Main content

Oracle's cloudy day in the sun

Denis Pombriant Profile picture for user denis_pombriant May 8, 2017
Oracle's recent mixed analyst and media event at the company's HQ reinforced the mixed, hybrid cloud message. How well did the message go down?

Thomas Kurian - Oracle
Thomas Kurian - Oracle

Last week Oracle invited a gaggle of independent analysts and journalists to a briefing day in which company executives sought to educate us on the finer points of their business. And why not? This is part of a carefully orchestrated series of media events that both showcase the different aspects of Oracle technology while also reinforcing its relentless push towards a cloud only world.

The company’s financials are looking better today than they have in years; in part because as co-CEO Mark Hurd said they’ve been through the difficult period of converting their business to the cloud and are now showing results.

On the day of our meeting, the popular tech/finance site, Seeking Alpha, rated the 40-year old company’s stock a buy noting that for the last five years the stock hadn’t done much,

From 2012 to 2017 Oracle's revenue stagnated near the level of $37-38 billion, with diluted EPS growing primarily due to decreasing number of shares outstanding. As a result, the company's share price fluctuated around $40, from 2014 to the beginning of 2017.

However it also said, the share price had experienced an “increase of 16% year-to-date.”

The five years of treading water corresponds with Oracle’s decision to convert its substantial application base to the cloud, a process that’s still ongoing, and to convince its 400,000+ customers that it’s time to move onto new technology.

Oracle will be in the conversion process for about the next decade but the time of investing in new products with little to show for it is over. Most of the company’s customers are still reliant on Oracle’s premise-based solutions and that fact alone drives the rollout of a cloud solution set that might resemble Salesforce’s in its broad outlines but in fact is tailor made to support a huge conversion of the company’s legacy customers to the cloud.

As Thomas Kurian, president of product development, explained the company’s strategy and products support three major deployment modes—premise based, cloud at customer (essentially private cloud computing) and public cloud. It will be necessary to support all three if Oracle expects to bring a substantial majority of its customers along with little defection.

Salesforce, NetSuite, Financial Force, Apttus, Xactly, Zuora, and almost any other cloud business software provider you can name (yes, we’re skipping Microsoft and SAP for the moment) had the luxury of being born in the cloud and of only attending to software needs that they pioneered and for which there are few if any analogs on prem. Like Oracle, SAP and Microsoft both have sizable legacy bases so their strategies are similar.

Converting a legacy enterprise customer base is not something that happens over night. There are complexities, contingencies, and memories in the enterprise that must be attended to. Most important of these are the memories of twenty years ago when every IT shop changed its financials to accommodate the dreaded four digit date format. The people who worked day and night before and after the conversions are now managers and the memory of a major rip and replace might have added PTSD to the IT vernacular. At any rate it is enough to slow the migration to the cloud from a stampede to a stately march.

No wonder Mark Hurd loves to be quoted saying that taking the enterprise to the cloud will be a slow and steady effort,

I expect the era of hybrid computing to go on for a decade,

he told us. He’s right for many reasons.

It also explains parts of Hurd’s strategy to be number one in all phases of cloud computing including IaaS, SaaS, and PaaS. Infrastructure and platform are not typically things that born in the cloud vendors market or hang their hats on. When you buy Salesforce, for instance, you buy (okay, subscribe to) the apps with the assumption that infrastructure just comes along as a necessity that the customer doesn’t have to think about.

But if you’re a legacy customer infrastructure could not be more important. Frequently, legacy IT will need to first replicate its data center in the cloud before it can consider optimizations that come with new cloud apps that can be bought or built in house. That’s why infrastructure and platform are so important and why Oracle has set its goals so high. As a practical matter infrastructure is a commodity, all you need do is look at Google’s and Amazon’s efforts. But without infrastructure to move customers to, it would be much harder to retain Oracle’s customers hence the focus on infrastructure.

Platform is more of an area of contention between Salesforce and its rivals because the company has made its development facility a product too. Oracle’s strategy which of necessity requires transplanting enterprise data centers to the cloud needs to focus on infrastructure and platform as well as apps and that means big conversations about things other vendors skip. Because of this, Oracle has developed an elegant solution that stresses high availability cloud support.

According to Kurian there are 10,000 people at work building the cloud portfolio and Hurd boasted that spending on R&D was in excess of $5 billion in the current fiscal year, up from about $3.5 billion when he came to the company.

Kurian explained in detail how the company doesn’t deploy single data centers but three at a time in what it calls an availability domain. In these domains, each data center is supplied by a separate electric grid and Oracle is careful to avoid earthquake zones. Hopefully they have a minimum elevation above current sea level as a hedge against global warming (sorry, Miami). All regions use a dense fiber network designed to be fault tolerant and offer high bandwidth. And all of Oracle’s compute and storage hardware is available through the cloud as well, a point that the company uses to demonstrate superior performance over its near infrastructure competitors.

Kurian said that the availability domain is basically a flat structure and can be expanded enabling the company to land at a basic cost and expand as demand increases. Most interesting to me Kurian revealed that the company had developed management software for each domain that it refers to as a data center operating system.

My take

Oracle’s recent quarterly results have shown a company at the beginning of a hockey stick curve thanks to the availability of its cloud products. For example, Oracle now calls itself the fastest growing cloud company, a statement that will need unpacking because it can easily show high growth off of a small base; any company can - and then there is the vexed question of how software companies run their 'counts.'

But the recent quarterly results show 1125 new customers—that’s net new not conversions from the legacy base—and 908 expansions. While these numbers are good, it would be nice to know how fast the legacy base is converting or at least taking on parts of the cloud solution set.

That highlights an opportunity and a challenge. Cloud computing has made it possible for this enterprise vendor to attract mid-market companies that were harder to attract before.

Companies like Twitter and Trek, the bicycle maker that might not even have formal IT shops to sell into are now in Oracle’s stable. At the same time, Oracle is still selling aggressively into the enterprise and last week announced a major deal with a long time customer AT&T to provision its cloud field service dispatching services for 70,000 field service technicians.

The fact Oracle is able to traverse these widely differing customer types (and sizes) speaks to a flexibility in approach that I know our own Den Howlett appreciates.

So Oracle has entered a phase between on-premise and cloud computing with a strong emphasis on cloud done its way. Future comparisons between it and its rivals must take this into account. For instance, a new customer for infrastructure should not be counted the same as one that buys cloud apps. But since we started this talking about revenue, the only thing that counts is that number. Undoubtedly we’ll have interesting wars of words trying to figure out what a cloud customer is. The answer is, it’s complicated.

A grey colored placeholder image