If you’re one of the geniuses who was calling for selling Oracle shares ahead of the earnings call yesterday, you might want to recheck your calculations. On your slide rule.
The headline on the press release is all you need to begin the self-recriminations: “Q4 FY19 GAAP EPS UP 36% TO $1.07 and NON-GAAP EPS UP 23% TO $1.16”
What were they thinking? More importantly, what were they thinking with?
I’m not going to quote every statistic, and there were bright and not-so-bright spots in the news but bottom line, there was more that was good than not or as Co-CEO Safra Katz put it,
“In Q4, our non-GAAP operating income grew 7% in constant currency—which drove EPS well above the high end of my guidance.”
“Our high-margin Fusion and NetSuite cloud applications businesses are growing rapidly, while we downsize our low-margin legacy hardware business. The net result of this shift away from commodity hardware to cloud applications was a Q4 non-GAAP operating margin of 47%, the highest we’ve seen in five years.”
So what’s going on? Oracle is executing a pretty run of the mill pivot of its business, if run of the mill is a phrase you could ever use when betting the farm, but there it is. When the pivot started, Oracle had little more than ambition to reference. It had no cloud datacenters and the company with the most experience running the Oracle DB in the cloud was, uhm, Salesforce.
At that time, then CEO Larry Ellison, told anyone who would listen that the pivot would take about 10 years. That’s right, ten years to move the majority of 400,000+ customers and their data to cloud apps and a new database built for the purpose complete with self-driving capabilities.
So, where are we?
New data in G2 Crowd’s Enterprise Grid Report for Relational Databases/Spring 2019, shows that 17 percent of Oracle’s database is in the cloud, 83 percent on the ground. In other words, about where Mr. Ellison might have predicted.
Of course, G2 Crowd also reports Oracle’s nemesis Amazon Relational Database Service at 100 percent cloud, but it also indicates that the median number of users bought by Amazon customers is 37 users vs. Oracle’s 1,750. Even Amazon Aurora only clocks in at 175 users. In other words, the cloud DB competition is largely still in the sandbox. I’m not casting shadows at Amazon, their Prime service is pretty good and at least they haven’t offered up a cryptocurrency yet.
Apps are a different story and there’s work to do there. Even assuming a report like Satisfaction Ratings for CRM, also a G2 Crowd product, ranks Salesforce with 84 percent likely to recommend and a Net Promoter Score of 41 against Siebel 66 percent/-15 and Oracle Engagement Cloud 74 percent/-8 and Oracle On Demand 63 percent/-22.
Quick word on NPS, the scale runs from -100 to +100 so a zero is pretty good and 41 is awesome.
A full look at the G2 Crowd CRM report indicates that the larger and better-established companies struggle with keeping their growing customer list happy while the newer entrants like Salesflare 92 percent/78, and BPM’Online 98 percent/79 have gaudy, almost unearthly numbers. G2 Crowd says averages are 79 percent/31. So Salesforce’s 84 percent/41 is extra impressive.
My two bits
Oracle is a bell weather for the industry so we’d rather see them doing well than not. That said, they’re also riding a secular change in tech that’s aggressively moving to adopt commoditizing solutions like cloud, analytics and machine learning. In this they are on track rather than a leading outlier. Nonetheless, other market leaders have stumbled during this kind of transition so it’s good to see Oracle’s numbers especially in the current economic environment.
All this notwithstanding, economic storm clouds are on the horizon. A story in the New York Times pins at least some of the problems on the US trade war. Miscalculation beyond its control, in several dimensions, could upset Oracle’s applecart and other warnings for a bumpy 2020 are also making the rounds. It’s impossible to say how it all turns out, of course, but if there’s a way to make the numbers, the team at ORCL seems capable.