Alphabet Inc. (GOOGL) Earnings Call Transcript & Summary

And maybe I'll just pivot there as you've taken us right to the next key element, I think, of your strategy, as I remember it, is certainly around industry verticals, as you've talked about. And can you talk a bit about it more broadly? And then you mentioned retail and financial services specifically, I'd like to jump into financial services because it's most relevant to us. But certainly, can you talk about vertical strategies, how you chose them, to rank them and how you're building them out and how they compete?

Yes. Our technology applies to all industries, but we see different industries focused on different needs, and we're trying to deliver solutions for those industries. Our primary industries are public sector, financial services, retail, health care, manufacturing and communications and media. If you look at financial services, we break financial services into consumer and institutional banking, wealth management, capital markets and insurance. And there are different requirements in each of them. Just to give you an illustration, our goal is to use large-scale data processing to streamline business processes within an organization, to transform the customer interface and to allow people to build new products and sources of revenue. So when we look at, for example, insurance, we're working with a number of the leading insurance companies around the world around some of their important processes. One example is, how do you change underwriting? How do you streamline how claims are processed? And the third is, how do you put a digital agent in front of the consumer because they may not feel comfortable going to an insurance agent office still. And so these are examples of how we're taking our base technology and then applying it to different domains in different industries. And we are seeing, because of that, enterprise customers always want their problem solved. The more applied you can make it, the easier it is for them to adopt it, and we're very focused on that kind of an approach.

How big of a challenge to say when you're working with an enterprise client to say, yes, thanks for helping me re-engineer this process that saves time and money, but you're not allowed to sell it to the next competitor out there? Is that something that's more and more a part of agreements with enterprise clients or your core operating premise is, the cloud is going to be available to all industry participants and our IP will be separated out and kind of repackaged?

The vast majority of clients recognize that if we maintain the IP and are able to enhance it, it will move faster than it was for just for them. So in the vast majority of cases, people say, as long as you work and give me an advantage by starting with me, and I'm part of your early adopter programs and other things. They recognize the value that comes from having shared innovation with other people, and they find the benefits of that. So we've been generally pleased with how we're able to solve our major customer problems while at the same time maintaining an overall curve of innovation that helps all the players in that industry.

And if I jump to financial services, and obviously, a heavily regulated industry. How are you thinking about dealing with regulators as you move more and more banks or insurance companies onto Google Cloud services, the regulator will look at greater and greater systemic risk, which is their job, with those operations. How do you manage this growth in this ecosystem with the regulatory systemic risk?

It's a great question. We completed regulatory audits in 39 countries. And so we're very familiar with them because we work with many leading financial institutions. Our approach boils down to 3 important elements. One is evolving a platform to address some of the requirements that regulators have. A lot of the work that we've done on this notion of multi-cloud, being able to use Google's technology on other cloud platforms and in a data center of a customer, has provided a significant benefit in financial services around this notion of concentration risk, which regulators are very concerned about. The second is, we've put in place operational practices and policies. For example, many European financial institutions want their data to be maintained in Europe. Many financial institutions have certain access requirements to who has access to their data. And we put in place lots of operational practices to govern the use of data, particularly financial services data. And then lastly, we go through regular audits with regulators, both financial regulators in a country or in the case of Europe, also with the European Central Bank. And we have people in our policy organization that work super closely with them to understand where they want regulation to go and then evolving our technology to be able to support those regulatory requirements. In general, as cloud becomes a more and more important part of the IT landscape in industries and in different countries, we expect to get more requirements of this kind, and we are very committed to meeting that.

So you've obviously thought through this. That's really good because I can see the questions coming in from our regulator as we start to move these services over. So you started to pivot then again to another key component of, as I understand it, the Google strategy around multi-cloud. And you're going to rebrand it and relaunch Anthos, which I believe is the Latin derivation of the flower. And it was the multi-cloud kind of management capability. Can you talk a little bit about your multi-cloud strategy and how that differentiates you from maybe some of the uni-cloud competitors as they would maybe self-describe out there?

I use this example. Imagine if you were back in 1997 and the Internet was happening, and you said certain web pages were only available in Microsoft's browser and other web pages were only available in Netscape's browser and other web pages were only available in Firefox's browser. The Internet would not have succeeded because it would have been fragmented. So our approach has been, if you want to enable organizations, the primary barrier to adoption of cloud in organizations is skill set. And a big part of it, I don't think organizations say, I want to have 3 different cloud teams, one for Google, one for Amazon, one for Microsoft. Why don't I have one team and learn and give them the choice of where they want to deploy. So our approach with multi-cloud is, there should be no complexity in providing customers the ability to use the cloud provider of their choice, but to have them do it without being locked into that provider. And the way to do it is to build an open programming model for how people can build, deploy, manage and secure workloads. And that's a lot of the work we've done. We've then taken our data and analytics technology that many, many people find very differentiated, our machine learning tools, and made them available on other platforms. And the reason is people said, I've chosen someone else to put my data in, can you bring your data processing to my data as opposed to the other way around? And we feel that if you do right by customers and give them the choice, it will eventually get you more success. As the regulatory landscape has evolved and the notion of sovereignty comes up, the notion of, hey, I want to have a regional presence. Even, for example, in the public sector, depending on changes in the classification of certain kinds of information, people have to use different clouds. The use of multi-cloud gives them choice and flexibility. And we've seen really strong interest from customers to this approach we've taken.

