Be the Cloud, Don't buy the Cloud

Two men named Smith boarded a plane in 1953. They had no idea that their chance meeting would result in redefining the airline industry and effectively, transportation for the human race.

As the story goes, Blair Smith sat next to C.R Smith, who was the President of American Airlines. Their conversation inevitably turned to the airline business and the inefficiency in a manual reservation system. By the end of 1965, their brainstorm, SABRE was handling 7,500 reservations an hour and items that used to take 90 minutes to process were being done in seconds. SABRE’s compelling capability was the ability to manage inventory in real time, and make it accessible to agents globally.

Eventually, deregulation forced airlines to turnover their reservation systems to independent companies. As of today, there are 3 major GDS (Global Distribution System) providers, who make this information (schedules, fares, seats, ticket records, etc) available to all who subscribe.

While Warren Buffett has famously observed that the airline industry in aggregate, has never turned a profit, the GDS providers have fared quite well. Before Sabre Holdings went private in 2007, they were reporting earnings of ~$200M/year.

Why the disparity in fortunes? Deregulation forced airlines to ‘buy’ the cloud, instead of ‘be’ the cloud. Thereby forcing them into a business of revenue per seat mile, instead of a data and information business.


A more modern example comes from Google, who made the prescient decision to 'be' the cloud for their data and information business. Their ability to organize and operate a vast network of servers and applications, with efficiency and speed, provides a level of differentiation that makes Google, well Google.

Their extensive investment enables them to handle 3B+ searches per day, index 20B+ we pages daily, offer free storage to 425M Gmail users, etc. More importantly, it enables them to customize their infrastructure to their unique needs. On example is their unique approach to cooling, which is estimated to reduce their electricity loss by ~15%. That generates real savings to be reinvested. Note: Google PUE (power usage effectiveness) is a best in class 1.2 (2.0, meaning half your power is wasted, is widely considered to be a good mark).

Google then takes the innovation one step further, to drive differentiation via workload management. They are able to instantly determine which machine can most efficiently process any given workload, optimizing the overall investment. Said another way, the data center acts as one giant supercomputer.
Google, realizing what business they are in (data and information), made the decision to 'be' the cloud. This is despite the fact that any of 20 possible hosting/cloud providers would have underbid the deal to win Google and it would have appeared very inexpensive in the short term.

However, what appears inexpensive in the short run by 'buying' the cloud, is often lost in the long term by not 'being' the cloud. Lest we forget the lesson learned from the airline industry.


Be the Cloud or Buy the Cloud. I believe this decision starts with the question often espoused by Clayton Christensen, "What business am I in?" If you are in the data or information business and/or data is the key enabler of your business, then I believe you have to ‘be’ the cloud. While every company relies on data, there is a big difference between those that rely on it and those that may cease to exist without it.

Note: I found alot of good information to help with the examples above. A particularly insightful one came from Wired.

1 comment:

  1. Agree Rob... Cloud's a journey not a destination.