SAC4: University-Wide Web CMS Implementation - Failure, Then Success

[Intro Music]

Announcer: You’re listening to one in a series of presentations from the 2008 HighEdWeb Conference in Springfield, Missouri.

James Buratti: So we’re from Texas State University. We have about 29,000 students this semester so we’re a large university compared to some in Texas that only puts us top the fifth largest. So we’re kind of walk you through our journey on the different CMS’s we view. Some of the trials and tribulations, all that take home stuff, all those do’s and don’ts that’s in your notes.

So two weeks from now you’re like what did he say, “Pull out in that little folder?” That’s the notebook that they gave you it’s all there. What you’re going to see today is kind of more the time line human side of that politics, bad decisions that were made, good decisions that were made, and see kind of a timeline flows. You can see how we got to where we are going. When we started thinking about this presentation, where we’re at now is a very good happy place and so we look back and it’s like, “I don’t want to look back because I’m remembering a bad dream.”

It’s like well, we want other people to avoid that pain and so we were trying to think of what we can do, what we can present here so that we can make something good out of something bad so I’m going to start with a little analogy. Reminds me of a story of a gentleman who was going into get life insurance and on the form you had to fill out when his parents passed away, and what they passed away from and unfortunately his father was a very bad man and he had been hung for his crimes. So he thought about this for a while and when he turned in the form, it read that his father had passed away at a young age of 45 taking part in a public demonstration when the stage unexpectedly collapsed below him. [Laughter] So you can make something good out of something bad and that’s what we’re going to try and do here today. So here’s our crew.

Sean McMains: So I’m Sean McMains, I head up the technical side of our CMS implementation as well as a couple of other projects. We got about 6 programmers we’re working on this and a few other tasks that we’ll talk about briefly later on. My business starts at Alpha Geek and that’s a pretty good description of what I do. I do want to ask one favor, have you guys were using Twitter and I haven’t actually asked these guys about this so bare with me.

But I’d like for everybody who’s not in this session to feel really bad about missing it so if you all would post the most outrageous lies you can think of. A couple of suggestions to get you started, I can’t believe they brought in the Dallas Cowboy cheerleaders. They’ve made a really good point about usability by setting fire to that goat so go with it.

[Laughter]

James Buratti: Oops, wrong one, got to learn my buttons. I’m James Buratti. If I have to pick a title it would be user advocate. My official title at the university is University Webmaster. I’ve been in Texas State since about 2001 prior to that I was at the Ohio State University since about 95 working on the web so I’ve been at this a little while. Then we move way forward in time.

Jeff Snider: Yeah, pass all these old guys and incidentally I’ve been at the university for longer than these two and yet one of the gang and it’s kind of weird how that worked out. As my little tag line here says I have most of my program experience with Pearl on the GATO project. I’m server administrator. I’ve done some programming and almost all of the CSS and HTML and the front end of it.

I’m also very involved in our online learning management system in our university so I’m one of the flip-floppers between projects.

James Buratti: All right, taking you up through time prior to 2003, basically everything we quoted we hand rolled. It was HTML, somebody typing away at a keyboard me or Jeff or our previous university webmaster Fozzy, is all done by hand. Moving forward we actually get to our first site that we launched with a CMS and Jeff can tell you about that, he did all the work.

Jeff Snider: Yeah, so we started seeing around with some stuff about CMS’s and Web Gooey is a pro-based one so that’s where I decided to start with. And it was also really user friendly. Our users really took to it really quickly. This particular site was our telecom site inside of the IT Division so it’s kind of an in-house inside kind of job. But it went really well. I kind of prove a point with a lot of people.

James Buratti: And it also had legs until it did not actually go away until the implementation we’re going to talk about that we’re at now so it’s…

[Cross-talk]

Jeff Snider: Yeah, we actually still to this day have some Web Gooey sites hanging around.

James Buratti: They don’t want to give them up. About that same time that year 2003 at the university is thinking big. We’ve been talking about CMS’s. We got a few implementations out there so we go with the CMS project and the uppity-ups decide we don’t have the expertise in house we need to hire project manager for that. Our project manager we’re in Austin, there’s a lot of competition for smart geeky people near Austin Texas OK. So we chose what in the end I think in either it is on the committee would say, “Not the best of three not so super qualified people,” and we wanted to roll that back and there’s one of those political decisions.

