For investors, this “we’re bigger than BEA” litany raises the following questions:
· What’s taking Oracle so long to achieve its goal?
· How is Oracle finally going to put this obsession behind it?
· Why does Oracle’s place vis a vis BEA matter to investors?
First, it has taken Oracle some time to catch BEA because Oracle originally tried a different technical approach to the “middleware” market. Although Oracle tells investors and prospects that it “only entered the middleware market” four years ago, in fact, in 1996/1997 Oracle was one of the first software providers to offer application/web server code (most people’s definition of middleware, along with MOMs, ESBs, portals and transaction monitors). The product was an object-technology-standard-based code set providing what Oracle called a Network Computing Architecture [NCA]. Although NCA was an admirable technical effort, information-technology [IT] developers and administrators wanted deployment code such as the open-source Apache HTTP server, the WebLogic application server acquired by BEA in 1998, and the IBM (IBM) WebSphere application server announced around the same time as Oracle NCA and all based on a different technology.
It took Oracle a couple of tries after its object-technology NCA development work to get its present Apache/WebLogic/WebSphere-like product lineup in place. And by then (Oracle’s mythical “four years ago”) IT developers, even satisfied Oracle relational database users, were leery that the Oracle9iAS product lineup, based on application server container technology licensed from IronFlare, would not be Oracle’s final approach. Even Oracle’s own application development group, the people in charge of eBusiness Suite, moved slowly to the IronFlare-based platform. (As for Oracle’s “four years ago” claim, when you push back on them they say they mean Oracle only entered the “middleware suite” market four years ago.)
However, and answering the second question, Oracle is finally going to achieve its long-sought goal of passing BEA because of how it defines “middleware,” how it defines “pass, and possibly even how it defines “new.” This week Oracle said that it will "pass BEA in total middleware new license sales later this year." Since "this year" is over from a calendar perspective, Oracle must mean it will “pass BEA” some time in Oracle’s fiscal year, ending May 31, 2007. Legitimately, Oracle is having success up-selling to its acquired applications user base what it calls middleware (for example, Siebel Analytics, other business intelligence products, collaboration software, data hubs, Oracle Forms and Reports software) along with those products in its lineup that most people would agree are middleware (for example, the OC4J application servers, identity management, portals).
Of course—asterisk--BEA does not sell any products in the first category so Oracle’s comparison with BEA is apples to oranges to begin with. And—asterisk asterisk—to “pass” BEA Oracle will likely be comparing its historically high final fiscal quarter, ending May 31, 2007, with BEA’s historically low first fiscal quarter, ending April 30, 2007. Here’s how the "we-passed-BEA” arithmetic will likely work out:
· In the quarter ending April 30, 2006, BEA had $132.4 million in “license fees.”
· In the quarter ending May 31, 2006, Oracle had almost $1.5 billion in “new software licenses” in database/middleware.
· To beat BEA in new license sales in the comparable quarter in 2007, Oracle will need about 10% of its SEC reported database/middleware new software license revenue to be for middleware.
Of course, we’ll have to take Oracle’s word for it when and if it “passes BEA” in total middleware new license sales because Oracle announced in September 2006 that it would stop breaking out middleware license results from database license results, after reporting them separately for s
ome time, making investor research on this subject more difficult.
That’s not to say separating the Oracle middleware results out from the database/middleware total is impossible. Modeling results and what-ifs based on past Oracle SEC-filed statements, quarterly-call transcripts, and some other assumptions, Oracle is still probably a year or more away from “passing BEA” on a more meaningful trailing-12-month basis. For example, using the 10% multiplier mentioned above, Oracle recorded about $375 million in “total middleware new license sales” in the 12 months ending November 30, 2006 whereas BEA recorded $559 million in the 12 months ending October 31, 2006 (Note: BEA is in the midst of the same stock-option-related delay in filing 10Qs as many other companies and has said it will be restating quarterly results but such a restatement is unlikely to affect the top line).
Perhaps the right multiplier to use for estimating Oracle middleware is 12% or 13% instead of 10% but I doubt it for a couple of reasons: (1) the numbers don’t compute when triangulated against historical statements about Oracle database license size and growth, and (2) Oracle would have already “passed BEA” in May 2006 if the multiplier were much higher than 10%. And if that had happened, I’m sure we would have heard about it.
However, all the math and apples-oranges mind bending aside, unless BEA puts its cash hoard to work on an acquisition binge, Oracle will eventually pass BEA because of Oracle’s much broader middleware product portfolio.
So, the third and most important question for investors is “Why does Oracle’s place vis a vis BEA matter?” The answer is “it does not,” at least as long as BEA participates in only a subset of the middleware market as Oracle defines it. The middleware market goal posts are moving just as Oracle gets to the goal line. Oracle’s position in the middleware market vs. IBM, Microsoft, and SAP is much more important in analyzing Oracle’s long-term potential. Oracle’s problem in this comparison is that, as with its applications lineup, because of both acquisitions and legacy middleware development/marketing efforts such as Forms, Oracle has multiple middleware architectures to support and try to rationalize. That was true of IBM as well after its purchase of Crossworlds and other companies but IBM has a 3-year head start in pulling its middleware products together. SAP’s and Microsoft’s middleware offerings are much crisper in terms of a single architecture and cohesive offering (although both have evolved over time).
Even in terms of middleware market share, Oracle likely suffers vs. IBM, Microsoft (MSFT) and SAP (SAP). In a recent IDC press release about the middleware market narrowly defined (even more narrowly than I would define it), IDC ranked the market leaders as IBM, BEA, Oracle and Microsoft. But if one were to estimate the value of the application server/transaction management functionality (which is what IDC said it was measuring) that is built into IBM’s, Microsoft’s, Oracle’s and SAP’s applications as if they were sold separately, SAP would likely take BEA’s place as number two and Microsoft might also rank ahead of Oracle (depending on how you came down in a no-longer-useful single- vs. multiple-user operating system debate).
This expanded view of middleware is very important to Oracle (and therefore Oracle’s continued investment value) because it plans to use its rationalized middleware offering to normalize its six or more application architectures over the next 3-4 years and to further upsell and crosssell to its base. If Fusion doesn’t do the job, Oracle is unlikely to turn to Microsoft or SAP for help. Given the history of IBM with the Oracle J.D. Edwards products and ill-fated plans for IBM to work with the Oracle PeopleSoft products, perhaps Oracle would make a middleware deal with the devil. But the real easy answer for Oracle as a fallback middleware position would be BEA. So Oracle better not congratulate itself too strongly when it eventually crosses that goal line.
Disclosure: Author has no position in above-mentioned stocks.