There’s an odd phenomenon that happens to me at Google Next conferences, Google’s annual event to tout its cloud computing platform. It happened to me at last year’s show, and it happened again this year.
I go into the event full of skepticism about Google’s hopes for its cloud computing platform. The search giant is clearly a laggard in cloud computing. Its toughest critics suggest it may never be more than a supporting player. And yet after I spend time at its annual conference, watch demos and listen to Google execs, I become convinced: Google Cloud will be a force to be reckoned with.
I look back at my Google Next piece from last year and I’m almost sheepish – I was so enthusiastic about GCP’s chances, a distinctly minority opinion at the time. (In my defense, my Google Next article from 2015 was restrained.) Mind you, I’ve seen a lot of tech presentations; I’m properly jaded. But, even with a year to regain my doubts – well-founded doubts, to be sure – this year's Google Cloud conference again won me over.
First, the skepticism. Cloud leader Amazon Web Services began offering IT infrastructure services in 2006, essentially pushing the IT industry into the new world of cloud. Microsoft, slow at first given its legacy of packaged software, saw the critical importance of cloud and invested deeply in Azure. Google, meanwhile, appeared to be largely sleepwalking. Google Cloud Platform has only been seriously competitive in the enterprise market for the last few years.
This Kubernetes display seemed popular among Next attendees
Estimates of 2016 cloud revenue from Deutsche Bank sum up the horse race: AWS earned $12.22 billion, with 2.42 billion for Microsoft Azure and $900 million for Google. (Figures don’t include sales of cloud-based productivity tools, which for Microsoft is huge.) Deutsche Bank forecasts that both Google and Microsoft will double their cloud sales over the next two years, with Microsoft growing at a faster rate that Google. In other words, GCP’s underdog status won’t change in the foreseeable future.
The revenue numbers not only reflect Google’s late start, but prompt a still more challenging question. What exactly is it that gives Google Cloud an advantage? Amazon has the all-important first mover advantage, and now releases new whiz-bang cloud tools at a dizzying pace. Microsoft Azure benefits from Microsoft’s deep legacy relationship with businesses based on the Windows environment.
But Google? Hmmm, well, it has a depth of technical expertise that is unsurpassed by any company on the planet. But AWS has a clever habit of finding what’s commercial in high tech and offering it to businesses in a way that’s geared for their current needs. Microsoft is no slouch in this area either, and has an edge over AWS in hybrid cloud, a model adored by large enterprise.
Underneath the marketing noise, Google Cloud’s positives going forward boil down to two key areas: Google’s profound strength in artificial intelligence (and the related fields of data analytics and machine learning), and its newfound deep desire to romance enterprise clients.
AI and Analytics: Google’s commitment to artificial intelligence and machine learning will be key competitive advantage; incorporating AI will not be optional for companies in the years ahead. The year 2017 feels like a watershed in this regard; AI has reached omnipresence, defining our social media feeds, driving our cars, playing a key role in retail and medicine. Google has a lot to draw on is in this area. It develops DeepMind, the AI monolith that bested a human player in Go. And Google’s well-regarded TensorFlow is an open source library of machine learning tools.
In her keynote speech, Fei-Fei Lei, Google Cloud’s Chief Scientist of AI/ML, talked about “the democratization of AI,” claiming that Google will be a leader in enabling businesses and organizations at all levels to use this powerful technology. Google will likely succeed at this, given its breadth of AI research and its tendency to lower prices for cloud customers. As an example, Google’s Vision API is now generally available. The demo of this AI tool showed how a computer can “watch” a video – a significant achievement, the cutting edge of advanced image recognition. Google’s menu of sophisticated AI and ML tools will interest startups and more mature companies as well. And Google’s strength in data analytics is massive; the company, after all, develops an algorithm that catalogs the entire Internet.
Promoting GCP technology
Newfound Enterprise Focus: There’s an opinion that often voiced about Google: its internal company culture is geared for exceptionally high end, sophisticated tech. This is the company that creates self-driving cars and tweaks neural networks for deep AI projects. But catering to the boring needs of regular mainstream businesses? Not so much.
“Google makes phenomenal technology, but their strength and their weakness is that everything is very ‘Googly,’” said Stu Miniman, an analyst with Wikibon who attended Google Next. Translated: it’s geared for the hyper geek. Google’s attitude is perceived, he explained, as “we make the best stuff, and the smartest people know how to use it.”
Another longtime tech pundit I spoke to at the event, who asked that he not be named, put it thusly: “With Google, they can talk to you a long time and you won’t really know what they said. They’re smart people, they’re smarter than you.”
But there were signs at the conference that this is changing. Google SVP Diane Greene, hired in November 2015 to remake Google Cloud, is GCP's greatest asset. Having co-founded and built VMware – which revolutionized enterprise IT – Greene has the enterprise DNA that Google so lacks. If anyone can turn the GCP into a leader, it’s her. (Incidentally, in her keynote, she said, “I look forward to the day when this audience is 50 percent women.”)