Well, we need to get started, let’s go with this guy.” We’re in a university, what do we do? We form a committee so we actually form a really good committee. We get people all across the university at all levels that we know we’re going to need buying from a very talented group of people and then we promptly ignore them, all right another mistake.

So they come to meetings and the meetings are just death marches. It’s PowerPoints from the project manager with updates. And then when the committee is asked to so something, the work that they do isn’t seized upon it, isn’t used productively so they quit coming. They’ve got things to do with their time. So again, we’re kind of stumbling here right out of the gate.

Jeff Snider: Yeah, this committee our best example of what they did is they came up with this brilliant web style guy that never sort of like it’s still on a drawer somewhere, it was never released.

Audience 1: Why not?

[Laughter]

Jeff Buratti: Look at that we’re even looking at each other. Politics, if I have a better answer I’d gave it to you. It never made it passed the certain point in the administration and we were at a level where we can find out why. No idea.

So moving ahead here, our university went through a name change where public schools so that actually went through the state legislature. Texas State Legislature meets in effectively every two years so they made this change and between when they declare our name would change and the start of school was five weeks.

Jeff Snider: Yeah, we were told that we had to get this new website with our university’s name up by September 1st. So we went through the concept design the implementation, server setup, and content migration in five weeks using the CMS.

James Buratti: And a lot of it was the CMS like we got through that project we were like, “My God, we can do really good things in this really quickly like CMS is a great. This is going to be awesome once we have the big one. The one that we really want to…

[Cross-talk]

Jeff Snider: Yeah, this was the first that we’ve been able to create a website and drive all of our content into it in three days?

James Buratti: Yeah, record time and we we’re getting call times so that helped.

Jeff Snider: Yeah, a couple of midnight.

James Buratti: Sean you probably best…

Sean McMains: Yeah, I think your microphone is pointing sideways by the way. So in this process we’ve got our new CMS project manager in placed and they’ve been going through a whole litany of various Content Management Systems. The people who are involved that showed and marked previous position towards Enterprise Software. My standard joke about Enterprise Software is that a software that costs a million dollars and doesn’t work when you take it out of the box and there are various options that we look at, “Yeah kind of fit that description.” The two finalists, we narrowed it down to finally were Vignette and Documentum that was based on a list of requirements that sounded very good to VPs but didn’t actually very much relation to the day to day work of maintaining content.

So we had things like programmatic workflow for content approval. We had things like integrating with portal system, commercial support available, integration with other business applications. All of these things sound really good but as we found out later on don’t really have a lot to do with maintaining a website a lot of the time.

Jeff Snider: And Sean pointed out the project team’s previous position to Enterprise create software, after much persistent harping on the project lead we finally got at least one open source Content Management System in there. It was kind of a token gesture. It was immediately dinged for the support because that was one of our big things. It seemed like we are really scared of open source we just almost need it to spend lots and lots and lots of money.

[Laughter]

Sean McMains: So based on that list of criteria we finally made a selection. We went with Vignette and I don’t think I have anything more to say about that slide. [Laughter] We began staffing up for the project. After we had the project manager in there, we went ahead and hire on five, six more people who were full time dedicated to this CMS project.

So obviously this is something that somebody with a checkbook feels like it’s a very important thing. We also had four people who are already employed by the university who are lending some of their time to the project as well. So we had a considerable staff dedicated to this so it wasn’t for lack of effort that the horror story that’s about to unfold happened. We also started a parade of consultants.

One of the difficult things about working with Vignette for us was that all of the technical information is covered by an NDA. If any of you have been doing iPhone Development recently you may have some sentiments about NDA as I mirrored this pretty closely because we weren’t allowed to discuss anything pretty much to do with our experience with the product. People weren’t able to write, there are no O’Reilly books on Vignette because of the NDA. There are no websites where you can exchange technical information because of the NDA.

So we decided that one way that we could get access to technical information was to hire consultants so there are lots of people out there who are happy to take our great big shiny checks for this role. I guess our checks aren’t as big and shiny as some in the private sector so we ended up having a consultant who is on for several months and then he ended up moving on to another opportunity. Got another consultant, he moved on after a few months, and then a third one who went on to start his own company after a few months. So we eventually realized that relying on outside expertise really was not going to service well for this project.

