Why Oracle CEO Larry Ellison had to go ... Except he hasn't

Silicon Valley's veteran seadog in piratical Putin impression

Ah, yes. Larry's glory years: the Network Computer, JavaFX, er...

You don't get to be CEO of a multi-billion dollar company without taking some wrong turns. Of course, you are going to make decisions that don't pan out – Larry's done that: he launched Oracle's Network Computer that failed. There was the Oracle collaboration suite that failed. And he backed JavaFX, the interface technology inherited with Sun in an era when devices are native and the web is HTML.

Larry's made plenty of good calls, too.

It was Ellison, Oracle's official history tells us, who came across something called relational database technology while thumbing through an IBM research book. He was working with his soon-to-be Oracle co-founders Bob Miner and Ed Oates back when flares were the last word in style, and the three of them worked for Ampex.

IBM's Edgar Codd conceived the idea of relational databases, and then published his thoughts in 1970. But Big Blue didn't develop a relational product until 13 years later. IBM feared that relational would kill its existing mainframe database business.

Instead, it was Larry and a rival named Michael Stonebraker with his Ingres and then Postgres companies who were first to market an RDBMS. Thereafter Oracle under Larry elbowed everybody else aside, and now Oracle is around half of the world's RDBMS market.

Under Larry, Oracle expanded out of the core database product, and pushed app design to feed the database engine.

Oracle was forward thinking on technology: writing its database in the then-relatively-new cross-platform language C, going for maximum market reach by working on mainframe, mini and PC; adding OLTP in 1989, Java in 1998 and running Linux in 1999. Oracle was the first big software maker to bet on the new open-source operating system.

These were moves that kept Oracle growing. IBM and Microsoft were late on RDBMS, the latter cutting off its potential by putting its RDBMS product SQL Server only on Windows.

Oracle was – and remains – expensive yet pragmatic. Yes, there was, and is, a price list, but if you can go mano a mano with your red-meat-eating Oracle rep in his black Agent Smith suit, you get the discount.

I remember at one Oracle OpenWorld in San Francisco, California, when a customer during an open mic Q&A with Papa Larry complained about his company's extortionate prices. Larry's advice? Rip a processor out of your server if you want to pay less.

You didn't get what you wanted by asking with Oracle; you got what you wanted by hard-nosed haggling.

Recent years have seen a different Larry, and a different Oracle.

Growth has been through acquisition: 84 companies since 2005, a grand tour that commenced with PeopleSoft and took out rival BEA Systems and then Sun. Oracle moved into new markets and took on customers, offering them yet more of its existing technology while hooking them into perpetual software maintenance. Oracle Borg'd technology where appropriate, killing its application server for the superior BEA WebLogic for example.

No change is change. This is the software biz, after all

The big question after "why" is "what's next?" First, let's put what happened on Thursday in context.

Larry has not left the company he led for nearly forty years, he's merely stepped down from day-to-day responsibility for running the business. This isn't like Steve Ballmer quitting as Microsoft CEO and then leaving Microsoft's board. Larry's presence remains.

He's been elected executive chairman of the board by the board, with incumbent Jeff Henley now his deputy. He remains Oracle's biggest shareholder, too. Supposedly running things are Larry's tennis-mate Hurd, who got his job thanks to Larry's personal patronage, and Catz – Larry's loyal deputy and the Oracle's chancellor of the exchequer. She's a company veteran.

So don't expect strategic change. You can certainly forget big decisions about tackling the biggest issue that needs resolving – what to do about the hardware business. Larry will continue to make the power decisions – it's just that Hurd and Catz will take the flak until the time is judged right for him to return. As one blogger put it on Thursday:

OH: "Larry Ellison stepping down to become CTO of Oracle is like Vladimir Putin stepping down to become Prime Minister of Russia."

— Bryan Cantrill (@bcantrill) September 18, 2014

Can Oracle continue with two CEOs? Yes, on paper. Larry's big business apps rival SAP has had joint CEOs for years, only recently letting its chiefs fly solo.

Joint CEOs for Oracle, though, is likely a short-term arrangement for a number of reasons: the fact that Larry is still present, and the fact that there's been no planet-wide search for a replacement. If this is a long-term appointment, Catz and Hurd must prove they can competently run the corporate supertanker that is Oracle. If this is the case, there will be an incentive by both to out-do the other.

Of the two, Catz should be considered the keeper of the Oracle flame. She'd also become tech's third serving top-tier CEO who is a woman, after HP's Meg Whitman and IBM's Ginni Rometty. Either way, as Oracle's single biggest shareholder, Ellison would have a considerable say in who, if anybody, is next.

None of which, though, gets Oracle out of the hole Larry's put it in: a flat software business and late entry to cloud with a stubborn hardware anchor. Digging out of that pit will take some time. Or an act of God. ®