- by Paul Murphy, -
Novell just announced its purchase of SuSe for $210 million - $50 million of it direct from an IBM investment in newly issued convertible preferred stock. This deal raises two very difficult questions: what's Novell up to? and why, if the people behind SuSe wanted to cash out, didn't they do an IPO?
I know that I don't know, but I suspect that this will eventually turn out to be a pivotal transaction in the history of Linux, one fraught with potentials which nearly all of us will eventually wish we had understood at the time.
Novell ran on Unix as early as SunOS 4.1.4 in, I believe, 1989 and perhaps before that. It never sold well, of course, because adding Novell to Unix only makes sense if you already have Novell -in which case you're not likely to have Unix. Nevertheless, Novell's current technical direction is clearly aimed at remaking Novell as a Linux based Windows Server competitor and so it seemed reasonable that they would eventually either buy a distribution or make one. Since Red Hat's culture and market cap put it out of reach and issues associated with the SCO lawsuit could prevent Novell working with the former Caldera Linux team, I expected them to buy Xandros as the obvious complement to Ximian.
SuSe was, therefore, a surprise - and one that raises more questions than it answers.
First, Novell's record with acquisitions is generally dismal. David Every, in an essay on the history of DR-DOS memorably describes them as having "the septic touch" in that they tend to destroy whatever they buy. For example, Novell completely failed to realize on the value of DR-DOS and eventually tossed it to Caldera -which then sued Microsoft over its destruction and won a settlement minimally estimated at $200 million. The repeat performance, this time with AT&T's Unix rights, should eventually net a considerably larger settlement but, again, not for Novell's shareholders.
Novell's relationship with WordPerfect was different in scale but not outcome. The other acquisitions amounted to little in terms of the company's net worth, but in this case they paid top dollar for a company on the verge of collapse and only narrowly survived the experience by refocusing on Netware and selling WordPerfect at a considerable loss. In this case to Corel, a company which later developed and sold its own Linux distribution, as Xandros, to the same people (New York's Linux Global Partners LLC) who funded Ximian's development and were recently instrumental in selling that to Novell.
The current CEO came to Novell through an acquisition - of Cambridge Technology Partners, a contract developer which died from excellence. (Seriously: CTP's business schtick was to sell fixed price, fixed budget, IT development projects in the boardroom whatever the objections of the incumbent IT group and then hand over operational support to that incumbent IT group. CTP generally did its job, but the people they handed over to were usually able to show that CTP's apparent successes were really dispicable failures and the executives who had hired them correspondingly of dubious ancestry, doubtful judgement and questionable character.)
In each case most of the people who worked for the acquired company now work elsewhere. Immediately after the CTP merger, for example, Novell had 2,400 more employees than it does now with relatively none of the CTP people left outside the executive suite.
There's a reason for this record: the fundamental Novell product and its market is too narrowly focussed to provide a base for expansion into fields like consulting or Office applications. Thus it might make perfect sense for Novell to use a Linux kernel within Netware but it's hard to see how it would make any more sense for someone with Linux/Samba installed to add a Netware license now than it did for a SunOS user in 1989.
That the people at SuSe wanted an all cash deal suggests that they may have been aware of the track record but, possibly unfortunately for them, EU approval is not assured -meaning that they may have shown themselves willing to sell the company on the cheap but will ultimately be left holding a badly wounded SuSe with little or no cash compensation.
The IBM/SuSe partnership is important to a wide range of national infrastructure projects in Germany, all sold in large part on the value of having a German owned, German developed server-to-desktop counterweight to America's Microsoft. For example, the cities initiative (in German) under which Munich made its decision to go with Linux is part of a national plan to develop and support a European alternative to American software dominance.
At this point we don't know what assurances, if any, Novell has from the EU authorities on this transaction but it seems more far likely than not that there are serious people meeting in Bonn and Brussels right now with the goal of stopping this deal --and their enthusiasm for letting American companies take over strategic IT assets at home is likely to hit serious new lows if, or rather when, the interim Iraqi council gets US approval to formally repudidate Saddam's debts to France and Germany.
How then, did this deal come about and what will its consequences be?
On the Novell side Occam's razor suggests that simply taking them at their word - and their track record - is the most reasonable. Pending other evidence, therefore, we should believe that Novell is sincere, will complete the deal despite EU opposition, and will once again display "the septic touch;" thus taking SuSe out of play, tremendously strengthening Red Hat, and opening the way for another European distribution to develop major mind share. For example, Groupe Bull, which frequently acts as an IBM surrogate in France, will probably expand its partnership with Mandrake (whose Linux already runs on Bull's rebadged IBM servers) to spin-off a new German distribution.
On SuSe's side, however, the more sincere we believe Novell to be, the harder it is to explain the price accepted.
Red Hat has about 600 employees, a cumulative net profit since start-up considerably below total commissions paid to date on 1.9 billion share transfers, a recently restated "back to basics" business plan guaranteed to further antagonize developers, and a market cap of about $1.8 billion at the end of October.
In contrast SuSe only has about 390 employees but has been picking up market share both from new Linux converts and from the Red Hat installed base, has significant partnerships in place, major, and promising, sales initiatives underway, and a reputation for fair dealing that's given it tremendous customer and developer loyalty. Its owners, however, sold the company for only $210 million - less than 12% of Red Hat's market cap.
Given present management, known product quality, and the buildup the company has been getting in the European press, it's difficult to see how an IPO, whether treated as German or pan-European, wouldn't have netted five times that despite the bubble death of Frankfort's Neuer Markt and the German institutional preference for large traditional companies run by the scions of noble, or at least rich, families.
Equally interestingly, you have to be a conspiracy theorist not to see IBM's board ante -the $50 million new investment in Novell- as anything but defensive: meaning that SuSe's closest business partner didn't see this coming in time to prepare a better contingency plan. That, of course, would be odd under any circumstances but what makes this truly weird is the strong rumor that Sun turned the deal down before Novell got into the picture because, if true, the deal cannot have come as a surprise to IBM.
So why did the owners sell SuSe to an American company for considerably less than it seems to be worth while sending their premier partner scrambling for continuation cover?
One idea, which will appeal to the conspiracy minded, is that IBM could be getting ready to lay off the SCOsource lawsuit on SuSe as the primary vector through which AIX source got into the Linux development tree. In that situation Sun certainly, and Novell probably, have the AT&T rights needed to provide cover and so constitute the entire market - except possibly for Lucent and Avaya neither of whom would likely be interested. A Sun decision to turn it down, would therefore produce a firesale to Novell.
It seems far fetched and, as I said, I don't know; but I'm quite sure that we'll all eventually be smacking ourselves upside the head and wishing we'd seen "the obvious" at the time this happened.