James Buratti: The group begins working on the main site redesigned with the new Content Management System. We had just gone through redesign not much more than a year earlier and the reason I says is the eyes off the ball because we’re pretty on a Content Management System the most useless site as far as being able to replicate what you’re doing is your university home page. It’s a one off. Yes, it’s a very important one off but there’s nothing in it that’s really templated that you can share.

It’s like if you build something for a college. I can use that on another college and another college and another college but we immediately through our resources at the home page and I’ll call this the kind of project manager shiny object like, “Woo, home page I will make my mark. Bam! Everybody wants that and this isn’t the last time were going to see this occur. So, we immediately start throwing all of our resources that what at the time considered the wrong thing and I was proven right now.

Jeff Snider: Yeah, I think most of us agreed the site that we had at the time really wasn’t broken. It didn’t need this.

James Buratti: By again, it’s shiny. It’s very shiny and it’s especially visible to the people that are signed on the checks so again some political considerations. We begin working on the information architecture, that’s something that I did. Basically we go out, we start looking at all of our departments, anybody that might go into the system.

We’re looking at other universities not only in Texas, in the US, around the world that where very similar in what we do. Our departments are very similar on how they do it. So, we start putting together information architecture. We start bringing in the people that are going to use this system.

What would you expect to see in your academic departmental site? What would you expect to see in your VP site? Putting architecture is in front of them trying them out, testing them. So that process begins here as well which is more the core of what we’re really going to do with the CMS? At the same time, the project manager is really getting caught up in marketing research, OK.

So he’s designing a home page. He’s going all over Texas talking to juniors and seniors and high school and wanting to know what they want to know about a university home page and doing what is a lot of enrolment management and marketing type stuff which is not what he got hired to do. And when I say going all over Texas, that’s a commitment we’re talking from the border to Dallas. You spend a lot of time in the car so he’s kind of zooming around and our programmers are kind of starting to go a little crazy.

Sean McMains: Yeah, so by this point we’ve had several months of experience developing for Vignette even with our consultants it has been a nightmarish experience. The Vignette is a company that has grown over time by buying other companies and that’s reflected in their API. Their Portal API is actually pretty nice because they bought a company called “Epicentric” who had some good engineering. Their CMS API was much harder for us to deal with.

They had all kinds of very peculiar design decisions in there including each content item had multiple unique identifier. Some API calls would require that you use one ID. Some would require that you use another ID, and getting from one to the other was always pretty difficult. We also tried to lean heavily on the much vaunted support organization that we had paid and again was one of our big considerations in making the decision but they typically took days to get back with us and as often as not their response would be, “Oh, you should really hire a professional services for that,” which wasn’t free as you might imagine.

So we are getting more and more frustrated with working with this product feeling like we can’t deliver the features that we want. We can’t fix the bugs that we want and at this point our blood is beginning to boil.

James Buratti: At the same time, we’ve got a select group of users that are actually trying to start using this system to try and start building stuff out and it is painful as well. Like this is uploading images has more steps in AA I mean these things that should be simple are not, URLs are a nightmare. The things that it publicly shows to people as a URL are disaster. Jeff pointed out that this was the URL to our admissions site. How do you like to give that to somebody over the phone?

[Laughter]

It takes 20 minutes and importantly there was no Mac support in the rich text editor there that they provided actually it didn’t even worked on the Mac so there was no Mac support.  Worst we have a heavy Mac contingent in our university. We have a huge instructional and communication design department so that was a big deal. We knew right there ding, there were a bunch of users that weren’t going to be happy with us.

Jeff Snider: So when we first started to use this thing to see what we could do with it in implementation for our server infrastructure, we put together a set up that required two data centers and nine servers to attune of over $100,000, and that was step one. With that system, with nine servers involved we still have page low times of two and three seconds maybe sometimes four and that was not under low that was in our testing conditions when we were just fiddling with it. We rebuilt, re-architected the thing to eventually involve 14 servers and even when it was working, the software would randomly spew tract stack traces across the page which didn’t do very well for our public phase for the university. It also showed up blank pages, half pages, and all sorts of just random intermittent errors.

James Buratti: Very exciting.

