Rolling your own CMS just doesn't make sense

Here at FierceContentManagement, we write about content management systems all the live long day, so when someone suggests all open source and commercial content management systems are basically useless, it makes us take notice. That's precisely what Clay Johnson [1] of the Sunlight Foundation [2] did when he posted a blog entry last week called Content Management Systems Just Don't Work [3]. It's a bold statement, but one which Johnson makes based on his many years of experience working with technology and he is no slouch, let me tell you.

He was one of the founding members of Blue State Digital [4] , the consulting firm born out of Howard Dean's 2004 presidential campaign, the first campaign to make use of the Internet as a successful fund raising vehicle (and which later gained fame as the firm that designed the immensely successful Internet strategy for candidate Barack Obama). At the Sunlight Foundation he helps develop technologies with the goal of making government more transparent, all good as far as I can tell.

No real agenda, just experience

Johnson doesn't seem to have a vendetta or an agenda when he says CMSs don't work. He's worked with them and he's talked to a lot of people who work with them and he says there is a lot of frustration around using them. His blog entry, in fact, was a reaction to the White House's decision to use Drupal [5] to power the Recovery.gov [6] web site, a decision he believes that will line the pockets of Drupal consultants, rather than lead to more government transparency. His answer to all of these perceived CMS negatives? Roll your own using a development framework like Ruby On Rails [7] or Django [8]. While I can understand his point of view--software can be frustrating and difficult to implement, especially a CMS--rolling your own CMS has its own set of issues and frankly I think, in spite of the natural limitations you will find with any out-of-the-box software, creating your own system is not the answer.

Too many opinions

Johnson says part of the problem with an out-of-the-box CMS is that they are born of too many opinions, which were thought up by some developer outside of your organization and that naturally these decisions aren't the ones you would necessarily make. It's an entirely logical argument, especially coming from a programmer, but the fact is that as anyone who has worked on implementing enterprise software can tell you, everyone has an opinion about what it should look like. Just last week I wrote in this space [9] about the on-going tension that exists between business units and IT (and there are a range of opinions within those entities). If you are fortunate enough to get executive sponsorship for your project, you might have someone driving the technology vision (or at least building consensus among the many constituencies affected by enterprise software), but you can't escape those opinions.

It seems that what Johnson is looking for is a benevolent dictator-programmer who can drive the vision and make the decisions he perceives to be correct, but just as he points out, some of the decisions about registration, or workflow, or the administrative interface can make users crazy, it doesn't mean that because it was created from scratch that you are any less likely to face these issues. Software is software. It doesn't matter if Clay Johnson created it or a vendor. The only difference I can see is that Johnson is happy with the decisions he makes, which is fine of course, but it doesn't mean that his users won't have similar complaints about the end result.

You're not in the CMS business and you probably shouldn't be

Johnson states that he comes from the point of view of organizations like governments and non-profits, but even when we are talking about private enterprise his fundamental beliefs would seem to apply, and that is that needs change over time and a framework is more flexible to address those changes. It's a solid argument, but still for any organization to build its own system in this fashion takes even more leadership and consensus building than identifying an existing package that meets the essential needs of the organization.

I understand that CMSs can be frustrating, that it's difficult to suit the needs of everyone, that there will be components and pieces that don't work as you wish and convoluted workflows, but building your own requires your organization to be in the content management software business and that makes even less sense to me, especially when there are so many choices out there that already exist. - Ron [10]