In my opinion, investors should mostly ignore the hype around “cloud” computing.
There are different definitions of so-called “cloud” computing. In a literal sense, cloud computing refers to computers attached to a network. Such technology has been around since the 1960s. The current interest in cloud computing largely has to do with the pervasiveness of fast Internet connections. “Cloud” software can be thought of as “software that requires a fast Internet connection”. The widespread adoption of broadband Internet allows certain forms of cloud technology to become more viable. For example, backing up large amounts of data over an Internet connection now makes sense.
However, all software companies are largely on the same footing when it comes to cloud computing. If cloud computing makes sense for a particular market segment, anybody can (re-)design their applications to take advantage of fast Internet connections. Software has always been an arms race between competitors improving their product with new features. Cloud computing is simply part of that arms race. Personally, I don’t see cloud computing as being a paradigm shift like the Internet was.
Defining cloud computing
One form of cloud computing is delivering software client via a web browser instead of a desktop application. This technology is not new. Microsoft’s Hotmail email product has a cloud-based user interface while the (old) Microsoft Outlook client is a traditional desktop application. Both products existed and were popular before “cloud” became a trend. The popularity of both Hotmail and Outlook demonstrate that both models can achieve high market share.
The ability to leverage an Internet connection is both an advantage and disadvantage. The Internet allows for updates to be rapidly propagated. The software developer can force all of the user base onto a very limited number of software versions. This makes customer support a lot easier as the software developer does not have to worry about what version the customer is currently using. The disadvantages of requiring an Internet connection are:
- The Internet connection itself causes performance limitations. All Internet connections have some “lag” or some delay between user interactions and when the user interface responds to the last user command.
- Some potential customers live in rural areas with limited or no Internet. They may have some form of Internet (satellite, wireless) that is very expensive and/or unreliable.
- When traveling on a cruise ship or airplane, a user may not necessarily have a fast Internet connection.
- The user’s Internet speed may have limited bandwidth. Over time, this disadvantage will become less and less of an issue due to Moore’s Law. Certain types of software will increasingly shift towards cloud computing as this relative disadvantage eventually becomes a non-issue.
For the most part, more and more software will shift towards using various forms of cloud computing.
New monetization models
Currently, there is a lot of hype surrounding SaaS or “software as a service”. The subscription pricing model has been around for a very long time. SaaS generally implicitly refers to subscription-based software that is also delivered via the Internet. One of the benefits of the subscription model is the ability for the software developer to force all users onto a very limited number of software versions. This is not possible if customers can opt not to pay for software upgrades.
The subscription model simplifies the marketing and cash flow issues of a software company. Under the traditional model, companies have to constantly get users excited about new versions of the software. The marketing department of a company may pressure the development team to build features that can easily be marketed. Software users may be unwilling to upgrade their software unless they understand that the new version has compelling features that they want. This can be detrimental to the quality of the company’s software if the developers spend too much time building easily-marketable features versus useful features that are difficult to market. For example, users dislike buggy software. However, advertising the improved stability of the new version implies that the old version was buggy and released too early.
The subscription model also has disadvantages depending on the anti-piracy measures implemented alongside it. Adobe faced a huge backlash when its authentication servers failed and its subscription users were prevented from using software that they paid for.
The prevalence of fast Internet also allows for other new business models such as “freemium” (free software where users can pay money for advantages or the premium version of the software), advertising-supported software, installers that try to foist spyware/toolbars/freemium software onto the end user, etc. etc.
Where cloud computing makes sense
One of the more compelling features of cloud computing is removing the need for the customer to configure and install software. Salesforce (NYSE:CRM) for example allows small businesses to enjoy the benefits of customer relationship management software without having to spend a lot of IT resources on managing and configuring a server, data backup, security, etc. etc. The IT burden on the customer is greatly reduced as Salesforce essentially handles those services for their customers. Salesforce can administer servers more efficiently due to scale.
For large enterprises, I do not believe that cloud computing is a major advantage for Salesforce customers. Most of the IT burden with an enterprise Salesforce deployment is in getting Salesforce to integrate with other software that the company uses. There is a large and growing industry of Salesforce IT consultants that fulfill this need.
Marketing and management hype
The marketing departments of many software companies have jumped aboard the cloud hypetrain and throw the word “cloud” around liberally. If it works, I wouldn’t blame the marketers for doing their job well. Their job is to generate hype and excitement over a company’s software to drive sales.
Many management teams will try to support the operating business by saying good things about their company’s products and services. They may echo what their marketing departments are saying about cloud computing. For publicly-traded companies, I don’t think that hype over cloud computing should be a reason to short a particular stock. If management is constantly talking about the “cloud”, they are not necessarily trying to mislead investors. They understand that the press pays attention to the company’s investor communications. Because of this, they may change their shareholder communications to try to support the underlying business.
The future is cloudy
As an investing theme, it’s obvious that faster Internet connections will allow software to get better and better. However, rapidly improving technology does not necessarily translate to shareholder performance. Industries like airlines, textiles, and computer hardware manufacturing have seen dramatic leaps in technology without generating wonderful returns for shareholders.
- It may be unclear as to whether or not a particular software niche will generate good returns for shareholders. For example, RSS readers may not be a profitable market given Google’s decision to exit the market despite the massive popularity of its Google Reader services.
- It may be unclear as to who the future winners and losers will be.
- It may be unclear as to whether or not a particular moat is durable. For example, ICQ and Myspace seemed to exhibit strong network effects until they didn’t.
- It may be unclear as to how long the cash flows for a particular product will last. The history of software is littered with the corpses of products that previously enjoyed massive market share.
In general, I think that the future of most technology companies are highly uncertain. While I have started my own software business, I am definitely not smart enough to predict the future of most software companies. I would not have predicted that a company making BASIC interpreters for the Altair computer would later become Microsoft. Nowadays, nobody makes Altair computers so the market for Altair software is quite dead.
Microsoft’s (ex-)CEO Bill Gates even write a book (The Road Ahead) making predictions about the future of computers. This book is filled with predictions that did not come true. This happened despite the author’s deep understanding of computer science (Bill Gates bothered to read Donald Knuth’s tomes on computer science) and his experience as the CEO of a major and highly successful software company. If Bill Gates (arguably the world’s best software CEO) isn’t very good at predicting the future of computers, I doubt that I could do a better job than him. I can see why Warren Buffett likes to invest in companies that face little risk of technological disruption (e.g. food brands, railroads, etc.). In any case, I believe that technology companies are generally very difficult to predict regardless of your understanding of cloud computing or the underlying technology.