I got back from San Francisco a few days ago, after attending Oracle’s annual conference, and have been ruminating ever since. You can see a summary here (I wrote some news coverage for CRM magazine), but that’s journalism—there’s no place for my own opinions. What follows are the impressions that don’t belong in a neutral-toned article. Nothing horrible, but still inappropriate for news articles outside of the New York Post.
I should point out first, however, that Oracle paid for my flight and hotel accommodations, and treated me to a couple of meals to boot. The Oracle analyst relations people are top notch, and the company is smart enough to allow them to do their job without undue interference.
1. Oracle Is Too Big. This should not surprise anybody. I don’t even think it’s necessarily a problem—economies of scale are important for industries valued in the billions of dollars and touching every facet of life in the developed world. It becomes a problem when you have a diverse array of products to display, but only one opportunity to do so.
OOW11 was two conferences, one for hardware and one for software. Unfortunately, the two conferences were co-located and ran consecutively, so all the hardware people got their content first, and then all the software people got theirs. This was especially evident at Larry Ellison’s two keynotes. The first, on Sunday evening, was a drool session for server wonks, with nary a bone thrown to the applications crowd. The result? A number of walkouts, and scads of Twitter heckling. The second, on Wednesday afternoon, introduced enterprise social networking tools and the Oracle Public Cloud—a big deal for apps people, useless for server people. More walkouts, and probably some heckling as well.
This approach isn’t likely to change, either. Throwing two events would be more expensive than throwing one big one, and Oracle’s new Engineered Systems initiative will bind hardware and software even tighter. Should it succeed, there will be even less reason to separate the shows.
2. Larry Ellison Doesn’t Get People. He’s an extremely sharp fellow, this Larry Ellison. He’s passionate about Oracle technology, and he’s an absolute shark for business. But he hasn’t figured out these flesh creatures around him. Sunday’s widely-ridiculed talk about nuts and bolts—Oracle Exadata, Exalogic, and new Exalytics servers, and the new SPARC SuperCluster general purpose megaserver—was passionate enough to hold my interest for a while, despite being irrelevant to my immediate needs. It made a strong business case for the devices, and was really a love letter to the technology. No doubt about it: Oracle has some very sexy tech, and it runs the world.
But the whole thing was numbers. “Ten times faster than X! One-fifth the power consumption of Y!” Complete failure to engage people, to tell a story that sold these behemoths on anything but raw capability. Attendees of OOW10 said it was like he picked up right where he left off the previous year, with as little regard for the audience as he exhibited then.
By comparison, the Wednesday keynote showed us a different Ellison. His presentation, likely due in part to the poor reception from Sunday and the gaffe of cancelling Marc Benioff’s scheduled Tuesday address, was lively, pojnted, and full of humor and fire. I don’t know Larry Ellison personally, but I’ve observed him over the years and seen him speak on several occasions. This was the first time I felt he was human, and I liked it. He rose to the occasion, introducing a family of Cloud apps whose relevance to individual users as well as the enterprises that employ them was clear.
This second address wasn’t perfect. It was largely devoid of specifics and, coming on the last full day of the conference, left little opportunity to get more information, or even build up much buzz. Introducing the Oracle Social Network and Oracle Public Cloud earlier on would have given us industry analysts and reporters a chance to talk amongst ourselves, dig for details, and basically do Oracle’s PR work for it. Instead, we spent three days begging for scraps, and Oracle leaders like Anthony Lye and Steve Miranda were reduced to telling us “there’s an announcement coming on Wednesday, and we can’t talk about it.”
3. If You’re Going to Give Free Press to the Competition, Do It on Your Terms. If you heard a loud bang on Tuesday night, it was Oracle shooting itself in the foot. Marc Benioff, chairman of Salesforce.com, was scheduled to give an address at OOW11 on Wednesday, as he’d done for at least the previous two years. At the last moment, Oracle (Larry Ellison) cancelled the address—changed the time, actually, to Thursday 8 a.m., after many attendees would have already left. This, combined with Marc having a prior commitment in that time slot, effectively killed the session.
Perhaps inviting Benioff in the first place was a bad move. He’d been critical of Oracle’s strategy and products in years past, and there’s no reason to think this year would have been different. The social tech Oracle was introducing would put the two companies into more direct competition, so providing a podium could be a risk.
As bad as it might have been, blocking the address was a HORRIBLE move. Marc Benioff is a master of the public address. Every word out of his mouth sells Salesforce.com and the vision of Cloud computing. Ellison’s actions removed any constraint Benioff might have had to be a gracious guest; they cast Benioff in the role of injured party and Ellison in the role of jerk; and they cost him money by making him have to get a different venue at the last minute. In other words, shit just got real.
Marc was able to use his time to attack Oracle much more fully than he could as part of the Open World calendar, pitting Salesforce’s fully-formed and successful Cloud model against Oracle’s still-unannounced one. Larry’s rebuttal in his own Wednesday address was intriguing and pointed, but it didn’t have enough meat on its bones. For the first time in history, Oracle was fighting outside its weight class. It’s believed Larry set this debacle in motion on his own initiative, which means there’s nobody else to blame. Stupid move from a very smart man.
4. Follow Your Announcements With Facts. I am really looking forward to seeing OPC and OSN in action. Their introduction alone was worth my attendance, and 2012 is going to be all the more exciting for the social technology crowd because of it.
The old show biz mantra is to always leave the audience wanting more. OOW11 took it to an unreasonable extreme, but it left us wanting anything. I can tell you little more about Oracle’s social and Cloud initiatives than that they exist, and there are some early adopters. I can’t name the early adopters because of NDA. I can’t tell you what the apps do, because the demo was sparse and the people who could tell us more were gagged. I can’t even tell you when to expect to see them in the real world, because the company’s official line is that no release schedule has been set beyond “over the next several weeks.” This is not how a company generates buzz. It’s a great way to make us industry watchers very suspicious of what we’re being shown.
Time spent by top executives deflecting questions could have been spent arming us with the facts we need to get the message out, all by tweaking the announcement date a few days. Now we have to beg for follow-on briefings and demos when available, hoping that satisfying our curiosity will wash this bad taste from out mouths.
5. Please Invite Me Back Next Year. Oracle is an incredibly important company. Even when Open World is a misfire, it provides valuable information, access, and networking. My criticisms are honest, and I offer them in the hopes you’ll make OOW12 much better for us, and thus for yourselves. Prove me wrong about what I perceive as your mistakes. I look forward to it.
Tags: cloud computing, enterprise 2.0, Open World, Oracle, Salesforce.com, social computing
