On collaboration and creativity
Quentin Hardy: I was looking at a tape today of Steve Jobs in 1991...
Prabhakar Raghavan: Yeah, yeah.
Hardy: When he's selling the next machine and talking about the value of interpersonal computing.
Hardy: And what he's trying to sell there is the value of networking, the value of collaboration...
Raghavan: Yeah, yeah.
Hardy: The value of multimedia inside this, computing power.
Hardy: All those elements. Part of the difficulty is he's selling a $12,000 machine.
Hardy: And we're doing this for pennies.
Raghavan: If you follow the chain of antecedents, right? I would actually go back to Mark Weiser at Xerox PARC in the late '70s, right? And of course, that was a lab high in system. It was probably hundreds of thousands of dollars to — to make the basics of what you just said, come to life.
Raghavan: The operating cost of doing this in the cloud is negligible at this point, which brings us an interesting challenge. If I were to grant you infinite computing, can you make the user's life any better than it is today? Yes, you can load things a little faster, but can you fundamentally change the way they work? The ability to collaboratively stay in one place, I think, is a nice instance of that because it multiplies the velocity of the participants, right? By a quantum leap. We at Google constantly have to challenge ourselves with that question.
Hardy: We've talked about how only 5% of the time are people doing truly creative work...
Raghavan: Creative work, right.
Hardy: We're not talking about delivering 100% creative work because that's just not human, but, if you got it up to 10%, you'd double the amount of time and people in the organization.
Raghavan: Exactly, and I think we are on our way there.
On G Suite and Google Cloud
Hardy: I think, inside Google Cloud, there's a sense of cloud computing, and then there's G Suite. But really they're quite intimately interconnected, and computation and the productivity tools are part in parcel.
Raghavan: Totally. The best way to start thinking about it is the Cloud Platform and G Suite have slightly different but overlapping audiences. The G Suite has its audience, all the employees in the company, including, I'll remind you, many people who did not have desks and desktop productivity before, right? There are field workers, security guards, store associates — they're increasingly licensing our products because that's productivity for them. Now, there's a subset of these employees who are developers, right? Who develop tools and systems for others to use, including their own employees, but also their customers. Those run our GCP. So the audiences are slightly different, but they're not disjoint. And it's often the same decision maker who is pulling all this together. We can focus on, this is what GCP builds and this is what G Suite builds, but, if you think of what gets added on top in terms of applications, the opportunities multiplicate in many fold, right? And that's what really we should look for.
Hardy: The original mission of Google is to organize information...
Hardy: Part of the mission of G Suite is not just to organize, but allow the information to be shared in better ways.
Raghavan: Yes. When you're saying information, it's not just raw information that you and I alter, necessarily, but it's a composite of everything in the company, including transactions, with the ultimate goal of making the enterprise more efficient, right?
Hardy: So then we're really trying to create this overall environment in which information can be shared appropriately, with appropriate rules and security, with machines, and people, and software.
Raghavan: Right, and the thesis underlying this is when information moves at a velocity to wherever it can, and you can throw enough computing power to make sense of it all, we all win.
Hardy: Why is that?
Raghavan: Because, as individuals, and as organizations, we can spend more of our work cycles being creative, building new products and services, and that's good for everybody.
On security within G Suite and Google Cloud
Hardy: As we talk about sharing a lot of information across a lot of geographies...
Hardy: And to a lot of individuals, I think some people might wonder, "Am I building a less secure environment?"
Raghavan: People get so worried about putting their data in the cloud, right? For a thousand years, people have been putting their money in the cloud. It's called a bank, and it goes there. They believe it. They're secure. They've never seen the money. All they get is this piece of paper at the end of the month saying, "This is how much money you have."
Hardy: And, likewise, there's a really big safe.
Hardy: There's a guy with a gun.
Raghavan: That's all true, right.
Hardy: These are things you tend not to have under your bed.
Raghavan: Exactly, at Google, we have a thousand security professionals who are guarding our customers' data. A bunch of things will evolve over time. One is people's perceptions, and regulations will evolve to be more accepting of the data in the cloud. People will get that the cloud is actually at least as safe as their cellar room with a key that's only available to three people, or whatever it is. The logical inconsistency is when people believe their on-premise world is secure, when, in fact, their information assets are distributed over a thousand computers and 10,000 thumb drives, none of which they've seen in their lives, right? How could that possibly be more secure?
Raghavan on productivity
Raghavan: As we bring together the data of entire organizations, all of a sudden, we're not only responsible for making people productive as individuals. We get to make teams and organizations more productive in their entirety. All of us do a range of work, ranging from extremely mundane stuff, to extremely creative work. Now, perhaps 5% of a worker's day, a worker's life, is spent on this truly creative work. And, shockingly, perhaps, 95% of it is spent on relatively mundane mechanized work, like searching for information, scheduling meetings. People clear their inboxes and think they've done work, but really it's not creative work. Once machines in the cloud are privy to all of this information, they can automatically schedule meetings, because that was an action item from before. They can block time on your calendar because you have an all hands presentation coming tomorrow, and you didn't think to prepare for it. These are autonomous actions that, if you had the time and the bandwidth, you would be spending on it, but you shouldn't have to, right?
Hardy: Right, and, likewise, the machines aren't dictating what we should do. They're automating according to rules we set. And it's not like Google or some other cloud provider is dictating how you have to work...
Hardy: You're setting your standards.
Raghavan: Right, I think one of the delicate balances in this business — as we bring more automation and intelligence into these tools — is to protect the user's understanding that they remain in control. Today, about one in eight email responses is machine written. In other words, machines read your incoming email, suggest responses, and you're going to pick one of them, one in eight times on average, and you're off to the races.