Sean McMains: Meantime, our project manager is also getting really enamored with other shiny objects around the university. So not only are we working on the CMS that’s sucking as dry and making our lives miserable but also we decided, “Oh, we’re going to do portals now. Let’s build the portal and have all the applications wield into that, and let’s do electronic portfolios and do that for certification processes. And let’s make it the Content Management System into an institutional knowledge repository where everything that everybody at the university knows is in one giant database.”

And of course that doesn’t make our life any easier so we’re getting scattered in a hundred different directions. Our project manager is wanting to cure cancer and the common cold and things are getting even nuttier and nuttier. If you see on the next slide we continue to spend more and more money and our team morale gets lower and lower over time.

[Laughter]

James Buratti: Those probably aren’t really stiff enough to actually explain what happened but they are an approximate…

[Cross-talk]

Jeff Snider: Yeah, it’s really demoralizing to watch hundreds of thousands of dollars walk out the door that aren’t going to you.

James Buratti: Fortunately we have really smart people and Jeff comes up with something and kind of saves the day for a while.

Jeff Snider: Yeah, so seeing all of the problems that we were going to have, just trying to put this out is our main sigh which of course was the first set that we were going to do. It was going to crash and burn horribly and at least the geeks we knew that that was going to happen. So I put together a piece of software that would essentially act as a layer on top that would in many circumstances possible show good data back to the user as quickly as possible. So this software cached it and spit it up significantly.

It was configurable to tell what was an error so we could put in a match on the stack traces and blank pages and things like that and it would throw it up as if it didn’t even happened. This piece of software was probably responsible for Vignette lasting on our campus as long as it did.

Sean McMains: And I appreciate that.

[Cross-talk]

Jeff Snider: Every time I say that I turn to Sean and to James and say, “Sorry guys.”

[Laughter]

Sean McMains: Thanks.

James Buratti: It made it look like it was working whether it was working or not.

Sean McMains: And it’s worth pointing out that this is something that we still use even today even though we have a much better system in place underneath it that’s been really valuable.

James Buratti: Yeah. So this main site that we have been working on has gone through all the regular processes that you know the user testing and all that and so nonetheless we do a soft launch on this site. So we basically linked to it from our home page and tell people, “Hey, go tell us what you think of this,” and immediately we got a lot of feedback. I guess say we have 29,000 students feedback is not a problem and so we got a lot of defects reported immediately and we’re like, “Great! That’s why we soft launch we can go on and fix these things before we throw it out there to the real world.”

So almost two months later, February 23rd we’re like ok. The project manager does the hard launch on this and very few of those problems have been fixed not because weren’t incompetent but because so many of them were kind of beyond our control. They were just things we could not get in and make work right so we said, “Well, it’s OK. We’re going to keep plugging away on that and will have the really good version of it coming summer 2005 so no problem it will be good.”

At this point, our information architecture and our kind of common template navigation has been completed. Everybody has bedded this. We think it’s really solid. The way this is going to work is that when you get a new blank site from us instead of just being nothing there if you’re an academic department you go in and you have all these pages. You have your information architecture filled out, it’s blank, and it’s up to you to drop in the content.

Remove let’s say you don’t have a graduate program. Remove those pages that might apply to your graduate program. Add in the stuff but it gives you this great big leap on where you’re going and we’ve got those dialed in for colleges, academic departments, service departments, and VP units and those we’re still using to this day. Also at this time our graphic design department begins working on our template design for these sites that I’ve just talked about. We decided that we’re going to offer five templates with five color variations of those templates so even though everybody is going to know you’re Texas State, you’re going to have something to choose from and they’re not all going to look cookie cutter alike.

Sean McMains: But things continue downhill. The programmers are getting angrier and angrier about this system and having to deal with it on a daily basis. We keep presenting, training new users in this and they almost without fails they, “No, we’re not going to use that.” It’s a nightmare experience for them. The teams starting to get upset at each other and very frustrated with the situation that things are just going downhill worse and worse.

James Buratti: Nonetheless, March 1st, 2006, only two years after the start of the project we launched our first templated website for somebody in the college of health, project manager claims success. March 3rd, project manager heads to greener pasture with the resume that says, “Success!” He got out of the door two days later. It was his price, his part was done he was gone so meanwhile, we’re still all kind of like. Fortunately for us we got a project manager at the university that accepts to take on that what is kind of a stinker of the project.

He’s actually was already my immediate boss and now these guy’s boss. His name is Mike Farris. He’d been the director of our educational technology group. He’s been doing software in Higher Ed for over 20 years so he knows how these things work.

