exposed. It’s now your problem to defend the company. How will you handle that?

We've actually been putting out metrics on how relatively few of our queries actually show ads. It's the most commercial ones, right?

Yeah, the ones that matter.

The ones that might apply what we call commercial intent. But you also asked about the allegations of preferencing. This comes from a bunch of premises, not all of which I agree with. So there's a premise that says, oh, many Google queries don't send a click out to the web. But this flies in the face of the fact that every single year since our birth, we've been sending more traffic to the web than the previous year. It's also the case that a lot of times the answer to your query is put up there on the page. For instance, if you ask, “What's today's weather?” I show you a module of the weather, I don't send you off somewhere else. These questions will linger. But yes, I will firmly rebut them.

There’s a hunger in DC to break up big tech companies, and in Google’s case some people are saying the Doubleclick merger never should have happened and it should be undone. Is it possible to remove Doubleclick from Google?

First of all, I will remind you that our third-party ad tech plays a critical role in keeping journalism alive.

You mean, like my publication, WIRED.

So I think it's a good cause.

But people are tired of data-based advertising. One effect of the spat between Apple and Facebook is the increased awareness of tracking. When I type something into the Google search engine, I know that I'm going to see ads related to that, unless I go incognito. And now there's an issue about whether I'm entirely private using incognito. Is your view changing on how people get tracked from the information that they use in their search queries?

I hear a lot of my best friends and family say exactly what you're saying—I typed the query in Google, I went to this website, and now the ads for the website followed me. Most commonly, that is because the website cookied you.

Well my ISP might be watching what I do.

That's a great point. Because many of those companies are also ad-tech players. My view, and this is now Google's view, is that the practice of direct third-party cookie tracking isn't what users want. So we have renounced third-party cookies. That said, we want to preserve some semblance of a nice advertising experience, not random ads showing up at you, which is why we believe in technology that aggregates segments of users in massive numbers like thousands. And so the advertisements that come at you are not necessarily from that one query or website.

I never thought I’d hear a person from Google describe thousands as a massive number.

You'd be surprised at how small some of these numbers are in the industry. So thousands, in comparison, is actually very big.

OK, so here's another fun subject: ethics and diversity. Google has suffered setbacks here, like in the firing of two AI ethics researchers, including one of color, and some employees have quit during this controversy. Do you regret any actions that Google took in those cases? And what are you trying to do to rebuild trust?

We have this famously open culture where people are expressive. I actually think it's vital to preserve that. There is no circumstance under which I would say let's toss that out and become a conventional company, because that is not what is going to lead us to innovate and serve humanity. So first and foremost, we do want our employees to be expressive and open and to debate. That's how I think we're going to get the best and brightest minds. And so no question, we want to continue that tradition. If that entails building more trust, yeah, we're committed to that. Now in the particular cases you're talking about, do I regret this or that action? I won’t comment because there's a lot of detail that is not public, and I cannot go into that.

But to the first point, as a longtime observer, I do see a shift in Google’s culture, as evidenced by the close of the traditional TGIF meetings. I just updated my 2011 book about Google and concluded that Google was a more conventional company, certainly more than it used to be.

It's hard to dispute that things have changed. But I still think that we hire the best and the brightest and unleash them. I hope we do better than most. That's my goal and that of my peers.

Part of your portfolio is commerce. Will Google do a cryptocurrency?

I'm not going to answer. It's more speculation than I'm comfortable with right now.

So you won't say no.

I won't say yes or no.

Another business you run is maps. What’s your vision there?

I want our maps to be built on by far the best model of the 3D world around us, that we interact with every single day. For example, we now bring AR into maps. That's an instance where the experience we provide to you is far richer than the plain old map that we had 15 years ago, which was essentially a paper map stuck on a screen. There are lots of people growing up who have never seen a paper map. What's the right experience for them? To still have that simple map, or should they just hold up the phone and say go there, that's your hotel? But the rightness of that information is critical. Is it crowded there now? Is it open for takeout? During Covid, we had to make literally millions of updates to maps just on opening hours. We've done that through a variety of techniques. Have you heard of Duplex?

I certainly have.

We use Duplex to call businesses.

So you have robots making those calls and getting information from people? I hope you disclose it’s a robot on the other end.

“This is the Google Duplex robot. We're curious whether you've changed your hours.” And there's a little bit of dialog. “What about on Saturdays?” We captured that, and we made millions of updates.

Have you been back in the office?

I have, but not on a regular basis. We have these open pods that we put out on our lawns. I've been to a couple of those. I think society will have to undergo a stepping back of gradual confidence building. That will go in fits and starts, especially as suddenly infections go up in certain parts of the world or new variants are found. But I think, over time, we will build back up.

Some people think that post-Covid the whole employee-pampering culture with dozens of cafés and on-campus massages and dry cleaning will never come back. Is that all over?

I wouldn't call it an employee-pampering culture. I think we want our campuses to be really vibrant centers where employees want to come back to. So I hope we don't give up.

How much do you interact with Larry Page now?

He does surface, typically at board meetings. Other than that, there is a very rare email where he gets involved and makes some very pointed remark on a product feature. When he says something, it's deeply insightful. And you're like, “Oh my God, what did he just say? I need to think hard about it!”

Tell me about your relationship with Sundar Pichai, your boss. You have similar backgrounds, having gone to the same institute in India. How do you work together?

I don't know if being from similar backgrounds necessarily helps. We have similarities, and we have our differences. I have more of a technical pedigree, he has more of a business pedigree than I do. And so I can sometimes get caught up in the details of the model. He's very deliberate, always staging things out from so many different angles. I try to be as thoughtful as he is.

Before he was the CEO of Alphabet, he was the CEO of Google. I know you’re not in charge of YouTube or Google Cloud, but you’re kind of the CEO of Google now. Right?

I wouldn't claim to be any such thing.

Well, you have search, ads, commerce, maps. That used to be all of Google.

I think the way I'd interpret that is Google has grown to be so much more than it used to be.

Way back in your career you used to work for IBM. Even though IBM was dominant in the ’80s, it had nowhere near the impact that the big tech companies have now. Did you ever think tech companies would have such valuations and such power over people?

You put it as power over people. I think much more than in terms of power. I honestly wake up every day thinking of the responsibility to people.

But as you know, there's a lot more skepticism now. When a company like Google comes up with a technological advance, people question it.

Some of that inquisitiveness, even scrutiny, is understandable. It is up to us to convey the magic and wonderment of what we do, so people don't get worried about things they don't understand. Increasing understanding is definitely to everybody's benefit. What we're doing is pretty freaking awesome. And it is upon us to help educate the public on how amazing it is.

More Great WIRED Stories