Hobson’s Choice

While it’s difficult to engender sympathy for one of the most powerful technology companies in the industry, you’ve got to feel a bit of a twinge for IBM’s Software group at the moment. The Oracle/PeopleSoft saga places them in a rather awkward position. According to evidence submitted at trial, IBM views an Oracle takeover as a threat to its software business. Although such a conclusion is not exactly rocket science, viewing those words written on the virtual equivalent of IBM stationary was noteworthy in the least.

So you have to wonder, in the wake of IBM’s announcement of a huge strategic alliance with PeopleSoft, what will the company do to protect its investment of hundreds of millions of dollars in technical and marketing resources in the event that Oracle veers closer to its goal?

You might recall that five years ago, IBM made a smart decision to exit the applications business, an area it never understood. If you used computers in the 80s and had the misfortune to use IBM’s DisplayWrite word processor, which forced you to navigate three or four menus to simply move or copy a word or phrase, you’d understand what we mean.

Instead, IBM Software has focused on infrastructure. Consequently, application vendors can take advantage of IBM’s market clout with little fear of being undercut with competing IBM product. And if you look to today’s leaders in the software industry — Microsoft and IBM — a huge pillar of their success has been the ability to cultivate a huge ecosystem of third party solution providers.

So maybe you can appreciate the tug of war occuring at IBM Software. Why break from a strategy that’s delivered? On the other hand, they might not have a choice. PeopleSoft gives IBM their best shot at middleware for the ERP market, because SAP is already developing competing product. If PeopleSoft’s history, how do they replace all that potential business, not to mention accounting for all the money invested in the relationship?

In the past 24 hours since announcement of the PeopleSoft alliance, IBM has faced plenty of questioning about whether they’ll finally break down and buy PeopleSoft. And IBM will continue facing those questions unless DOJ wins an appeal or some other white knight comes to the rescue.

The Myth of Open Systems

When was the last time that you didn’t hear a vendor call their system “open?” Although a buzzword today, it’s hard to believe that at one time, the term once meant something. Prior to the client/server era, most systems were closed, meaning, the only way to pry data from a legacy system onto a PC or relational database was through a flat text file extract, if that.

In theory, open meant vendor-independent. In practice, it meant systems that could readily access information across platforms. With client/server, you got a hint of open: relational databases that used various dialects of SQL, servers with various flavors of UNIX, and clients with one or two flavors of Windows. As reflected by the explosion of third party software, that created real progress. The Internet then blew matters wide open with standard networking, navigation, and the universal HTML browser. Now, if you could engineer that first back end connection, you could open up even the most proprietary of legacy systems.

The debate wasn’t over however. Religious wars broke out, primarily over Microsoft’s client dominance. They intensified with the emergence of Java, which following litigation, Microsoft answered with .NET. The Sun/Microsoft détente notwithstanding, a recent panel session where IBM and Microsoft representatives kept exchanging digs reminded us that the debate over who is more “open” continues to rage.

Folks, it’s time to slay this monster once and for all. Software will never be 100% open. Vendors want to open their software just enough to make it interoperable, but not to the point of making it easily replaceable. Therefore, while BEA WebLogic and IBM WebSphere are both J2EE-compliant, it’s quite difficult to switch from one to the other. Furthermore, Microsoft has little reason to publish Windows source code, while Oracle has scant incentive to make its database open source.

Vendors get far more generous with source code where the product itself isn’t strategic, but overall market control is. For instance, the user interface features of IBM’s Eclipse and Sun’s NetBeans Java environment are rather trivial, but when released to the open source community, they become stalking horses for third party support (IBM won thanks to WebSphere’s overwhelming presence). Similarly, C# means little to Microsoft’s core business, but donated to the international standards community, it provides a wedge into Java’s base.

Forget about “open.” Everybody’s products are open to a point. Instead, think interoperability, which is exactly why vendors are coalescing around web services.

Politics Not Numbers (II)

By now the news is sinking in about the U.S. District Court’s refusal to halt Oracle’s bid for PeopleSoft. As we’ve noted previously, we never believed that DOJ had a case. With core ERP growth prospects limited, the field is overdue for consolidation. On numbers alone, a combined Oracle/PeopleSoft combination would barely exceed 15%, according to recent Gartner figures.

But the case was about popularity, not numbers. The only problem with that approach is that U.S. anti-trust law is about the opposite. Yes, the market long ago rejected Oracle applications in favor of SAP, and to a lesser extent PeopleSoft. Nobody ever liked Oracle for its high-pressure sales tactics and forced upgrade marches, an impression that didn’t change after Oracle was forced by legal discovery to reveal that, yes, it really did plan to liquidate PeopleSoft’s products and organization. Ironically, while PeopleSoft is cast as the victim, it behaved virtually identically as it digested J.D. Edwards. Maybe it didn’t draw the scrutiny because those most affected were AS/400 customers, a segment that the market has largely forgotten about anyway.

Obviously this isn’t the last battle, but the only way for PeopleSoft to escape this situation alive is for Oracle’s own numbers to tank and for PeopleSoft to miraculously grow market share for a product whose future existence is questionable. Our hunch is that this is the beginning of the end.

Beyond the fact that antitrust law is not written for customers, we draw several conclusions from this soap opera:

1. For customers: Keep behaving like investors, evaluating vendor viability when you buy product. While you don’t have to stick with household names, you should (A) have an escape plan, (B) document all logic and interfaces, and (C) hope beyond hope that the brilliant startup that you’ve discovered eventually gets acquired by IBM, SAP, Cisco, or Microsoft. We hate to say that, but it’s true.

2. For vendors, don’t shoot for the upper right hand side of Gartner’s Magic Quadrant. Strive instead to become the kind of household brand that nobody gets fired for buying. We hate saying that, too.

3. For Oracle, your bid for PeopleSoft will be a pyric victory. If you actually pull it off, make it a reverse acquisition, where your product is yanked in favor of PeopleSoft. But seriously folks, could you really imagine Larry Ellison behaving so sensibly?