Sunday, April 30, 2006

Sun Marketing, Innovation and Commoditization

I have a couple friends who work for Sun. One is well-placed up in the Sun food chain and the other has been heavily involved in Java standards for a number of years. Both are loyal to Sun.

Below is an edited snippet of email conversation between one of my friends and me regarding, at first, the departure of Scott McNealy as CEO, and then a discussion about the strategy of Sun as it has been unwinding over the past year or so. I asked him what he thought of Sun's strategy as it relates to cheaper servers. His response was, "The low-power server is a big deal, I think. If you are just buying 1 of them, you don't care. If you are going to buy 100 of them, you care a lot: the air handling costs really do add up. It is certainly a unique strategy to try to market a server (not just a chip) as low-power, but I think that we are doing so because we were listening to customers."

I'm including my response because it sums up my current analysis of Sun:

"but I think that we are doing so because we were listening to customers.."

That's an interesting dynamic, for three reasons:

1. Are Sun's customers listening to Sun?

2. In the book, The Innovator's Dilemma, the argument is made that innovators who listen to customer feedback can actually inhibit innovation because customers tend to only think in terms of incremental performance improvement with incremental cost reductions. Customers typically can only conceive of existing technologies once and then create demand that drives downward price pressure along with incremental performance/feature improvements. For an innovative company like Sun, listening to customers could (not saying it will) inhibit innovation. Sun needs to be careful not to allow themselves to be pulled into Dell's commodity server market because I don't think Sun has anywhere near the manufacturing efficiency that Dell does. That would mean loses for Sun. Sun cannot compete against Dell on the basis of price. Neither should Sun compete against Dell with the server as a commodity because that conflicts with Sun's innovation DNA.

3. Last year, when I was visiting my buddy Scott, we had dinner with a Sun exec. He made an interesting observation: he said that Sun has been an incredible innovator from the beginning but that Sun was terrible at marketing. He contrasted Sun with Microsoft, who he characterized as primarily a marketing machine and a technology innovator secondarily. He might have even scoffed at the notion of anything from Redmond as being innovative. I silently disagreed but listened to his point.

Bringing these three together might lead to the following idea: Sun may be taking marketing more seriously because they recognize that marketing is important to capturing more market share. I read Sun press releases and Schwartz' blog and both sources tell me that Sun is successful at large computing deals, particularly with developing nations and academics. Listening to customers is part of that but so is leading customers to better technology.

The links in Jonathan's blog tell a story of U-B's decision to go with Dell servers even though they were (if I remember correctly) Sun customers as well. U-B coulnd't afford to power and cool the Dell servers so they could only run a portion of them. What is interesting about this article is that the University of Buffalo didn't consider electricity consumption and cooling needs as part of the total decision-making process. I think this is in large part because most accounting types want ROI and TCO calculations that don't consider cooling and powering servers. Plus, as I wrote here: ROI and TCO calcs are accounting gibberish that frequently have little value in IT decision-making.

The question to ask here is Why? And I think the answer is because a lot of decision-makers aren't thinking of power and cooling issues because they view electricity and cooling as overhead expenses not direct costs. Overhead expenses are difficult to include in these kinds of calculations because they are... well, overhead and overhead is hard to tie to a specific point. You can get around it with an allocation of overhead but that's not entirely accurate. I think Jonathan refers to a per square foot cost for chilled server rooms but again, that allocation rate is based on assumptions about cooling and power needs, labor, cost of real estate, etc.

My point is that listening to customers is good and its good that Sun is taking customer input more seriously but what Sun really needs to accomplish is to sensitize decision-makers on the value of lower power consumption, particularly as you mentioned, with 100 servers or 1,000 servers. I think the mistake Sun marketing is making is by trying to pitch the new servers as eco-friendly, however. The only people who care about eco-friendliness are the bearded Birkenstock-wearing, Prius-driving IT guys.

My theory right now is that businesses see the server room as a commodity. They see value-add from IT purchases on the desktop in terms of what the dektop enables from a functionality and business process perspective. This is because software is on the desktop where users see the functionality. Once the philosophical and business decision of platform is made (Windows, Linux, OS-400, Solaris, etc.), it is easy to view the hardware of the server infrastructure as a commodity. This is due to two reasons: decision makers see only the desktop as evidence of IT, which makes the cold room invisible (except when a server goes down); and there is relative parity of price:performance ratios across server competitors and their products.

So, Sun is trying to compete on a selection criteria that isn't part of the thought process for a lot of organizations. Because of the emphasis on ROI in purchase decisions, IT guys need to justify up-front costs of server purchases because it is assumed that the back-end costs of server ownership are equal. The notion of eco-friendliness is not valuable to most people. Dollars, however, are, but from a marketing standpoint, it's a bit of a trick to get people to consider electricity and cooling factors, not only in terms of cost, but also in terms of capacity, i.e. do we have enough electricity and cooling to take these boxes live, because these costs are simply seen as overhead, not direct costs.

I think Sun has come up with a very cool innovation: more computing power, less electrical consumption, less heat generation. The challenge is to get existing and new customers to see it as a value-add feature/benefit set that puts Sun ahead of Dell instead of just something nice you do for the environment.