Sunday, December 29, 2013

Thoughts on Vertical Integration: The Cloud and (Gasp!) Ma Bell

Although it has been a couple of months since the hysteria of re:Invent, there has been a great deal of consternation by hoards of people in enterprise IT over the emergence of Amazon AWS. The fear is that AWS and it's main competitors will substantially disrupt the market for enterprise infrastructure and create the next big oligopoly or even monopoly in the technology industry: The Cloud. In this weekend's Barrons, we find Tiernan Ray making the near term case for mediocrity, if not eventual doom for the big enterprise technology players (Note: subscription may be required):

No Silver Lining in the Cloud for Blue-Chip Techs

The key concern among investors - both venture and public market -  and users everywhere really centers around the notion of AWS being a "dominant exchange". This is what happens when the game becomes one where the entire solution set is owned by a single player. In this scenario, scale matters most, and customers can no longer see an incentive to go anywhere else for their needs. 

Well, rather than try an imagine a bleak and scary future where all IT is dominated by a large evil corporate entity, it turns out we can put our fantasies aside and just open a few history books. If you are an American of a certain age, or even a current or former resident of select countries, you have probably lived in a world that was eerily similar to what everyone is afraid of. In the twentieth century, the paradigm of technological vertical integration has to be the old, pre-breakup, Bell System - also known as American Telephone & Telegraph. It turns out that looking at AT&T's business model and history is hugely instructive for those who are planning strategies in the Everything-As-A-Service world of the early 21st century.

Voice-As-A-Service circa 1900

If it weren't for some bad behavior and a bunch of concomitant anti-trust lawsuits, the story of AT&T would be an example of a great American capitalist success. Famously, Alexander Graham Bell invented the telephone in the late 19th century and set out to commercialize his discovery with his eponymous company. In order to deliver the service, the Bell System became an end-to-end solution for making voice calls to other Bell customers. This meant that they provided the telephones to their customers, they strung the wires on the streets, built and maintained central offices where operators would help patch point-to-point connections for customer calls, manufactured all of the equipment that the operators and customers used, and serviced all parts of the system in case of trouble.

To make the story interesting, the company was an aspiring monopolist. They refused to interconnect with other telephone providers unless they sold out to Bell, and they set the standard for how interconnection would work when they owned such a subsidiary. This got them into trouble when they bought out their biggest competitor, Western Union, and attracted the attention of the anti-trust authorities of the pre-World War I federal government. Eventually, AT&T was granted a legal monopoly and became a regulated entity. What is fascinating to consider, however, is how they operated within those constraints for 60 years. While the world has changed a lot in the last 100 years, the recent discourse around the cloud has had me making a comparison: Is The Cloud - or AWS, in particular - destined to be the new Bell System?

The tradeoffs of this extreme level of vertical integration are particularly interesting. As a result of owning everything, The Bell System all actually worked. The phone system, the switch gear, the telephones, and everything along the way was overbuilt and extremely reliable. One struggles to remember a time in the 1970s when their phone service was unreliable or when they needed to troubleshoot a telephone problem. On the other hand, the pace of innovation was stiflingly slow. If it weren't for telecom deregulation in the 80s and 90s, we would not have meaningfully fast data network access available today. We would not have VoIP. Long distance calling would be inexplicably expensive. Most importantly for most of us, the economy would be bereft of trillions of dollars in market value that has been derived from IP networking.

When one considers the direction of the products and technology in the 80's, it is obvious that AT&T did not perish as a result of anti-trust action, but rather from obsolescence, just like every other tech company. They were so busy attending to that enormous customer base, that they could not address the needs of their more innovative customers. In fact, any enterprises that had advanced telecom needs had already found ways to work around AT&T long before it was broken up. Media companies were sending transmissions via satellite links, and large enterprises were buying private switching equipment from a slew of competing companies that catered more to their needs than the one-size-fits-all approach offered by "the phone company".  If it weren't for their status as a legal monopoly, they would have been disrupted much sooner.

Is IT-As-A-Service Destined to Remake Enterprise IT?

Looking at the Cloud, the key innovation that the established enterprise players are finding disruptive is really a business model. Turning a product into a service in a way that makes economic sense is a ground-up endeavor. Setting that aside, this generation of technologies built for The Cloud are just as easily disrupted as previous generations. Whether or not the establishment players will be able to hang on to their customers hinges on the same factors that have always been important: Will the incumbents be able to adjust to the evolving needs of IT, or will someone new take the business away? Most importantly, do AWS and the other service vendors provide any meaningfully new technologies that need to be consumed as a service rather than a product? It is still far to early to tell. Certainly, it is not yet time to sell the incumbents short. The next few years will be interesting, though.

No comments:

Post a Comment