He gets a big bonus hopefully to manage his Stinker. And fortunately for us he is a busy trusting guy. He basically has no time to micromanage. He’s like, “Do what you guys need to do to make this work and if you need support come back to me.” That’s actually him solving a problem. He is an aikido instructor in his spare time so not somebody you want to screw around with.

So were convinced were going to have a good CMS. We’ve had good CMS’s. We have this wonderful experience with web viewing we know we can do this so we start doing our branding work on what our system is going to be so we started working on branding with logo color type. We have some really great design people in our department and…

Jeff Snider: And at this point we knew we have to do something else because previously we haven’t had any service distinct branding so most of the campus knew of our project as Vignette. And that had so much baggage with it, we just had to move on from that.

James Buratti: Yeah. So, at this point, we’re August 1st 2006, we’ve missed a gazillion deadlines that have been promised to everyone and anyone. We’ve missed site launches. We have missed feature launches. We basically don’t have a stable server environment. We do not have a useful CMS, things are not going well.

Sean McMains: And at this point all of us were down on the trend have realized that if we’re going to stay at this project and not actually commit homicide and be jailed, there were going to have to make some changes. What we started to do is have little secret black ops meeting over in the student center in the subway hidden behind the menu so that the vice president of IT does not stumble across us while we’re having these discussions. And we ask what in an ideal world would a CMS look like?

So we start evaluating some of the other CMS’s that are out there and come up with a list of about a dozen. By this point we’re a couple of years into a Sakai implementation as well, and we have a good taste for open source. We’ve used it in a number of other places too so we’re excited to look at those options which weren’t really considered the first go around. So we go through the list and we sort of assemble our new requirements matrix which instead of the great big long abstract one consist of number one has to be super simple for the user because the person who’s going to be doing this is on average going to be an administrative assistant who does it for 30 minutes a week and doesn’t have time to learn complicated weird stuff.

Number two it has to be easy to program so that we can actually implement the features that people are asking for. The ease of programming also works in our favor potentially with an open source product because if we come across bugs in the vendor’s code, we can track those down, make the fix, submit it back to the vendor rather than just sort of being stuck with somebody’s black box that we don’t have visibility into how to fix so this process goes on.

In the meantime, Vignette gets wind of this and sends their Crack Commando Squad down to try to convince us that we really, really want to stay with them. We sit down with them and using this few four letter words as possible layout the case for the change that we’re considering. They think about it for a few minutes and then sighed exasperatedly. Threw their pencils down on the table and then refused to pick up the launch tab for us this time around.

James Buratti: So that made our next decision much easier.

[Laughter]

Sean McMains: So after some consideration, we ended up going with the product called “Magnolia.” Magnolia is made by a company called “Magnolia.” They are out of Switzerland and their core product is that they have sort of a two tier product. The core thing is called “The Community Edition,” and it’s a totally open source projects licensed under GPL so you can take it and you can do any kind of crazy stuff you want with it that’s consistent with the GPL license.

They’ve also layered an Enterprise Edition on top of that which uses all of the Community Edition code plus adds a few new modules on for things like LDAP and fora, document management services and various other sort of more Enterprise Features. The experience with Magnolia is generally a good one. They have some excellent engineering. Their API is a comparative pleasure to work with after the whole Vignette rebuttal and they’ve got a good engineering organization that’s responsive because there’s no NDA.

There is a place for technical users can get together and can exchange information online and we can post it on our web blogs without fear lawyers descending like little vultures and coming to pick our flesh point.

James Buratti: All right so December 1st, we made that decision November 1st a month a later our new CMS is launched, and month later our new CMS is launched. We have our branding. You can see there that is Luchador if you’re not from Texas that most of you aren’t that’s basically a Mexican wrestler. They participated in Lucha Libre which means free fight. If you still don’t know what I’m talking about go rent "Nacho Libre".

We branded this powerful content manager. We play off of this a lot. We want people to have a good time. We also have a lot of people that are obviously seriously gun shy so we want to explain to them that we are not trying to constrict what they’re doing. We are trying to make what they are doing easier to do.

We want to give them assistance, they can go in and I just like Lucho Libre, have a free fight. Go crazy, go and have fun. So we immediately get in to a ton of training and change management that goes in and talking to anybody that wants to talk to us. This is a volunteer process this is not top down.

