“If you liked Hurd,” began the email from an HP employee, “Then you’re going to love Leo!”

Yes, it was sarcasm on his part.

Mark Hurd was a darling with Wall Street and a demon to HP employees.  Hurd did what imported CEOs often do, which is cut expenses by cutting employees and burning deadwood.  HP employees lived under unemployment fears but HP’s bottom line improved.  Some souls who still fondly remember Bill and Dave and a kinder, gentler and more stable HP didn’t much care for Hurd (or Carly for that matter).  Where as HP once stood for innovation and nearly family-style employee relations, it now more closely resembled Oracle in its ruthlessness.

So it surprised nobody in HP that Hurd landed at Oracle after HP auditors found a soft-core porn star on Hurd’s expense report.

What HP staffers and the larger tech and investment communities were surprised about was when Leo Apotheker — the former and suddenly ousted CEO of SAP — was selected to take Hurd’s place.  Leo oversaw SAP’s stumbling in the face of Oracle’s slash and burn acquisition spree, where they purchased most of SAP’s competitors.  Larry Ellison was on a mission to move from being an infrastructure software company making relatively low margins, to being a total stack company that sold everything from centralized applications to the solder joints on server mother boards.  Key though was the upper layer of the stack — applications, where fat margins and user lock-in reign.

Which explains, in part, Leo’s new job at HP.

Apotheker was quoted declaring that software is the “glue” that holds together the different parts of the company.  Giving him the benefit of doubts, I’ll assume he was talking about IT consumers and not HP, which has never been known for software (some HP employees forcibly remove HP add-on software from their laptops because of the latter’s notorious instability, though the former are occasionally as unstable).  What IT software HP sells today is all infrastructure, though they dabbled in the past with apps (Slate/Word/Desk on their extinct HP3000 minis, AllBase in middleware, and other market losers).  Bringing a software maven into a hardware and infrastructure company indicates that HP’s board thinks software is essential to their future prospects. If you are looking for a new software to help you run your business then you need SalonTouch Studio.

No doubt Leo thinks the same way.

The problem is that HP’s software hole is huge. They have a few components of the stack’s bottom, and Openview remains a darling of the network jockeys.  Yet HP has nearly nothing in the middle and certainly nothing on top.  Compare this with Oracle (who has all but completed their mission to have a complete vertical stack) and IBM who has had many software successes and acquisitions.  Yes, HP may have brought Leo on board to start backfilling HP’s empty software shelves, but they should have brought in Caterpillar instead.

Their direction is as admirable as it is audacious. The higher up the stack you go, the more margin you earn and the longer the lock-in you have over customers.  Since IT customers still prefer having one throat to choke when possible, having a complete vertical stack increases hardware sales and generally makes good marketing strategy, one which Larry Ellison launched long ago and has now achieved.  This is why HP makes nearly 50% more in revenue per each employee than Oracle, but Oracle profits almost twice as much per employee.  That and Larry scares his employees so badly that they work 25 hour days just to avoid being placed on the rack.

Interestingly SAP earns more money per employee than HP and IBM and almost as much as Oracle.

Despite early rumors, it is doubtful that HP and SAP would merge (though HP has successfully merged with big companies before).  Such a significant and transatlantic marriage would require a significant degree of effort and delay, and then only bring HP into parity with part of Oracle’s application empire.  To remain competitive in the IT market and get the fatter margins HP’s board appears to crave, HP will have to acquire many best of breed solutions and orchestrate integration (which their EDS division will protest since custom integrations are a big chunk of their revenue base).  With a relatively paltry $15B in cash (Oracle has $24B and Microsoft $37B) HP has to choose acquisitions wisely and mergers that make sense from end to end (like Oracle’s acquisition of PeopleSoft and Sun).

Which may be the wrong strategy.

One thing HP got right long ago was the commoditization of the hardware market. HP used to make several proprietary platforms that ran proprietary operating systems (HP1000/RTE, HP3000/MPE, HP9000/HPUX) but now they sell a lot of Linux-running x64 boxes.  Yes, there is room for super servers at the hub of a data center, but volume-wise, the world is commodity hardware and HP’s merger with Compaq helped them with that and dominating the x86 PC market.

This may well happen to software too.

To make a commodity out of enterprise server application software, you have to develop a solution that becomes de facto through one or more clear benefits.  Linux came to dominate server operating systems because it was ‘free’, stable and cross platform.  Nothing in the marketing playbook says this cannot happen to applications.  Already we see several open source applications making major dents in competitors.  SugarCRM continues to creep into offices, and other bright folks are integrating various open source suites for interoperability.  Oracle’s edge in top-tier enterprise applications could suffer the same way that Sun’s position in servers did as Linux and x64 servers eroded their competitive advantage.

HP could hasten the process, though odds are they won’t.  Leo Apotheker knows high-margin software and HP’s board picked him because of that.  But the alternative to going toe-to-toe with Oracle would be to eliminate Oracle’s income streams by enhancing, integrating and popularizing no-revenue alternatives to Siebel, PeopleSoft, EnterpriseOne, etc.  They could fork MySQL or even acquire EnterpriseDB and swipe Oracle database customers.  In short, HP could avoid planting their own fields and just burn Larry’s crops instead.

There is no marketing lesson because it is too soon to know with certainty what Leo Apotheker intends to attempt.  But we do know that HP has options, and it all depends on if they believe applications will become commodities as well as servers and operating systems and DBMSs.

If HP doesn’t, someone else will.