Can you talk a little bit more about some of the partners you've chosen to execute a multi-cloud, containerization partners, containerization strategy, because obviously, I think about those things here as we choose our cloud partners.

We have a broad set of partners. They fall in 4 buckets: the tools and developer tools that developers use to both deploy, but also package applications; second, tools that allow them to run workloads; third, tools to migrate data is a common requirement that people have when they want to move data around; and lastly, tools around security, how do you secure this environment? And we have partners in a number of these. An example, if you look at the security space, there's common requirements on how do I lock down my container, how do I provide container security? We have a very good partnership with Palo Alto Networks in that area. So there's quite a few things that we've been always clear on what do we want to enable, what tools do we provide of our own and what's the partner ecosystem that we want to bring to solve these problems.

And I'm going beyond my technical ability, but can a uni-cloud competitor kind of block that strategy or do you have the ability to run this multi-cloud environment independently?

We can run the multi-cloud environment independently. It abstracts out the underlying cloud provider's ability to block it.

Right. Right. I'm sure you've thought of that. So that's great because I think as you talk about regulatory, corporate strategy, small business strategy, the need to manage short and long-term risks, right, risks of porting and not being beholden to one development capability is something that we certainly think about and managing operational risk at the same time, so I think that it's a fantastic strategy. You touched on another theme that I wanted to discuss. And that's cybersecurity, something certainly that's omnipresent, but heightened during this pandemic. As we see this movement towards public cloud services and more processing capability, more data residing as we see a world with more endpoints in a world of 5G, which we'll touch on after. Can you talk about your security strategy and how you're protecting data, protecting partners and customers in this omnipresent cyber world?

Our philosophy around security, it boils down to some very basic tenets: number one, make sure that security is easy to operationalize; second, provide the tools to detect threats and risks; third, make it easy to automatically remediate some of these problems because historically, that's also been challenging; and lastly, allow people to choose how to configure their systems to minimize all the way from the endpoint to the cloud. And so I'll give you just a couple of examples. One of the common things we felt was, people would really like to encrypt data. And if it's encrypted, it's much more difficult for malicious people to get access to it. Historically, there's been 2 difficult things. One is making encryption possible was hard. You have to manage keys. There is a lot of complexity with that. The second is data was encrypted only at rest or in transit. While it's being processed, it was left wide open, so somebody could always open the front door and get access to it. So we came up with a solution to encrypt data all the time, at rest, in motion and while it's being processed. And we also made it super simple for people to do so it was not complicated for them to manage. A second example of something we felt very strongly about was making it easy to detect if you're in compliance with security practices. So we've taken security and convert it through configuration so that an operator can just look at your system and say, is it secure or not? Historically, you had to use lots of tools. It was super difficult for a CISO to get a sense, am I secure or not? And so we made it much simpler through our threat detection capability to identify you're running in a secure environment. For instance, if you're a health care system and loading health care data, are you maintaining HIPAA compliance because that today is very difficult to do. And lastly, we're working on some advanced solutions around cyber protection even for external attacks. As you can imagine, we have lots of technology at Google that we're applying to this domain because we think that is a significant area of further work that needs to be done as part of the industry itself.

It's really impressive that you've built out a capability not only to -- that you expect to protect data at rest, as you say, and in motion, but also while being processed within the machine, and I think that's really, really important, so you've obviously thought that through. When you talk about kind of 5G and how that's going to change where we're processing data, where we're making decisions, how does that impact kind of the cloud-based and how you think about security, how you think about capabilities in the world of 5G?

5G, given the investments telecommunications companies are making in 5G in the network itself, 5G networks provide not just lower latency, but also more bandwidth. And so given that it's lower latency, what we're doing is integrating what we call our mobile edge computing solution with many different carriers around the world into the network. So in the carrier network, you can actually have computation that runs in the network. And it allows applications, new kinds of applications that have very low latency to be available right in the network. So we see a lot of different solutions, and what we're doing is providing not just a technology platform with mobile edge, we're also providing a catalog of solutions for specific industries with a partner ecosystem to make it easy for customers to adopt it. I'll give you just a practical example. We have a large manufacturing company that runs a manufacturing shop floor. The robots that run the shop floor have to have a clean room environment. The control systems and the computers that talk to these robots, therefore, need to be close to the robots. But if they're in the factory floor, one of the challenges you have is technicians have to go in to repair the disks and replace the disk and other things. So they wanted us to put that into the network itself. So working with the telecommunications partner, we built that into the network. It's very close to the factory, so you get low latency, but you don't have the disruption of people walking in and repairing things in the factory floor and allows them to get the combination of better value to the carrier because the network is becoming a locus for application delivery and a better solution for the customer.

