Alphabet Presents at Citi 2017 Global Technology Conference (Transcript)

|
About: Alphabet, Inc. Cl C (GOOG)
by: SA Transcripts

Alphabet Inc. (NASDAQ:GOOG) Citi 2017 Global Technology Conference Call September 7, 2017 10:15 AM ET

Executives

Diane Greene – Senior Vice President-Google Inc.

Analysts

Mark May – Citigroup

Walter Pritchard – Citigroup

Mark May

Welcome, everyone, and thanks for attending Citi’s Global Technology Conference. I’m Mark May; I’m the Internet analyst at Citi. Today our pleasure to welcome Alphabet to the conference and Diane Greene. Also joining me on stage today will be my colleague, Walter Pritchard, who heads up our coverage of the software sector.

Again, my pleasure to welcome Diane Greene as our keynote today at the conference. Diane doesn’t really need any introduction. But as everyone knows, Diane founded VMware in 1999 and company was acquired in 2004. Alphabet, Google was fortunate enough to be able to hire Diane two years ago to help lead their effort in Cloud. And the company has really established itself as one of the leading Cloud providers in the world in a fairly short period of time.

We look forward to talking with Diane about some of the key accomplishments over the last couple of years, how key technologies like machine learning and analytics are really helping them to win business with key clients around the world and how her conversations with customers have evolved in what is still a fairly nascent and high-growth market.

So with that, Diane is going to kick us off with a few introductory slides. And then Walter and I will get into the Q&A session. We will have time at the end to take your questions, so please be prepared. Although, let’s stick to the Cloud, although should there be a lot of questions around search and advertising, we want to keep the conversation to the Cloud.

So with that, Diane, thanks again. I appreciate it.

Diane Greene

Great. Thank you. Hello, everybody. This mic is on. Yes.

So it’s great to be here. Cloud is just getting more exciting by the minute. I’ve been on the Alphabet board for seven years and then 20 months ago joined Google full-time to run the Cloud. And we’re still at the beginning of the Cloud. I mean the Cloud was originally a place for startups, a place for surplus capacity, sort of a cost savings thing. And now with every business becoming a data business, the Cloud, people are moving to the Cloud to be secure and they’re moving to the Cloud to gain competitive advantage.

And Google, for the last 20 years, 17 years, has been building out data centers all over the world, developing data analytics. You could argue we’ve embedded most of the major advances and modern big data processing, machine learning security and also developer environments for our some 30,000 engineers. We’ve done over 100,000 miles of fiber optic cable. We are adding a new data center region about once a month, just this – in the past few months, we opened London and Sydney. And next week, we’ll open Frankfurt and that will be followed by Sao Paulo.

We’ve got – we run approximately one-third of the world’s Internet traffic. And in over the last three years, we spent $29 billion in CapEx, continuing to build out, keeping up with the enormous growth we’re seeing.

I want to start with security because I strongly believe that is the most pervasive problem every company has. Data breach is so expensive. I think our health record is $400 a health record if you get breached. And so Google, with all the information we have and the need to keep our customers data private has had to be a pioneer, and we feel pretty good about where we are in security, not that we are vigilant with our 800-plus dedicated security engineers every day, but I would say all of Google’s some 30,000-plus engineers are kind of security engineers.

And we’ve built security into every layer of the system from proprietary purpose built chips like Titan that say that the hardware hasn’t been tampered with and you are running the binary you think you are to every layer of the stack on to the software, the text phishing attack and then in your mail and then in your mail and then says if someone tries – if you gave the password and someone tries to use it, says, are you sure this person should have your password or do you want to change your password.

And we’ve also developed perimeter-less security because if you just have a firewall protecting everything, if someone gets inside, you’re kind of in trouble. And so we can look at what you’re doing, where you’re coming from and what call you’re making so that many, many of our services, Google services are just out there on the open Internet because we can protect them much better than we could with a firewall. And you take something like a Chromebook with it. The Chrome OS has tiny attack surface updated in the Cloud constantly. You don’t get a ransomware attack and or something like Heartbleed a few years ago, Google had that patched before it was publicly announced.

