Centralized vs. Decentralized

[Michael Fienen]

I just wanted to start a discussion that I sort of kicked off on Twitter this morning. There's no real motivation behind it, just a good topic. It's interesting to see how others run websites, centralized or decentralized, and why. I generally feel that we decentralize as a solution to small web staffs, but we're turning over sites to people that are clueless in how to run them. It's the wrong solution to a problem that is made to look worse than it is. CMSs are supposed to "make it easy," but they don't fix bad habits and lack of knowledge. Using a CMS doesn't mean you know how to work on the web. The result, in my opinion, is an environment that is worse than just putting the extra load on a web office. And I say that as an Army of One already.

I'm a supporter of the centralized model, or the moderately decentralized setup for bigger institutions (site might be broken down by school, but still not handed off to the lowest levels). This is especially true with a CMS where for us, the people that know the web and the system, it's often times easier and faster for us to do something rather than help someone else do it. I'm also not convinced that departments *must* have a hand in the growth of their site. We can easily help them craft identity and sell themselves without handing the keys of the kingdom over to them. They may want control, but do they really know what to do with that power once they have it? Of course not.

I know the theory that you want to let the "content experts" write their stuff, but even as we say it, it's often not the case. The office secretary is not the "content expert." And being a "content expert" doesn't mean that you know how to either translate or format it for the web. Really, how many people work on the web in your decentralized environment that you *really* think get it (as a percentage of the total number of editors)? Does that make it worth it? The content experts should provide the web experts with the bullet points that need to get listed, nothing more. That allows much better long term benefits: consistency, quality, findability.

Discuss. I might take some of this and expand upon it on .eduGuru if people would like to hear more.

[Tim Nekritz]

We went to a CMS during our 2003-04 redesign, which includes a version of organized-chaos decentralization.

Good: Our content improved immensely, as a team of cross-campus content authors I chaired created 100+ new sites. We became audience-focused, much more toward prospective students than ever before. Consistency of look and organization were a huge improvement (old site had many pages that didn't even mention the college by name and one that used purple script font on a lilac background, unintentionally plumbing new depths of unreadability).

Bad: We never really told them what made for good content. We bought them a fish (a CMS, a bit of training) instead of teaching them to fish. Updates often were handed down to already-overworked department secretaries. Content became stale, and we're not really in a position (currently) to step in and help, only try to urge them to keep updated. Some become frustrated they "want to do more" though when we ask what "more" is, no clear answers emerge.

Robert Brosnan of Seton Hall made a great comment at Stamats SIM Tech: He assures better content by educating his contributors. This includes such things as basic SEO, usability, etc. ... an awesome idea. I really want to implement that here in some ways as we proceed through a redesign.

[Jay Collier]

Why does the question have to be either/or? Can't content gathering and updating be decentralized, with professional editors in queue for quality control, either before or after publishing?

[Michael Fienen]

Sure, but that to me is by and large still centralized, because the changes happen via a core group of trained web editors then. I sort of mentioned that in the first post, where the "content experts" could pass on the core knowledge, and have it actually crafted and written by someone that knows how to do that best for the web. I'm personally not a fan of the "we'll let them write it and then we'll clean it up later," because that's something that sounds good, but I've never seen executed with any consistency or success.

It's like, I presented the issue sort of black and white. There is a lot of gray, but I find it to be either dark gray or light gray, not much gray-gray, so usually the strategy leans specifically towards one pole, even if it may have some aspects of the other.

[Jay Collier]

I think you've hit the nail on the head.

"I'm personally not a fan of the 'we'll let them write it and then we'll clean it up later,' because that's something that sounds good, but I've never seen executed with any consistency or success."

With some sites, more DIY is possible (decentralized), with others DFY is the model (centralized). Both coexist in our content ecosystems.

The better you recruit and train content developers, the higher the quality of the submitted material. Over time, writing for online media becomes more important in the hiring and selection of people who are working at the department level.

Check out Dartmouth's integrated model. Many departments, many editors, various levels of support.

As to staffing models, here's a recommendation for central staffing skills.

[Luke Wertz]

Something that I think (hope) will contribute to the solution is documentation. We launched a custom built CMS, gave the "power to the people" and, for the most part, it's turned out okay. I would consider our system to be fairly decentralized, and as we've onboarded more and more users, we're finding that it is starting to fray at the ends, so we're circling back around and providing documentation, tutorials, and consultations for our CMS users. Don't get me wrong: it's not as easy as putting out a wiki and sitting back while our data generators consume our tutorials... there is a lot of hand-holding involved in the process. But, so far, it's been the only solution that we've found to work with any kind of consistency. Oh, and, we also have a professional content writer who gets a list of our top-trafficked pages for review/rewrite.

And, to answer your question of the percentage of users who "get it": 2%. One user out of ever 50 in our system develops consistently good content for the web without a ton of assistance from our web office.

[Charlie Triplett]

I'm actually writing about an aspect of this on my blog, something about the manic depressive cycle of departments and how that affects your web presence.

Decentralization with a CMS, as you point out, is a double edged sword. Giving users the ability to maintain content they're responsible for sounds like a great idea— but that assumes they're actually being responsible for their content in the first place, doesn't it?

