On Open Source Analysis: Response to Armadgeddon

The folks at Armadgeddon have some justifiable questions about the Open Source Analysis model, which I believe I have answered before, but evidently not to their satisfaction. Before I proceed though I want to say a little bit about Gartner, a company which I increasingly respect. Sure there are things about the company which I am not a fan of, but that’s expected of any large corporate in any sector. I do think open source analysis and transparency will change its business, but the wheels aren’t going to fall off the firm because of broad structural changes brought about by the internet. They are doing the right things- such as hiring people like my good friend Thomas Otter, a blogger and enterprise software maven. I recently made my first referral to the firm. Am I allowed to admit to some childish rhetoric in the past? If so I hold my hand up now. Of course Gartner has had its own moments of weakness. All in all though – while I think RedMonk is an exciting, up and coming company that will hopefully change the business for the better, that doesn’t remove the need for incumbents, particularly ones that help people do their jobs, and cover their asses. Gartner’s credibility is a wonderful asset for both the firm and its customers, and I wish them well. Bear in mind that when I likened Gartner to a mainframe that wasn’t an insult. My career has been predicated on watching this wonderful platform evolve into something more powerful because its more open.

So here is the Q&A.

Q: Reputation/credibility lies with individual analysts or firms (and not with the “ecosystem”)
**An AR professional emailed us that the main issue with “one-man-boutiques” for them was the lack of branding/the confusion in the marketplace.

A: Reputation and credibility in Open Source doesn’t rest with the ecosystem so much as the distributions and individuals. Red Hat is a brand, a packager, and a trusted platform for deployment. Nobody just trusts open source because it is open source. They trust it because they have experience of a particular platform, they can introspect both the code and the community. Transparency certainly can engender more trust. Just because a firms labels itself open source doesn’t lead to trust. Indeed the industry is littered with examples of open source and mistrust. You’re judged by your actions, not a label. That said- I do work happily and naturally, notably with Freeform Dynamics, because it is somewhat aligned in its approaches and methods. Freeform does open source quantitative analysis, while RedMonk works on open source qualitative. We would like to think that we are a distribution of open industry knowledge, and a trusted one at that. To that end we always try to cite the sources of our ideas, rather than claiming them all as our inspiration. Because finally the community owns the trust, not the firm. Without our community RedMonk is literally nothing. Some pay us, some keep us honest, some make us more intelligent, some make direct contributions, others fix bugs. On the other hand bear in mind that MySQL is undoubtedly open source, but its not a shared development model. We don’t have to accept contributions from outside just because they are “open source”. That would be absurd. Our view is that their is an exuberant over-abundance of great analysis out there, the majority of which doesn’t even come from analysts. It is not our goal to be an aggregator of everyone’s analysis. That would be impossible. I know that the likes of the IIAR and James McGovern have recommended that we all go into one big M&A frenzy, and roll up all the UK independents to create a “credible” and “easy to work with challenger”. But in my humble opinion RedMonk is already credible and easy to work with. In the end brands create trust – that much certainly isn’t changing.

Q: The output of independent analysts varies considerably as it is not like a single source code (although the emergence of aggregators such as IT-Analysis.com is a major step forward). **As Richard Stiennon puts it in the comments, “open source research would imply a wikipedia like collaboration on market research, […] the results however would be FREE to use by anyone.”

A: The output of analysts does vary widely, just as the output of different developers does. That’s something any distribution manages. Indeed we’ve seen many communities have problems because people complained their code wasn’t being accepted. With reference to this chap Richard Stiennon, who claims we don’t understand open source, but has never spoken to us about it, I think he raises a point worth talking to. Let me say once again that MySQL is a model for us, and they do all the core development themselves. To claim we’re not open source because we’re not a wiki seems to me to show the lack of understanding of open source and how it works. Not everyone can “edit” an open source code base- that would be absurd because it would quickly lead to stability problems. Even wikipedia has a cadre of editors to ensure quality. Open source is not a free for all and never has been. Linus has been called a benevolent dictator for good reason. That said we have experimented with using wikis and Creative Commons licenses with limited success in our Compliance Oriented Architecture (COA). Forgive us for not executing on the opportunity as successfully as we might have. But, and this is a story for a different day, its precisely because the COA was free and not behind the firewall that I first began a constructive dialogue with SAP, which is now a valued client. COA was used to underpin discussions at SAP customer roundtables, and that was before they ever paid us a dime. I think that’s a fairly successful open source mode, where charging for services is decoupled from distributing the source code.

Q: Consulting to users remains the exception for independent analysts and it is not proven can the open source model can be financially viable for them. **Vendors value this user insight more than everything else, this is however currently the privilege of Gartner, Forrester, Ovum, etc….**

A: I would agree there are strong barriers to entry for independents in offering end user consulting but the question of financial viability is I believe a different question. The simple fact is that at RedMonk we have a great many enterprise contributors to our knowledge base. Architects and developers are our base, and many of them work at Fortune 500 firms. In some cases they ask us for help, in some cases they help to educate us as part of our ongoing research agendas. People seem to forget that when Gartner consultants work with end user customers, the analysts are also learning, and this knowledge transfer underpins their business. It is a wonderful privilege to be paid to learn from someone. My point is that the transfer of knowledge is bi-directional. At RedMonk we celebrate the fact, and offer free advice and help to those in our community that have made us smarter. They don’t pay us – but we can turn the insights into money in other ways, such as vendor consulting. Its a model I like.

One more question I hear a lot.

Q: How can users trust RedMonk when it would never bite the hand that feeds it (for that read major client IBM).

A: Firstly I would like to say that we feed ourselves. We’ve worked incredibly hard to get to where we are, over five years, and turned away a lot of money that we felt might impinge on our independence, from a broad range of sources. You only need to look out our published output to see that we regularly call out IBM for getting things wrong. Indeed some of the work I am most proud of is when those we critique adjust their strategies accordingly. If we do have a bias its towards open source, open standards, lower barriers to participation and community-based approaches; not towards those that pay us the most. IBM has also made huge contributions to Linux – does that mean Linus does what they tell him? They may buy some influence, its true. IBM gets the new patronage model pretty well, and increasingly works collaboratively with a range of constituencies, including analysts. I would argue though that its the collaboration, not the money, that really wins the influence. Yesterday someone from another vendor claimed that IBM were hypocrites in their work with analysts. Not in my experience. They just work closely with us, whether from big or small firms. Who sets the agenda though? The industry does, the market does – not IBM, not RedMonk, Not Gartner, not the CIO, not the grassroots developer. The agenda bubbles up, and bubbles down, and we all respond to that. It just so happens that RedMonk is particularly good at sniffing out new agendas before they become commonplace. We package these insights as part of our open source analysis model. That’s one reason our clients pay us.