Upstream vs. downstream complementarities in Apple-IBM deal

Despite coverage to the contrary, Tuesday's announcement that IBM will help sell Apple products to enterprise customers is long overdue and merely the latest example of cooperation between the firms spanning more than two decades.

The new thing is that — unlike previous deals — the cooperation announced by Apple CEO Tim Cook and IBM CEO Virginia Rometty reflects downstream complementarities rather than upstream ones.

In their seminal book Co-opetition, Brandenburger and Nalebuff defined complementarity between products X and Y as meaning that if someone bought X, then Y would be more valuable (and vice versa). This basic principle is behind nearly any positive-sum (“win-win”) strategic alliance today.

Yes, Apple’s early successes began to fade when IBM introduced its PC in August 1981, and its 1984 Macintosh introduction was aimed squarely at Big Blue and its user-unfriendly DOS PC. But the two companies have been cooperating far longer than they competed, largely through their cooperation in upstream components.

The big cooperation surprise came not in July 2014 but October 1991. Then Apple CEO John Sculley and IBM President Jack Kuehler announced that Apple would be using PowerPC CPUs based on IBM's proprietary RISC chips (Motorola was the third partner in the alliance).

At the same time, Apple and IBM launched two Silicon Valley-based software joint ventures based on a common rivalry with Microsoft. Taligent nearly killed Apple (and thus helped me find a new career) by siphoning off Apple’s top engineers to work on an operating system that it never shipped. Kaleida was intended to solve CD-ROM scripting challenges, but was swept aside by the emergence of Java and the commercial Internet.

A few years later — at the depth of Apple’s self-inflicted slide towards irrelevance — IBM helped Apple with problems creating hardware in its fastest-growing and most profitable segment, laptop computers. It sold its unique laptop hard disk to Apple (but not to HP) and also built the 1997 PowerBook 2400c for Apple at its IBM Japan division.

Fast forward to the 21st century. IBM could have begun selling Apple's hardware at any point since it divested its PC division in 2005. This is exactly why the company exited the market segment that it had created 24 years earlier: to get rid of a low-margin commodity hardware business and give it more flexibility to sell higher-margin integration services to large corporations.

The question is: what took them so long? The iPad came out in April 2010 and Steve Jobs has been gone for nearly three years. Ever since the Macintosh (1984) and particularly the LaserWriter (1985), Apple has been making products that would appeal to large companies, but lacked the sales, support and integration capabilities needed to address their customer’s complete requirements. (Only us Mac graybeards remember the 1988 Apple/DEC alliance that was intended to address these problems.)

Today, the two companies are not just looking over their shoulders at Microsoft, but also Google as well. The iPhone and iPad have already been widely adopted in big companies — spawning the IT acronym BYOD — but the new alliance should (like other successful downstream complementaries) generate incremental revenue growth for both parties.

However, there was one glaring omission in the latest Apple-IBM collaboration announcement: cloud computing. Apple has a retail presence with its true believers that is central to its integration strategies, but lacks the scale to compete with the industry leaders, Amazon and Google. IBM is their major competitor as a wholesale supplier, but (unlike Amazon and Google) does not compete with Apple’s retail offerings.

In the long run, Apple will unable to go it alone in cloud computing. We’ve all seen the risks that companies take relying on Amazon (cf. Netflix) or Google (cf. Samsung) as a supplier who is also a competitor. As in the PowerPC days, Apple should not only be leveraging IBM’s scale but working to attract others to its platform as the last honest broker in cloud computing.