We’ve also internally developed a very modern programming environment so it can be agile and separate the layers and engineers can move quickly. And so we developed – we have – we launched over – everything runs in containers in Google, we launched over 2 billion a day, Gmail, every service we have, Maps and so forth are sort of 8 billion-plus active user services. And we brought that to the Cloud. We open sourced it with its Kubernetes, and it’s now becoming the de facto way to run your services.

And you combine that with the containers where everything is taken care of for your software and then API management with Apigee and your developers can move very quickly, take advantage of the APIs out there, take advantage of the infrastructure, automating everything underneath. And so we’re pretty excited about what’s going on in the last few months, everybody announced support for Kubernetes very active.

Data analytics and machine learning, you take something like BigQuery, our data warehouse, when Spotify moved from one of our competitors, their 1.5 petabytes, I think, of – 1.5 billion files over, they got a 35x speed up using BigQuery or you take something like Spanner where we’ve combined our hardware expertise with atomic clocks in every data center to make it so you can do transactions, distribute it all over the world, so if someone is withdrawing on either side of the world, we won’t let an overdraft happen for example.

And then machine learning, Google has been doing machine learning, really Larry got it going right when he founded the company pretty much. And now we have built all these data models that we give – share with the world through APIs. And so there’s all the language APIs, there’s even jobs API, we just recently announced 30 more languages, for example, for the translate. We have natural language, video recognition, very popular.

And then, finally, Alan Eustace said if you are serious about software, you’re going to build the hardware. We have our own processor. We open sourced how we do our machine learning, TPU’s Google’s really pushing the world on openness. And the TPU is a processor that is for TensorFlow, Google’s open source which is the most active machine learning project on GitHub.

And then finally is how we support people in our Cloud for workplace collaboration and productivity. People are increasingly discovering the huge transformative effects it has on their companies to collaborate. I know a public company whose Board of Directors collaborates in Google – a large public company, collaborating Google Docs to ask questions and work with the CEO. And you can do things like search everybody’s documents and see who’s working on what.

And it’s proven to be an amazing suite and we’ve got 150,000 users of Verizon moving over there, Nielsen with 60,000 using Hangouts all the time, Colgate-Palmolive. And so really getting big enterprise customers moved over to gain that agility in their company, that collaboration. Pretty proud of what we’ve been massive effort to infuse it with machine learning, which is you see in things like Smart Reply and then Calendar.

So that is just a – I hope about a 5 minute quick overview of Google Cloud. I look forward to your questions. One other little fun thing I’ll just mention is if you took one of our big data center buildings and overlay the Stamford Stadium, you’d still see a little bit of the building. These things are massive. It’s a very exciting area to be in from the basic data center up to working with all these customers to share Google’s technology and see them take advantage of it. Thank you.

Question-and-Answer Session

Q - Mark May

Thanks, Diane. Google’s Cloud business kind of been moving faster than the Cloud market overall in the last couple of years. I wonder if you could think about your time so far at Google in stages and maybe talk about some of the key areas of focus that you’d look to build out maybe in – I don’t know if you want to think about the first year or the second year? And then what are the areas of focus for you going forward so we can get a sense of how the business is evolving over time?

Diane Greene

Sure. I think when I came in we were mostly focused on startups. And obviously, the enterprise opportunities, at least, the slide deck, I mean the enterprise opportunities what we are going after. And we’ve made huge strides. So we had to really get all the table stakes functionality and things like identity and access controls and networking configurations and all the compliance and regulatory. So we got that done. That’s – it’s all there. Any enterprise customer can be fully deployed on us now.