Hardy: In some ways, this is a very dramatic shift, but people are experiencing it in their consumer devices already when they use autoreply, or there's a spam filter inside their email.
Raghavan: Yep, yep.
Hardy: What makes it challenging, or a break from the norm, to institute this sort of automated behavior into the workplace?
On existing automation
Raghavan: People, in my observation, are somehow more reserved in how they use tools in the workplace, than they are as consumers and private citizens. And that's an aspect of change management. A year ago, we rolled out this feature in Google Drive called Quick Access. What we did was use machine learning to predict what file you want, even before you ask for it. When you come to Google Drive today, right up on top are our top predictions of which file you want to open. We're finding that, already, 50% — and this number is climbing — 50% of files opened in Google Drive are from this top panel of our prediction. Now, that's engendering a behavior change except that the nicest thing about this is most users haven't even internalized that there's a new feature here. It's like, yeah, you know, I came in, and the file was there, so I just took it and ran.
Hardy: Well, I think that's actually an important insight about any kind of change management, which is it should reward and, in learning it, you should feel a sense of agency...
Hardy: And a sense of power and control. In some clouds systems, there is a temptation just to take previous client server technology...
Hardy: And bring it inside the cloud.
Hardy: And you will probably gain a little efficiency, and you will save some money. What's the part you'll lose?
Raghavan: I think that misses the fundamental promise of the cloud, which is, again — when you go away from silos to aggregated information — what your team is about and what the mission's about is so much more holistic...
Raghavan: That is can optimize your productivity at a level you never could with sort of a climb server siloed system that's privy to half your information. The other half is sitting in a thumb drive somewhere, or in somebody's desktop. One of the things we eliminated with Google Docs 10 years ago, was the notion of version one with me, and version two with you, and the twain don't quite meet, right? There's one source of truth, and we're all working on it, and while, yes, we may think of a document as a physical entity, it's really the place in the cloud. Everything's happening there, right? And the team is virtually collaborating over there. Whether it's a document, a spreadsheet, a slide, any of these artifacts.
Raghavan on the benefits of changing processes
Raghavan: The most important thing for decision makers here to understand, is that these tools are getting so powerful that they're inextricable from the behavior of the people. Right? In other words, there should be a commitment to change, a commitment to change processes and work styles in the workplace, before you can make full use of these tools. Otherwise, you're going to buy our Docs and just use it as you did an old style, you know, editor.
Hardy: Once people learn how to work in a collaborative environment at a larger and larger scale, it can engender a level of trust and a level of mission around what people are doing that hadn't been as easy to come by in the past.
Raghavan: I think when you pick a large work group, partition it into teams, have them go do their thing, come back, and talk to each other and, you know, there have been walls in between, and nobody knew what the other side was coming in with. There's a lower level of class think instead of a different world, the world we live in, where these 100 people, as one of us is creating the project plan and some of the details, the others are already jumping in and saying, "Hey, this I'm not sure I agree with," right? There's an openness that it leads to.
Raghavan: And I've been at the tail end of projects that have entailed 100, 200 people, and things come out with much better consensus and much more trust. And even if I pull in the leads from 12 different teams, they're all nodding in unison and saying the same things, right? Which is a great feeling.
Hardy: If there are defects or difficulties, and you spot them early on...
Hardy: You tend to have a much lower cost of improvement...
Raghavan: That's right.
Hardy: And making something good. And in a shared document, you are heading towards a definition and a common understanding much faster than you are...
Raghavan: That's right.
Hardy: If you have five versions of an email going around.
Raghavan: Which, yes, exactly right. And, you know, people who have studied classical disciplines, like industrial engineering and optimization, recognize this because there's
all these local minima, and for everybody to find themselves in the same global minimum is, is always a challenge, right? And then you have to shake the organization around until that happens. Here, you don't have to because it happens through natural osmosis.
Hardy: For those who don't know about it, describe the Jamboard.
Raghavan: Okay, the Jamboard, as you'll see it, is a physical, digital whiteboard that you wheel in and you write on and so on. What makes it unique and different is everything you're doing in the Jamboard, whether it's freeform handwriting, or bringing an artifact from the web, from Google Drive, etc., is apparently happening on the board, but, in reality, happening in the cloud. So, the board actually sits in the cloud. What that means...
Hardy: The screen, it's a screen. It's an expression of what is being logged in the cloud...
Raghavan: Exactly, right.
Hardy: With everyone contributing. One of my screens might be a phone in a taxi. I'd be getting the information.
Raghavan: Exactly, right. You are rushing to the taxi. You need to participate in this Jam session. You pull out your tablet, your phone, and you're participating, right? And you can collaborate with anybody anywhere in the world, and people can jump in with a video conference, with a document. All the things we have in the Suite now, get pulled together in a collaborative session up in the cloud, and that's a super powerful metaphor.
Hardy: Where is this device going?
Raghavan: People want to be able to do more on the run from their tablet, from, you know, somewhat surprisingly, from their desktops. The expectation that your work group and the organization has on responsiveness, has gone up dramatically. None of us sort of blunders down a path, and then somebody comes and says no, no, no, go somewhere else. There's real time feedback and micro corrections, so an entire team and organization stays focused on one object.
Hardy: It can be a very clear way to communicate and to share information, and then move on to the next thing.
Raghavan: I think that's a great point. Meetings, as an example, should be one click, literally. You end. We are not sitting around waiting for the chair code and whatever else. Get your job done with real time collaboration, and you're out at your 25 minute mark or whatever, so you can go on to the next thing.