Paradigm shifts and bullshit
There's a lot of bullshit written about the cloud.
When something gets called a Paradigm Shift, it's a very good sign someone is trying to sell you something. Cloud is a good example - a marketing man's dream, but a customer's nightmare. Do you have a cloud strategy? What are you competitors doing with cloud? Is it all a <cue obligatory hot air joke>.
But cutting through the techno-babble, what does Cloud computing mean for your old-fashioned equity investor?
What is Cloud Computing? A recap
This summer in London, retro is cool. What with the Queen's Golden Jubilee (with obligatory retro-50s styling) and Bradley Wiggins (sideburns and all) sweeping all before him at the Olympics and the Tour de France, old school is in.
Cloud is just the latest example of this. Here's a potted history:
- Mainframe: In the early days of computing, all of your computing power resided in the basement in the mainframe. All you had at your desk was a "green screen" terminal which acted as a dumb portal. Connecting the two was a very low-bandwidth network connection. Note at this stage all the money was spent in the basement - millions of dollars on the mainframe, thousands on the terminal.
- ClientServer: This was fine in the age of green screen, however over time as computers got more powerful this could no longer keep up. You couldn't stream a modern graphically-intensive user interface from the mainframe to a dumb terminal. The LAN (local area network) simply couldn't keep up. This meant that more and more computing was done at the desktop - the move to client-server architecture. The big box in the basement went from becoming the brains of the operation (in the mainframe days) to being just a dumb file server. Note at this stage the money shifted from being spent in the basement to being spent on the terminal.
- Cloud: However step forward another decade and the network connection is no longer a bottleneck. Gigabit ethernet in the office and fiber broadband at home makes it possible to stream even graphically intensive content like movies and video games down the network. In addition client-server architectures are grossly inefficient - look around a modern open plan office and you will see dozens of multi-core Intel desktop with endless MIPs of power running basic word-processors and web browsers which use only a fraction of their capacity. It meant it was possible to move compute power back into the servers in the basement (or even better, in a remote Icelandic data center), and stream it to a dumb desktop. Just like old times. As I said, retro is back in this year. And note at this stage the spending shifts again from being spent at the terminal back to being spent in the basement.
And that, in a nutshell, is Cloud Computing:
Translating cloud bullshit to business models
One of the toughest parts of thematic investing is translating a big hairy intellectually impressive theme, and quantifying a) how it changes the business models of companies and then b) which companies will benefit from this change (and whose shares we should therefore buy).
There are three impacts of cloud computing which affect investors:
1) For customers it increases Return on Invested Capital: Forgive the Business School 101 here, but its worth remembering the primary job of a business is not simply to make money, it is to make more money than you put into it. If you buy an ice-cream van with a loan which costs you a $600 a month, but you only make $300 profit selling ice-creams, you are making an operating profit but its being chewed up by your financing costs. You are not producing a return that beats your cost of capital (the $600 /month loan).
That is how the EVA (Enterprise-Valued-Added) methodology of valuing companies works. It figures out how much capital is invested in a business (factories, machinery, land, upfront payments, inventory etc), and works out how that capital costs. It them looks at the post-tax returns produced by the business and asks - am I making more money from my business than it costs to run the business?
This is a remarkably sane way to value a company which isn't used much by equity investors because they prefer the alternate DCF (Discounted Cashflow) method which is a whole other debate entirely. Let's just say DCF was probably engineered by the same crowd who built the Tacoma Narrows Bridge.
Anyhow back to the cloud. When looking at a company on an EVA basis there are two ways to boost your Return on Invested Capital (ROIC). 1) You can make more profit. 2) You can deploy less capital. Cloud computing lets you achieve the second objective by converting upfront capital expenditure (buying computers and data-centres) into operating expenditure (instead of buying them, renting the data centres on a day-to-day basis). Your upfront capital costs are reduced and you make efficiency savings as you only buy what you need. Effectively this is old-school business outsourcing in a frock.
It also gives you a lot more flexibility as a business not only to use what you need, but to have it when you need it. A good example is the Guardian newspaper. They were about to get a dump of pdfs of UK parliamentary expenses which they needed to parse through very quickly (the rival Daily Telegraph had possessed the same information for months). In order to do this they were going to dump them online and crowdsource the parsing to members of the public. However to host this they needed a bunch of servers which would likely only be used for a few days. The choice was to spending $10k on new servers which would be unused afterwards, or to rent the space from Amazon's EC2 cloud computing service for $100. Guess what they did...
So for businesses who are smart in their use of cloud, they can boost returns on invested capital and the valuation of their businesses. In theory good for the equity investor who buys their stock. However there is a sting in the tail, which is that any out-performance is likely to be transitory. As a businesses competitors all adopt similar cloud methodologies any excess profits generated by this new more efficient way of doing business are likely to be eroded away by competition. So yes a higher ROIC, but only for a year or two which, to be honest, doesn't make that much difference to your average EVA valuation.
2) For vendors it gives them something new to sell: Actually most of the benefit created by the Cloud Computing theme goes to vendors of cloud software and services, rather than customers (heh, something that brings to mind Schwed's famous book Where are the Customers' Yachts?). This relates to what I wrote in the very first post in this blog, which is that the Tech sector is so special because the inexorable rise in compute power means that vendors always have something new to sell. This gives them structural pricing power and high volume-growth markets.
This is the most easily investable way to play Cloud Computing - buy the stocks exposed to these growth trends. Cloud poster-boys like Salesforce.com, Rackspace or even the UK's dear own little Telecity have down very well over the last few years on the back of this. It's easier for a share to go up, after all, when its customers are buying more. Especially when everyone else is seeing demand drop due to the recession.
3) For vendors it disrupts spending patterns: However there is also a dark side to this, which is that there will also be losers. Note how different computing architectures (mainframe, client-server, cloud) have changed where in an organisation IT spending goes to. In the mainframe all the dollars went into the basement. In the client-server era they were spent on the desktop, buying mainframes and paying for expensive systems-integrators like Accenture to make them all work together. In the cloud computing world the spending is going back into the basement.
Also in the world of cloud computing, IT spending gets concentrated in the hands of fewer and fewer buyers (the cloud computing providers). This means that at best the guys who sell IT equipment to them have a tougher price negotiation and at worst they get disintermediated entirely.
I always remember 7-8 years ago when a friend working at Google pointed out to me they were probably a top-5 global PC manufacturer, even though had never appeared on any PC vendor's ranking - this was because even then Google had figured out it was more efficient to build their own servers from components than to buy from Dell or HP. Which isn't pleasant if you make lots of money selling servers.
But more on that next time. Disintermediation is great for a short-seller, but quite painful for the subject.