And then we have to build out the go-to-market, which – certainly it’s the direct sales force and an immense number of customer engineers, Google quality customer engineers, which our customers value. We set up professional services. We set up an office of the CTO, Customer Reliability Engineering advances, we set up a lot. And then it was partnering because we really – the most important thing is our ecosystem. So we have big global strategic partnerships where people run on our Cloud and we integrate technically and we go to market together, and then just go-to-market partners. So building out our ecosystem has been a huge priority.

So we got all that done. And then we started looking at verticals. And now and taking our machine learning and applying it to verticals, and then taking Kubernetes and now we can support hybrid and lift and shift. And we bought Apigee to facilitate the API management. And then, I would say a big milestone, we’ve got in the top – Google Drive got in the top right-hand quadrant of Gartner and Forrester named us as the leader in PaaS, Platform-as-a-Service. So clearly, when I came in basically every customer partner discussion was is Google really serious about the enterprise, and I haven’t heard that in months now.

Mark May

Just a follow-up to that, Diane. You’ve obviously run a large enterprise software company in the past. I think you probably know very well what it takes to have a presence from a go-to-market perspective to be able to compete. You have a couple of competitors in the space, obviously, Microsoft has tens of thousands of sales and marketing professionals just by virtue of their legacy in that market.

Diane Greene

They did when I was running VMware.

Mark May

Right, exactly right. And then Amazon’s been building that out. Do you feel like the go-to-market is kind of the rate limiting step in Google’s Cloud growth or are there other factors as you look going forward or I guess, I want to focus in particular on the go-to-market and how much of a limiter is that on where you want to be in the next several years?

Diane Greene

Yes, well it is definitely an opportunity-rich environment and we are working hard and it’s kind of all hands on deck because we don’t turn anybody away. We do – we work with them and it’s very exciting for us. But we are building the leverage, where we build out training and certification, so we can train our partners, we can certify our partners. And we start – we announced a very deep strategic partner with SAP, which is going to give us a lot of leverage, and both SAP and Google are pretty excited about that partnership. And also some of the advanced data trustee thing that we’ll be doing. And then we are – you’ll see more and more partnering announcements. We recently announced Nutanix for the hybrid support.

And so I – we don’t have as many feet on the street, we’re growing very rapidly. I recently brought in a very seasoned sales executive, run double-digit billions of revenue somewhere else. And so we’ll continue that build-out, but it’s really focused on our partners and making them successful.

Mark May

And just on – you mentioned SAP. So your – you have the G Suite, which is a set of applications. You’re market leader in PaaS, you’re very competitive on the IS side. Sort of moving up the stack, how partnering with SAP, having some third-party apps, how do you think about that long-term in terms of the mix of business that Google wants to have in cloud and the importance of the Apple and SaaS?

Diane Greene

Well, G Suite is very strategic to us because you can do things like between BigQuery and Sheets, and Drive is a great place to put your files, getting greater all the time with all the machine learning and then Gmail with over – well over 1 billion users, it’s a pretty commonplace. So G Suite is the communication mechanisms that go with the Cloud, but – and the storage – individualized storage, team storages. But the SaaS vendors we’re partnering, we just announced a huge win with Marketo and that was a big deal, that’s an example of the kind of partner we’re going to assess vendor that will run on Google Cloud and take advantage of what we have.

And so we’re – we want to work with every SaaS vendor, and we think that’s a huge advantage for them to be on our cloud. Every time one of them moves they go, this is fantastic and they talk about all the things that have been improved for them and all the advantages.

So with SAP, it’s up and down the stack. I mean all of SAP, HANA runs on Google Cloud and more and more of SAP is running on Google Cloud. We are integrating G Suite with them, we are working with them on machine learning. And then what people don’t know is where doing a data trustee relationship. We did announce it, but I’ll just reiterate it where they over time with it will be their data center, we are the supplier, so that when they’re running in Germany, running our data center, running any workload, it will be a German company who has the data and not a United States company, which is a key milestone for us.