Right. Well, that's a great example, an exciting usage model. Can I jump a little bit to artificial intelligence? I know we're running out of time here, and I got a couple of topics I might try to squeeze in, in the last few minutes. AI has so much promise, so much capability. Can you talk about both some of the exciting applications you've seen in AI? Google has invested heavily outside of cloud and AI capability, reinforcement learning. Obviously, you'd think that some of that capability is going to be brought into the cloud. And what are some of the barriers you see to really leveraging this incredible technology? We talked about data earlier. And then the third question, I'll probably remind you the third question, a little bit about ethical AI and how we have to think as a society about using these tools. So a lot in there, but we'll definitely get through them.

We see AI being applied to support and augment humans in a variety of different industries and fields. As an example, we've done work using our image recognition to let radiologists have better tools to diagnose tumors. Because our image processing is so accurate, it can identify tumor cells. We've helped many manufacturing companies have image processing on top of their shop floor so that we can do quality inspection without human beings standing next to one another. It reduces the risk to the humans having to be on the shop floor, it also speeds up the manufacturing line. There are many, many such examples of places where artificial intelligence, whether that's image processing, understanding data, helping people understand conversations are all going to change the way that organizations function and use the value of the data in different ways. Now at the same time, we've also been very careful that the same technology that can be very powerful in one context can also potentially have negative consequences. So as an example, we do work on object detection. Object detection is, for instance, if you got a human being who is in distress in the water, for example, can you detect that so you can help the Coast Guard rescue them more easily? That's an example of a project we're doing in one part of the world. At the same time, the very same technology can be used to identify whether that person is a specific individual. And that technology of human identification has a number of potentially negative consequences. For instance, it could lead to diversity and inclusion bias. So we've been very careful on putting forward an ethical AI framework because we feel that when you have a very powerful technology, you have to have some principles that govern the use of that technology. And Google adopted something called ethical AI. We've put forward a set of principles that are meant to govern how we use AI in our products, how we expect people who are using our AI to use it, but also meant to be a foundation for discussion with other technology companies and policymakers on what's the right way that AI can be used in the world.

That's a great example. It's fascinating. Both sides of that equation, trying to help, but also trying to protect confidentiality and the ethical side. I'd like to just ask one more question. I know we have just a couple of minutes. One of the most important issues that we've all dealt within our society and as leaders is building a fairer, more inclusive society going forward. You touched on it just now. And can you just talk a little bit about kind of your thoughts on building a fairer, inclusive society and how you're leading within Google through this?

Our general view is technology is still something that does not distinguish which person and what race and what nationality has unique advantage to create the next technology. And so our view is, we need to be representative and provide opportunities for people of all backgrounds to be able to create new technology and innovate at Google. For us, it boils down to 4 things. Are we recruiting a diverse and inclusive population? Are we helping develop people of diverse backgrounds? Are we promoting and rewarding them? And are we retaining them? And we have very specific work going on in every one of those areas, working with our people operations team, our employees, and it's a matter of great importance to us because we want to be a place that builds representative technology as well as has a very representative organization. Some of the things we've had to change, for example, people often think about, and we have to educate our managers, that you don't hire for cultural fit. You hire to evolve the culture of the organization. And so don't just hire people who make you feel comfortable, but hire something that makes our organization stronger because it's more representative of the broader community in which we live and work. And so there's a lot of work that we're doing. We constantly are paying attention to it to see if there are ways that we can improve further as an organization. And we're learning also by discussing with many universities, with different organizations, and this is a matter of great importance to us.

Well said, I think that's the very definition of diversity, right, in different perspectives and not hiring our likeness and challenging ourselves to think differently. It's such an important point. Thomas, we could go on for another hour. I know we've taken a lot of your time, and we still appreciate the time you've spent with us today. And I think we've covered a lot of topics, but I hope everyone can feel the power of the technology you're building. How you're going about it is different, from multi-cloud and the verticals and the partners you've built will differentiate your capability. The future is going to transform small and large enterprises, bringing the power of AI, bringing the power of processing data and re-engineering processes within verticals to put into your capability is transformational technology. And we are so excited that you were able to talk to the conference today about some of the things you're building and have a good feel of the future that's at your fingertips and that you're building. So on behalf of everyone, a sincere thank you for a great conversation today.

Thank you so much for having me, Dave. It's a pleasure to join you.

Be well, and I'm sure we'll see you soon. On behalf of the entire audience, a sincere thank you.

Thank you very much.