Google Cloud sits a distant third in the cloud wars (ignoring the Chinese cloud companies, which are pretty much irrelevant to Western markets as they face a Huawei-or-the-highway reckoning). With a cloud infrastructure business generally estimated to be between a quarter or a third the size of Microsoft Azure, which in turn is about half the size of Amazon Web Services, Google has a lot of ground to make up. And they’ve been tilting at the infrastructure space since 2008.
You’d think AWS would have bigger worries than Google Cloud. Yet the recent AWS lawsuit against Brian Hall built up Google Cloud as a serious competitive threat to the industry leader. A thin-skinned AWS sued Hall, claiming he violated his non-compete agreement by joining Google Cloud after leaving AWS. The suit exalted Brian as a god-like marketer (he is pretty good). But it got ridiculous, focusing on Brian’s slide creation skills (which was ironic for a company that prides itself on not using slides – except when they do for hours on end, and then struggle with some of the basics), and conjectured (under oath) that just a few of Brian’s slides could radically reshape what is apparently a very precarious cloud industry.
The suit also highlighted the overly broad non-competes Amazon asks employees to sign and its haphazard enforcement of those agreements. After a month of terrible press for AWS that broadcast the risks any potential employee takes working for the company, they came to an undisclosed settlement and Brian promptly assumed his job at Google working for Thomas Kurian (Google Cloud’s newish leader and now recovering Oracle lifer).
I know and have worked with Brian (and am sure he’s horrified by this post), and supported him during the lawsuit (amicus Twitter). He’s an outstanding executive, a creative and clear-thinking marketer, and a high-integrity human. And his slide chops live up to their acclaim – I’ve seen him conjure up over $50 billion in value with slides on a Saturday afternoon that came to fruition and then some.
I’m increasingly fixated on deciphering company cultures (strategy becomes culture at successful technology companies, and then that culture inhibits the evolution of strategy) and using that perspective to handicap future prospects. Google Cloud is as good a place to start as any, as it has some deep and perhaps intractable cultural issues. As Brian tries to get Google Cloud out of their third-place rut, lets see what he is up against.
It’s a Hobby
Steve Jobs introduced the idea that trillion-dollar technology behemoths can have hobbies. By describing Apple TV as a hobby, the company restrained expectations and signaled it wasn’t core to Apple’s business. Apple TV was merely something to putter around with in the Apple Park garden after a particularly arduous week in the smartphone wars.
Google Cloud Platform is also a hobby, even if Google doesn’t admit it (they may not even realize it). When Google CEO Sundar Pichai gets up in the morning, the broader Google Cloud business (including G Suite) is, on a really good day, a distant fifth on his list of concerns. The four business segments ahead of it collectively constitute Google’s advertising franchise – search, YouTube, AdSense and Android – which are a mutually reinforcing magical money machine that generated about $150 billion in revenue last year (with very attractive margins).
Sundar’s attention is disproportionately on the advertising franchise as he circles the wagons against a pandemic ravaging (among other things) advertising spending, his fellow members of the avaricious online advertising oligopoly, and a barrage of regulatory and antitrust actions. The broader Google Cloud –both G Suite and Google Cloud Platform infrastructure services – was less than 7% of Google’s overall revenue in the first quarter of 2020. It just isn’t existential for the company.
Despite being a hobby, Apple TV exploits the same skill set and assets Apple employs in its day job. It is an integrated hardware and software product that leverages the company’s silicon, software, and services platforms. Apple TV sells to the same consumers through the same distribution channels, and deepens the customer’s ecosystem attachment to Apple.
Google Cloud not so much: it is an outlier and disjoint in almost every way from the rest of Google. It targets an alien and difficult to please customer in enterprise IT. It requires a completely different partner ecosystem and a completely different business model that is almost antithetical to the advertising business model. It requires dissimilar talent to build, sell, and support that is not particularly fungible with the rest of the company. Google Cloud even requires infrastructure increasingly divergent from Google’s alien mothership infrastructure.
The degree of difficulty to build an enterprise cloud business, never mind a market leader, is very high. It is even harder to build one within an advertising company. Expensive hobbies that cost real money by even Google standards are hard to sustain, so material progress is required. Google management appears to have put Google Cloud on a PIP, giving them a limited window in which to become a market leader or risk defunding. Can this peripheral business for Google rise to the top in cloud infrastructure? After over a decade in this market that hasn’t’t lived up to Google’s ambitions, what has to change and what can change to deliver a different result?
Losing Their “Googleyness”
Google in many ways invented cloud computing for their search business. And they did it in bold and unconventional fashion – not by using what were considered the most advanced computing technologies of the day – but by taking a greenfield and sui generis approach to construct an entirely new computing model melding novel software like containers, Borg, and MapReduce with commodity hardware.
What is the point of a Google Cloud that is losing its iconoclastic approach and turning into a me-too amalgam of AWS and Oracle slavishly chasing the market leaders? A product portfolio that increasingly sees itself through the lens of the atomistic AWS product line combined with an Oracle sales culture is – to put it mildly – hard to get excited about. And GOracle may be overindexing on databases at the expense of everything else. The last thing the world needs is another also-ran cloud company whose use of the word “enterprise” is inversely correlated with their market performance (after all, we already have both IBM and Oracle).
The “Googleyness” of the early Google Cloud was pioneering and unique enough to perhaps warrant overlooking their very underdeveloped customer skills, at least for a while. They might have matured beyond simply telling customers to get smarter and found a distinctive and complementary go-to-market approach, instead of remaking themselves in Oracle’s image. But that “we-come-from-the-future” inventiveness seems largely in the past, particularly amid an exodus of engineers out of the cloud group. The platitudinous new Google Cloud mission statement seems unlikely to uncork a deep well of future innovation (did someone win a bet for getting “digitally transform” into the final version or was that a placeholder they forgot to update or what?).
Google these days seems more focused on relitigating historical battles than pushing the infrastructure frontier. They are grappling with deep neuroses from both Hadoop and Kubernetes, having failed to reap the associated credit and financial rewards they believe they deserved. The company’s previous open source idealism has run smack into both ego and the financial constraints under which Google Cloud now operates.
Between Google Cloud’s hobby status and being in the midst of an identity crisis, Brian has his work cut out for him. But cloud infrastructure is an enormous market with lots of growth still ahead of it, Google is investing heavily, and perhaps with better slides, they could move up in the standings?