I mean, think about it: Someone in a department who put everything off to the last minute in a print environment making the undergrad catalog designer's life miserable is going to do the same thing to you— likely on a more consistent basis rather than once a year.

Handing over the keys means that they have to act like responsible people in our sight, which means we're still responsible for sweeping their work and making sure they conform (such a dirty word to some people).

For me, decentralizing things has worked well because I have the luxury to be a bit of a bastard about it. If I sense that someone isn't "getting it" or doesn't really want to learn this, I don't let them use the CMS. They need to find someone else in their department or research center. Content contributors are like your employees. If you don't think it's going to work out, don't hire them.

Also, when I train people on how the system works, I do so with the expressed rule that I reserve the right to meddle in their website after they update things. If I do have to meddle, I send them an email explaining why I changed their content and how to do better next time.

I don't think that departments should be responsible for developing their own website. That should be done at the college/school level. A typical department doesn't have the competencies required to find/hire/pay a good designer. Plus, when you get down to it— regardless of the org chart— departments all have the same things that are important: a) faculty, b) their research (relevant to grad students) and c) degree programs. There's just so rarely a reason for a department to not conform.

The con with centralizing content management, however, is that you're suddenly wrapped up in people's work flow for getting things done. If they're looking for an excuse why something isn't done, that web guy is a handy scapegoat or distraction from the real problem. That's such a stab in the back that can jade a person.

The answer to either scenario is leadership— which is NOT the ability to just set up the amazing web system for people to use and then bitching about the "lusers" who use it.

You've gotta actually motivate those people like you would your employees and make it worth their while. And that takes some time. Not everyone is going to be talented at content creation. Not everyone is going to be responsible. Some may need to be fired.

Here's a thought:

What if we treated contributors like employees, with those kinds of expectations. What if we started doing nice things like... I dunno... Send them a nice email or card when they do a good job? Have a award ceremony for your contributors (the Dundees)! When was the last time I sent a nice email to one of my content contributors for a job well done (or just for not hassling me about things)? I'm gonna start to-day!

[Chris Nixon]

Thanks for starting this discussion Michael. At the U of Arkansas we are currently decentralized. We do use a CMS, and have a group that does meet regularly comprised of the college webmasters and a few that represent our larger systems like BlackBoard and PeopleSoft.

Currently I am putting together a proposal to centralize the web staff under me. After talking with a few peers who have done this to great success, and after reading a good report titled "Investigation into the management of web content in Higher Education Institutions" written by the Social Issues Research Centre out of Oxford UK, it seems like this is a neccessary evolution to our web support in higher ed.

I think essentially the problem has stemmed from staff being added to address the web over time in an adhoc and unthoughtful way instead of looking at a holistic approach for the institution itself. When the idea that the homepage is not the marketing piece, but rather every department page on campus, it seems clear that this is not a department or college problem, but a University problem. Needing to be addressed from the top down.

Of course the CMS has been very helpful to make consistent code and design, but as mentioned here the quality of content has improved very little. The root there for us is that content on the web is just now being looked at as important by the individuals with the power to do anything about it. I think its unfortunate that discussions about the web, for us, keep coming back to things like 'what can you build' and "can you make this do that", instead of, "how can we create more engaging content that people actually want to read."

My hope is that with centralization comes a well rounded team that can focus on the core arts of the web, and can then branch out via the class room to empower the editors to write better content as someone also mentioned here. SAdly though the reality of centralization seems too far off in the distance. Politics, reporting lines, HR, etc.. are all big obsticles to this effort of success.

As a stop gap we are developing a certification process that will be supported by HR with the intent that we can put a carrot out there for people that want to learn how to do their job more effectively when it comes to the web. We hope to cover everything from content to programming.

Wish us luck?

[Chris Wiegman]

We're so decentralized that moving in the opposite direction would be a major hurdle as many units, including my own department, employ a web person or staff of their own. As such there are a number of us that run our own CMS systems and other systems on a level far below what most schools would see.

There are some benefits to this model, when people do get together there is a large amount of knowledge available for sharing. In addition, units with the know how can provide a level of service that would not otherwise be available.

On the flip side, for an incoming student browsing from one department to another is nearly impossible and in fact when doing so there are complaints that users think they have left the university all together.

How will they fix this? There are always a few ideas but never any sold answer for implementation. For example, a couple of years ago campus mandated a banner be put at the top of all pages yet provided little in the way of help in doing so. Currently, our campus has a committee to determine what constitutes a web page/site and general requirements for such sites. Even with such a group however there is no real method for enforcement as those who don't like the requirements will simply move to their own hosting thereby exaggerating the problem.

My question at this point is: are we an extreme example of decentralization? What have other school done to overcome these types of problems?

[Chris Nixon]

IMHO opinion you have to start with leadership from the very top, president or chancellor type. Step two is to put some power into a policy and make someone responsible for enforcing the policy. Also give them the tools and authority to enforce. Then you work with your group to develop the guidelines, not everyone has to agree but it helps to get buy in. Then you put into practice. Those that don't adhere get turned off, blocked, or integrated against their will.

I think one of the most important and often had discussions is "it's not YOUR website, it's the Universities, follow the rules or get out of the way"