Mark May

Interesting. Just on hybrid, you brought up Kubernetes, which I think maybe two years ago there were a lot of these standards out there, that one has definitely gelled as the leading container manager, container orchestrator. What I don’t see from Google is there’s products in the market that leverage Kubernetes from other vendors. There’s not a sort of hybrid Google offering.

What is the hybrid? Because it does seem like also in parallel to hybrid Cloud, you’ve got enterprises with large data centers, you have large data centers, everybody sees the advantage of what you built, but getting there and sort of mixing what they have with what you have is difficult. So what is the hybrid strategy with potentially Kubernetes as big part of that look like?

Diane Greene

Well, we want Kubernetes to run extremely well on-premise, to run extremely well in every Cloud. We are pretty certain that it’s going to run the best in our cloud.

Mark May

Right.

Diane Greene

I mean, we invented it and developed it. And we did announce some partnership around hybrid with Nutanix. There is a solution there. And we built solutions with our customers where they’re running Kubernetes on-prem and in our Cloud. And I would just say stay tuned, a lot more to come.

Mark May

Okay.

Diane Greene

And then, we are taking – we’re also doing some new development that take it to the next level and that will be open source, too.

Walter Pritchard

Maybe a question on machine learning, you talked about how important that is to winning business today. And any examples that you can provide about how that’s actually being used by enterprise customers today?

Diane Greene

Sure. Yes, machine learning is the one of the major disruptions going on in that once you get in site, I always say Google itself, which has been, which is the most carbon neutral data center – set of data centers in the world we’ve always prioritized that, so we optimized our data centers energy usage as much as humanly possible. And then we applied machine learning and then our cooling got a 40% advantage and humans are getting 2%, 3%, 40% advantage from the machine learning, that’s kind of amazing.

So we are working with basically every customer on machine learning. And we can even bring them into our advance system labs and work with them. The public customers I can talk about are like Airbus that did use the vision to build a model to take the clouds out of their images. Something they’ve never been able to do in 20 years or Ocado one of the biggest online retailer in the UK, they’re based in the UK. And they were able to use our machine learning to detect customer sentiments and know how to handle calls in their call center or FamilyMart over in Japan used – it’s a huge chain of neighborhood grocers where they stock at three times a day and they can – they are using our analytics and machine learning to figure out how to stock things to predict what to do. And there’s a lot more of that as you might imagine, I can’t talk about.

Mark May

Sure. Just on the earlier days of cloud still get some airplay, price reductions were a focus and everybody in the market was reducing price a lot. Google actually had some sort of a headline making price reductions a couple of years ago. Can you talk about what impact price reductions are having now on the market? And how does that drive your strategy around pricing in your – especially the infrastructure as a service kind of the more commodity type services that you are offering?

Diane Greene

So the question is about pricing and putting words in my mouth saying our services are commodity.

Mark May

Well, no, I’m saying we’ve seen price reductions at, I think we’d all agree that maybe the basic compute storage services are they’ve been around for the longest, they’re mature, and they’re the ones that have seen the greatest price reductions. I’m just wondering your view on maybe price elasticity of demand in the market and how that drives your strategy around pricing versus what some of your competitors may do.

Diane Greene

One thing is we’ve been a pioneer in the model for pricing. And so when customers look at the total cost, we end up being quite a bit less expensive partly because we charge by the minute, partly because we give automatic price reductions, the more you use, then we don’t require reserved instances, and partly because we have such great performance. I forget the name of the group, CloudHarmony or something like that, just published a study that showed us the number one performing cloud in our services. And so we do have a very low total cost, and it is becoming companies biggest cost item, so companies care.

And then, when you get into the big deals, there’s a lot of work because everybody’s sort of got some custom, there’s a lot of work that goes into our contract, that’s a large team of people for us. And so there’s a lot that goes on as you develop a proposal for a customer.

