My my little debate with Gartner was utterly unintentional, and I blame it entirely on Google reader. Late at night it brought me this gem of a quote about ‘cloud computing’

With failure and lack of interest [in cloud computing] the fall from Inflated Expectations [will soon] begin its decent to the Trough of Disillusionment. It is at this low point that interest wanes as experiments and implementations fail to deliver. Producers of the technology shake out or fail. The only thing that will save the technology at this point is for those who have been successful to improve their products.

Remembering the last two months of their headlines I got upset. They’d been cloud washing magic quadrants, and market sizing and making a big splash in the news each time. Now they were going to sabotage the same phenomena, grabbing headlines again, by pronouncing pilots ripe to begin failing en-mass? At this point my life-long addiction to playing ice hockey got the best of me.

But as any hockey player knows dropping the gloves is transitory—you get the frustration out, see who is tougher in front of the crowd, then get back to the business at hand in an orderly fashion. I had a little scuffle with the NY Times earlier, but now I’m very friendly on Twitter with Jonathan Zittrain reading his book, and planning podcast here on the angle with him.

The Root Cause?

As I think about  my Gartner frustration with a little more perspective and fewer # signs I’m more concerned with the systematic reasons for the problem.

Above all the #1 (shoot used a #) cause of dissonance is jamming data into models not built with it in mind. Comparing API driven Amazon and custom contract driven IBM hosting as ‘managed hosting providers’ makes almost no sense –to anyone. Primarily because Amazon is in no way a managed hosting provider, and IBM is in no way (yet) a cloud hosting provider.

Then to muddy the rankings by having five additional underlying attributes to the conjoint:

In brief, Amazon scores moderately because this is a Web hosting MQ, not a pure cloud MQ. They can’t address the full needs of the market’s five major use cases. Gartner’s clients weight their selection heavily towards managed hosting, and the MQ’s weights reflect that; Amazon doesn’t do managed hosting, and although they score extremely highly on the self-managed hosting, that’s just one of the five use cases.  –Lydia Leong, Gartner

Oh, I see now…..so what I’m really looking at is not just apples and oranges rated by which tastes better—but five dimensions where they have almost no overlap. This whole thing seems like it was designed by a committee over months and months—oh wait it was:

The [Cloud] title was debated broadly and passionately within Gartner, and the final title was the result of a lot of input and a lot of compromises.—Lydia Leong, Gartner

Do I need to get into why committee and design by group compromise don’t create breakthrough products or ideas? No, no I don’t—and if you need help go read everything by Seth Godin. (def check out the vid above, suggested by Seth on the topic)

I’m very thankful to Lydia for joining the conversation, and I’d encourage you to check out her blog where you can always get her latest thoughts and links.

Hype Curve Modeling Force Feed: Cloud Computing as a Technology

One thing the obviously Gartner insider guests kept hammering me on in the comments yesterday was that every technology follows their hype-cycle so gosh darn it its just cloud computing’s turn.

All technologies, whether ultimately successful or not, follow the Hype Cycle[!]

Dude, hold the phone. What hard-core cloud guy is calling it a technology?  The big dog on the cloud computing blogging block, @jamesurquhart, cleared this issue up for us months and months ago. He wrote:

Cloud computing is an operations model, not a technology.

When you run an application in a public or private cloud, there is no “cloud layer” that your software must pass through in order to leverage the physical infrastructure available to it. In the vast majority of cases, there is probably some virtualization involved, but the existence of hypervisors clearly does not make your data center resources into a cloud. Nor is the fact that Amazon EC2 uses Xen hypervisors the reason that they are a cloud.

What makes a cloud a cloud is the fact that the physical resources involved are operated to deliver abstracted IT resources on-demand, at scale, and (almost always) in a multi-tenant environment. It is how you use the technologies involved. For the most part, cloud computing uses the same management tools, operating systems, middleware, databases, server platforms, network cabling, storage arrays, and so on, that we have come to know and love over the last several decades.

He’s also recently come around to the idea that its also an application development model—but notice, again, not a technology. Cloud computing is a mindset. Its the same technologies as ever, but with a zealots focus on scale, speed, elasticity, flexibly and connectivity.

They’ve got a yearly (yes that’s how precise they are being) report hammer and every trend they see is a nail to be mapped to it.

Everything broadly discussed has a hype-cycle, no doubt—but there is a pretty high standard of contribution within the cloud blogging community and calling as important of a moment as the peak in cloud interest with no precision isn’t up to par.

Why analyst firms focus on graphics and surveys

Jeremiah Owyang departing Forrester yesterday, presaged by Ray Wang’s departure brings quotes like these

He is generous with content (and blogs and Tweets) when analysts are told to hide behind their firewalls. He says it as he sees it when most analysts have turned mealy-mouthed.

[…]

We’re also seeing the emergence of the personal brand – Ray and Jeremiah have that in spades. It’s something they work ridiculously long hours at, taking huge risks along the way. Why would they give all of that up to the larger firm when there is so much more they could do in the wider world? It’s a non sequitor.

Graphics, surveys, headlines are the staple of large impersonal research shops for a simple reason: they are more durable than trying to retain top 1% human talent. They perform a vital function in the enterprise buying cycle—someone to blame if things go wrong. “Well we went with the magic quadrant leader,” and then Gartner can point to their committee driven 5 factor mechanism for scoring and say “we were thorough.”

But in the age of Twitter and brilliant real time blogs why should we get our key, fundamental, and attitude changing insights from mechanistic modalities?

As I want to make clear this isn’t an intellectual superiority jab at the top people at Gartner, its a swipe at their SOPs inflexibility and hype jet-wash. No doubt they have ‘bozo’s’ there like any organization, but there are smart people too, and the good ones are blogging more and more, just ask @mjasay

Redmonk has always provided great blog content. But I’m surprised by how great Gartner’s blogs can be

Exactly Matt. Its not the people its the process—like working at SGI in 2005 no matter how smart you may be you are shackled by the game plan. Everything possible is funneled into existing branded images and maps—period—don’t ask questions.

(Disclosure. Matt’s company is ranked in an MQ here).

What happens when things are forced into a model

Surprisingly lots of people wrote me about the X:Y observation regarding the Magic Quadrants. It was an aside I threw in just from my personal experience studying them over the years. They responded in comments on that issue:

The one thing I agree with is how low-quality Gartner magic quadrants tend to have X=Y. Some of them are very good, and very useful. But some of them were probably defined where the attributes of X overlap with the attributes of Y – and I’d call that a FAIL….. X will tend toward Y. But I agree with you – if a Gartner MQ shows X and Y totally intertwined with all the vendors, it ain’t useful and something’s wrong.

I personally detest the X=Y regression trend because it tells me the MQ is really just a top 10 list in the disguise of a conjoint. Without some radical deviation why put it into a conjoint analysis? All the conjoint does it make it look BCG sexy, like you have done a lot of thinking. Its superb branding.

Let’s not be sheep. Let’s demand these models really show us something interesting.

I’ll leave you with a few MQ’s with a hap-hazard regression line drawn in red. You decide if it resembles something like a linear relationship between X & Y.