Mission critical open source and referenceability

A friend emailed me today to critique an earlier post on Gartner's apparent curious underhyping of open source and the subsequent discussion on Open Season. He made some good points, but I admit that I was surprised to see this:

You guys [MuleSource and Alfresco] talked a lot about who is buying your software, and how those customer names demonstrate that you're enterprise-ready. You are leaving out a serious piece of data and you're not reporting honestly until you include it. What are those customers doing with your software?

If [a large financial services company] is managing its trading systems with Mule, then...you're in the club and you ought to be saying so. If they're running the server that manages the employee complaint drop box in their Indian call center, then shut up. [Big vendors like Microsoft, Oracle, SAP, and IBM are] not the monster because we sell to [big companies] -- we're the monster because without [us], [these big companies] would stop working.

This is a good point, but it overlooks an obvious point:

It may not be the lack of mission-critical deployments that stops us from talking about them, but the very fact that there are mission-critical deployments. In other words, it may be easy for an SAP to convince a customer to go public about its IT choices, but it's very difficult if you're a small company, especially with the FUD that the big companies spread about open source.

This may well be proprietary vendors' most cunning tactic: spread FUD about open source such that the myriad of enterprises using it - for both non-critical and mission-critical applications - figure it's not worth the risk to go public with their technology choices. Referenceability, perhaps more than anything else, is slowing open source's advance.

If I could go public and tell the planet that Fortune 100 (Financial Services, Retail, etc.) companies are standardizing on Alfresco, in some cases ripping out multi-million dollar mission-critical applications to replace them with our software, you can be certain I would. If I could tell the world that SugarCRM is rapidly closing monster deals for company-critical, enterprise-wide CRM applications with Fortune 100 companies, of course I would. If I could tell you that there are some astounding applications being deployed on MuleSource's open-source ESB, I wouldn't hesitate.

And yet each of the above is 100% true. I just am not allowed to associate company names with the deployments. And so Financial Services Company X that is looking to replace its proprietary bloatware will take that much longer because it won't realize that its peers are already making that move.

Open source may start within enterprises at the departmental, non-mission critical level. But if it's any good, it doesn't stay there long.

Gartner hasn't been helping in this. As my friend notes:

The analysis that Gartner does is the one that its big clients want to see. That doesn't mean that they cook the books, but that they look at the class of deployments that drive the biggest revenue for their clients, and see how significantly those have been eroded by open source. Singling MySQL out in particular, that impact is not yet causing Oracle and IBM serious pain. So when you look at the placement of MySQL on the hype cycle, it reflects that fact.

Rather than focusing on the big vendors' pain, the analysts would do well to focus on the benefits that open source is delivering to an increasing array of real customers solving real-world, mission-critical problems. This would give a more accurate view of where we're at with open source.

A long way to go, but with massive momentum behind us.

Disclosure: I work for Alfresco and am an advisor to SugarCRM and MuleSource.