Walter Pritchard

I guess along those lines, given how transparent and how segmented your pricing is, how many of your customers are actually on the sort of rate card, if you will and versus more contractual pricing and how…

Diane Greene

Well, we certainly have a very big self-help online business, right? And maybe as those people get bigger and bigger, we might come back and discuss what they’re doing and how to optimize it. And then – but the top Fortune 5000, when they’re doing a full on Google Cloud, we get together and talk or a snap, formerly Snapchat, we get together and talk about where they’re going to run and how much are they going to know ahead of time where their loads are going to be and how spiky is it going to be and we work with our customers.

Walter Pritchard

So as an outsider, how do we translate when we see these kind of public rate card type price changes. How immediate does that flow through to customers that are on contractual terms? Do they see the immediate benefit most all of your customers from these recorded…

Diane Greene

Yes. I mean, we will certainly – yes, we honor our prices no matter what.

Walter Pritchard

Yes. Okay. Sure.

Mark May

We have some microphones I think in the audience we wanted to definitely give…

Diane Greene

Yes, I’d love your questions.

Mark May

…to have an opportunity. I see a hand. Yes, you got it, right there.

Unidentified Analyst

Thank you very much. So I guess Walter asked you about enterprise sales force, you said you’re partnering with Nutanix and someone else so I guess.

Diane Greene

SAP.

Unidentified Analyst

Yes, SAP. So I guess…

Diane Greene

We have more partners, we have sales force, books, we have more partners and more coming.

Unidentified Analyst

And I think everyone recognizes your – the amount of money you guys have spent and you guys are a leader in technology. But to Walter’s point, this point is do mean upgrading your enterprise sales force, let’s say two years down the road if Google Cloud is kind of still maintaining its current trajectory, Walter will probably – Walter and Mark will ask the same question, do you need more traditional enterprise sales force so we can…

Diane Greene

Well, I think the Cloud – so we are hiring incredibly aggressively. Alphabet is investing in Cloud. So it’s we are not sort of saying, oh, we’re done with our sales force and they are very Googley. We are building out a serious enterprise sales force. But I would argue that Cloud demands are different approach. It’s a very – if you think about what’s going on when a company moves to the Cloud, this is something they had in-house where they run it all and now they’re just building these assembling almost these applications and putting them in containers and we’re kind of doing everything else. That’s where it’s going. And so we need to partner with them.

And so you do need the salespeople to drive a contract and sort of project manage, if you will. But it’s – you also want a lot of technical people involved. That’s what the customers need. How did they take advantage of our data analytics and machine learning? How did they transform to be more of an agile container-based development? How did they migrate things? So – but we are building out serious enterprise people. I think it’s really the way I translate that is we’ve become extraordinarily customer-centric.

Everybody in Google Cloud is so focused on our customers, and it’s what we talk about every – and our partners every single day and share a lot of information about what’s going on with our customers and partners and that’s how we’re driving the business. We have our SLAs, and we track how we are doing very carefully. And I would say we’re getting some pretty high-quality people. And one advantage of – the many advantages to Google Cloud, one advantage that I see is we really don’t have any trouble hiring. People really like to come to Google, so that’s been very helpful.

Mark May

I think there’s a question there towards the middle.

Unidentified Analyst

Diane, you did a good job in describing a lot of the IP that has been developed inside of Google over the last five or ten years. And Kubernetes, for example, I think if that were a commercial software that would probably worth the entire enterprise value of VMware at its peak.

Diane Greene

Yes, it sounds like let’s say we have 80 company-sized products.

Unidentified Analyst

Yes. So can you just describe the decision to open source all that software and IP rather than monetize it internally? And what goes into that decision and what’s the kind of the longer term monetization opportunity for all that open source IP that’s created?