Onstage, Greene presented various large enterprise GCP clients. Most notably, to my eyes, Colgate-Palmolive. While it’ a highly successful company, no one has ever described Colgate as a hot bed for futuristic innovation. Greene’s spotlight on Colgate seemed to saying, Look, we can cater to the needs to plain-old businessy businesses. GCP enabled Colgate to deploy the G Suite of productivity tools “to 28,000 employees in less than six months.” Wow, touting cloud-based office software: that’s a wonderfully mundane thing to brag about in 2017.
Clearly, Greene is whipping GCP into shape as an enterprise competitor. Google Cloud has hired more new employees than any of Alphabet’s other divisions; it’s also working to bolster its partner community. Google claims that in competitive bid situations, it’s awarded the contract 50 percent of the time, an increase from prior years. Not surprisingly, Microsoft and AWS dispute this; an AWS executive told The Wall Street Journal that “We rarely see (Google) among the last two finalists for big deals—especially when it comes to larger customers or enterprises.”
Whatever the truth of that issue, GCP is rolling out a smorgasbord of tools to entice enterprise clients. For instance, a free Virtual Machine transfer service, which facilities transfer from public clouds or on-premise environments; Cloud Dataprep, which allows data to be more easily prepared for processing; and support for Microsoft’s SQL Server on GCP. In fact Google put out a byzantine slew of announcements at Next; 100 by its own count, an amount no regular human could digest. Apparently the strategy is “let’s overwhelm them with new stuff – we are that committed!”
Press conference with Google executives; SVP Diane Greene at far left.
Along these lines, Google pushed two points hard this year. First, it promoted itself as the most resilient, reliable cloud, offering five nines – 99.999 – of uptime. Greene said that GCP had been designated as “having the highest availability of any cloud over the course of 2016.” (The figures come from CloudHarmony, a division of Gartner.) Then, in a clear jab at AWS, which suffered a recent outage, said added, “I think 2017 will be promising, too.” The audience chuckled heartily.
Microsoft issued a statement to dispute Greene’s claim, pointing out that it has a larger number of regions globally: “When looking at average uptime across regions, rather than total downtime across a disproportionate amount of regions for each provider, Azure reliability is in line with that of the other cloud providers measured…” Microsoft added a zinger of its own: “What we hear from our customers is that uptime is a more useful measure of availability.” Ouch! Take that, GCP. Google, for its part, promoted its expanding geographical coverage.
GCP’s most aggressive tactic to woo business was its price cuts, the surest path to the heart of any business. The so-called “race to the bottom,” price-wise, has been years long and Google has lost no enthusiasm for it. It cut Google Compute Engine prices up to 8%, offered Committed Use discounts of up to 57%, and extended its free trial from 60 days to 12 months.
Google also promotes its ability to get new cloud customers up and running quickly. Carousell, a marketplace app based in Asia, transferred from a competing public cloud to GCP in one month. Productivity software firm Evernote required just 70 days.
But those rapid on-ramps don’t reflect the experience of many established enterprises. At the conference, I spoke with a potential cloud customer, who attended to check out Google’s offering. He didn’t want his name used. What matters to you, I asked? He’s looking for reliability, support, and wants to avoid vendor lock-in – this last was particularly important. He said it would take 3-4 years to migrate to the cloud. That long, I asked? Yes, and some companies take longer, he opined. “If some cloud company says it takes six weeks, they mean a little startup takes that long.”
Perhaps for a customer like this, one of Google’s big sales pitches will be particularly compelling. Google points out in its marketing material that we’re living in a multi-cloud world; customers use more than one cloud vendor. Translated: even if we don’t get all your cloud business, perhaps we’ll get some of it. There’s a pragmatic humility in this. It’s the kind of pitch you hear from a backbencher; you don’t hear AWS talking about other providers.
Yet Google is correct about the dominance of the multi-cloud model. Companies will (and already do) use several cloud vendors, based on what that vendor does best. A company might use the Azure platform to boost its legacy infrastructure, and simultaneously deploy to GCP to leverage Google’s data analytics and machine learning offerings.
The pioneering days of cloud computing, say 2010 or so, were about accessing a remote platform. Business welcomed the advantages of extending their cumbersome compute platform using an external provider. Startups could pretend to be established by renting major compute power. Now in 2017, it’s about accessing the next-gen tool set that cloud vendors offer, from cloud-based Big Data, niche vertical software, AI, even security tools. No internal data center can create all these tools on its own. Cloud computing is no longer a choice; to be competitive, businesses need these cutting edge tools from external providers.
Google will benefit from this al la carte strategy, which is enabled by a multi-cloud world. Other cloud vendors will offer similar tools, but certainly Google’s version of various key tech tools – particularly machine learning and data analytics – will be at the forefront.