Friday, January 20, 2006


Thoughts on Industry CTOs

Yesterday, I had the opportunity to talk with an industry analyst for a large analyst firm regarding ServiceMix and some code I am planning on contributing. The conversation was frustrating and enlightening at the same time. It was frustrating in that I learned why analyst firms are struggling with telling the real story behind open source...

In knowing that many of my peers run around the halls of corporate America with analyst reports as if they provide insight tells me that if analysts are telling the right stories, that many enterprises are doomed to be mediocre when it comes to understanding the real value proposition behind open source. When it comes to open source, many of the analyst firms still think in terms of vendors and not products. The lack of a vendor makes it difficult for them to capture marketshare, brings the matrix and even makes it difficult to rewrite portions of information borrowed from the web site.

One example is that I cite often is the wonderful folks over at Duke Energy. For those interested in .NET development, they actively participate in the community and have contributed a rock solid application framework. How come zero point zero analyst firms have done a case study on them? In fact, how come zero point zero analyst firms have not even covered any Fortune enterprise that has created their own project? There are several out there and the world should know about them.

I guess if you look at the typical analyst report format, it would be interesting to see how much power generating capability they have. I wonder if analyst firms were to actually consider Duke Energy and list their revenue figures that the .NET framework they have made open source would then be viewed as a market leader?

One of the things that absolutely pushed a button was the perception that commercial products automatically have more users. The analyst quoted the number of customers each enterprise service bus vendor had. He then followed up with a statement that suggested that ServiceMix couldn't possibly have as many customers. Oh really?

Should open source projects start keeping track of who uses them? Should industry analysts instead of simply relying on vendor provided numbers do their own homework to figure out usage? Should industry analysts participate in the open source community themselves by joining listservs to see first-hand what the discussion is all about? Should industry analysts actually attempt to install the software themselves and build small proofs of concepts and not rely on Powerpoint for all their information?

For those that don't know, ServiceMix is a highly scalable Enterprise Service Bus that has more features and functionality than any commercial ESB on the planet! From my own participation and observation in this particular community, I am of the believe that it has about 800 viable customers which is more than any of the commercial offerings in this space today.

Another open source project worthy of attention in the enterprise is Liferay which has been certified by several third parties to be the most secure portal platform in existence and is the only one certified to run on 384 CPUs. The rest simply run out of steam if you attempted to throw say 10,000 concurrent users within a single instance.

If analysts want to count paid customers, then I can honestly say that both Liferay and ServiceMix will shortly have over 5,000 paying customers. I have asked everyone in my second home of Biche Trinidad to pay exactly $1 TT to each of these projects. Of course, knowing both of the leads of these two projects they will more than likely donate all proceeds to a worthy charity...

It is simply a crime for enterprises to blindly follow research reports when they don't contain the entire picture. Enterprises need to demand of every analyst they talk with that both commercial and non-commercial open source projects are listed next to proprietary closed source vendor offerings. Putting them into separate reports is also unacceptable. I wonder if Stephen O'Grady of the high quality industry analyst firm Redmonk wouldn't mind putting together a list of ten things that enterprises should consider when purchasing analyst research? The continuing crime spree committed by most analyst firms needs to be halted immediately.

Over the past several weeks, I have noticed a pattern in that multiple industry CTOs had posted in their blog disappointment whenever they have participated in an enterprise bakeoff. While not having any insight into the reasons they lost I do wonder if they have actually figured out what their real job is? So many industry CTOs feel it is their job to ensure conceptual integrity to the product line. While this is important, there are tons of other things that are more important.

If you understand that folks within the enterprise obtain their "wisdom" through one of two channels then you would realize what your next action items are. If you know that folks in the enterprise achieve information via the industry analyst channel and you also know that analyst firms aren't of the right mindset and you are taking steps on a daily basis to change their communications, well...

Likewise, if you understand that google is also a research tool and that by engaging in dialogs with the community that also happen to get indexed by a variety of search engines, you may be hyperlinked and come up higher than those who don't. Maybe you should spend time reading tomes such as the Cluetrain Manifesto. After all, hyperlinks subvert hierarchy...

Maybe you may conclude that the industry analyst model is fundamentally busted. Maybe you may noodle ideas on how not to sit on the sideline but instead embrace the notion of Open Source Analysis. If the model is busted for you then you should rightfully assume it is also busted for your customers.

There are lots of analysts with integrity. I consider Jon Udell of Infoworld one of them. He too at times needs a course correction in that he is equally guilty of not testing non-commercial open source projects side-by-side closed-source proprietary product offerings in the Infoworld Test labs. Hopefully he will fix this in upcoming issues.

Maybe what customers really need is for vendors to figure out how to work with competitors so that real industry reference architectures emerge. I know that many of my peers would be filled with tears of joy if they could see one in the rules engine, BPM, ESB, and Identity Mangement space?

Maybe if analysts simply provided answers that customers ask in their blogs. I wonder how many analysts or even CTOs can provide insight into Enterprise Content Management or even thoughts on BPEL and ESB and what folks should be really considering? Maybe analysts and CTOs don't really care. Maybe they are comfortable with mediocrity...

<< Home
| | View blog reactions

This page is powered by Blogger. Isn't yours?