Diane Greene

You know, it’s somewhat of a principled approach where we feel if this is – if the benefit of this will be far greater if it runs everywhere, then it should be open sourced. And so you take our Kubernetes, it is – it’s – we are driven by creating customer value, we really are. And if Kubernetes is available to everybody to run anywhere, it really amplifies its value, and that’s why Kubernetes has one.

And we feel that we know how to do it better than anybody else and that’s our advantage. And so – and TensorFlow was the same way, we want to build the ecosystem, we wanted to get a huge community, but we can – we’ll continue accentuating the advantage to running it on our Cloud, but we want to make it available everywhere. So wherever there is something that is open sourceable and its advantage gets hugely amplified by being available everywhere, we’ll open source it.

And I think this game of trying to lock people into your Cloud, we’ll see. But it’s not a great customer decision to do that and as the appendages, I think of, Google Cloud become more and more apparent and we are so open, I am seeing customers find that pretty appealing. Yes. Go ahead, sorry.

Unidentified Analyst

Can you talk a little bit about the decision to build hardware for machine learning? And then kind of how do you see in the future used cases that will be better run on your TPUs versus GPUs?

Diane Greene

Well, first let me just be really clear, we have a lot of GPUs in Google Cloud, and we work very well with NVIDIA. And the TPU, so Google does an immense amount of machine learning internally as well as externally. And so for the big data machine learning, training and using the models, we saw an opportunity to build a custom chip that would give an order of magnitude performance advantage, which actually saves us a lot of money and also lets us do a better job on the machine learning because you could turn things around so quickly. And then it’s natural to put it in our cloud for our users.

So the really exciting thing about Google Cloud is we now have a vehicle to take Google’s really ongoing technical innovations and share them with the world with every company and every geography of every size. That’s pretty exciting. And so it’s a natural thing to share the TPUs and we’ll also take advantage of all the innovations coming out of NVIDIA as well.

Mark May

Maybe time for one more. Go ahead Harry.

Unidentified Analyst

Yes. Two quick questions. One, how would you compare you experience with VMware versus Google Cloud, where VMware you’re leading from the front and here it looks like you’re coming from behind? And then the second, do you think you need to make any acquisitions to accelerate your enterprise adoption?

Diane Greene

Yes. How do I compare being a head at VMware to being behind in number of customers, I wouldn’t say we’re behind in every category. And so I mean, there is actually a similarity, if you think about it, because when we came out with virtualization at VMware, nobody ran it, and it was the best way to run your system, but people didn’t see the value in it. And there were a lot of operating systems running native on every machine out there, so we were behind in a certain way. But we – once we got it out there and people realized the value, we were definitely ahead.

And I think in a – you could say there’s the same thing going on at Google where we have the technology that’s going to really automate things for people and give them a competitive advantage and on an ongoing basis, but we’re behind in terms of people saying, okay, I’m going to be here, although it’s – I’ve never seen – one thing that is different from VMware is the pace. I’ve never seen such a pace. And VMware was pretty rapid growth. I was there 11 years to $2 billion in revenue.

And – but – there is a similar kind of startup feel in Google Cloud, although we have immense resources and immense technology and very high-quality people. And now, we’ve got very replicable customer deployments that go across the board in every single vertical. So I’m pretty optimistic. But it’s – I guess, I sort of – I think competition isn’t a bad thing. And in a certain way, we had competition. It was against running in the old way. So well…

Mark May

I think the second question we had on acquisitions.

Diane Greene

Acquisitions. Yes, we are definitely acquiring a lot of them, been fairly small. The biggest acquisition we did was Apigee which has worked out really well because API management is a big, big deal in the modern enterprise and how companies work across with their customers and partners. And we are constantly on a lookout for any sized acquisition. We have – we’re constantly revisiting our strategy. And we will acquire if we can find a way to make help us accelerate things. Yes.

Mark May

I think have come to the end of our allotted time, but I’d like to thank Diane and the team at the Alphabet.

Walter Pritchard

It’s a real treat. Thank you, Diane. Thank you.