The university is not telling people to use this. Anybody who wants to talk to us will go in, will do an information session. We’ll tell them the university goals, our goals. We’ll give them a demonstration and then because I have them in there captive I do a 30 minute web best practice where it’s like even if you don’t want to use what we’re using here there are seven things that will make your website not suck.

All right, I don’t care if you use ours or not just please do this thing. And we emphasize user satisfaction over and over and over that’s what my crew does. We’re the only people that tell people in training, “If you don’t like this complain. If it doesn’t work complain.

If it doesn’t work the way you think it should work complain. Don’t put up crappy software even if we’re the ones providing it to you.” And then we have this wheeling at a previously burned which we kind of tried for a little while and we just gave up on that. [Laughter] They’re going to come to us and actually two years later, they all have eventually.

Jeff Snider: Something that’s worth repeating from that is one of our key tenants of the project is that we were not telling anybody that they have to do this. So many times at least in our university things can be dictated from top and we wanted to make sure that everybody was clear that if this made your life easier, great use it. If it didn’t, nobody was shoving you to the door and trying to get into it.         

James Buratti: And one other important lesson is brand the service not the product. 90% of people using our system don’t know its Magnolia. They don’t need to know. They don’t care. They want to make good websites and so we learned everybody knew about Vignette, they knew Vignette.

They knew the product but we don’t want to brand the product. We could hopefully pull the underpinnings out of this throw something else in there if it works better. They’ve still got GATO. All right, they don’t have Magnolia. They don’t have Vignette. They have GATO so brand the service not the product. Our service is simple easy to use, easy to build websites.

Sean McMains: Yeah, we referred to Magnolia as GATO 2.0 for a little while as we’re making that transition.

James Buratti: Support starts to ramp up for me. We have many new sites coming on quickly. We have move to January 15 so we’re only a month and half into this. Our training is expanding. We’re offering our basic training our entire training is two hours. Two hours, my mom can come in and learn this system, build a website that’s the way we’re shooting for, that’s what we went for. Also could have been in classroom for more than two hours our bladders can take it.

There’s just no biological point to it, all right. We get our launch process dialed in because we have lots of sites that are starting to ramp up. The users are building this very quickly. We basically go in and when somebody says, “They’re ready to launch,” we have this process.

We look at every page. We look at every link. We love student workers. I have talented student workers that do all of these for us. They look at every page, every link, every picture, we fix what we can so when they launch their site is as good as it can possibly be, all right.

And then we also add some brilliant apps that made my life so easy that these guys made for me. One in particular I want to point out was the application Sean build for me on rails I would call it a “Life Cycle App.” We track every site in the system, who the owner is, who the managers are, who can work in that site, where they are in the process, have they have an information session, have they gone through training or in development, are they in the process to launch? And it’s great because somebody can come running in with their hair and try and say, “Where’s the VP of whatever?”

And I can be like, “They went through training on the 9th and they’re in development. What else do you need to know?” “Who is it these five people?” So that sort of thing, having that data right at my finger tips has been just really, really, really helpful.

Sean McMains: So at this point, a lot of things that have been sucking stop sucking. The engineering team is getting much happier because we’re getting to do programming and we don’t hate it. Our server room has become strangely quiet and cool because we have scaled down from a 172 servers to four in our production system, I’m overselling a little bit. [Laughter] And the users are really happy because they really like the new interface and they’re able to get their heads around it very quickly and without much hassle. Meanwhile every story needs a villain. So our university has hired in the meantime an enrolment management Czar and he has decided that he is also in-charge of marketing.

The higher apps at the university are sort of backing him in this move and as part of his decision to put his brand on things, he decides that all the graphic design for all the websites needs to change. So conveniently enough we do have a CMS in place but these guys are not guys who have any web experience so they basically come to us and say, “Look I have a half page Photoshop file. Can you make every website on the university look like this?” And we have to do that so this sort there’s a fly in the ointment. Our technical challenge is like kind of fading into the distance a little bit but we’re having new political challenges as a result.

James Buratti: And of course they want a new home page.

Sean McMains: Yeah.

[Laughter]

James Buratti: All right.

Audience 2: Question, why is it a challenge?

Sean McMains: It’s a challenge because these folks were not good at listening and they have no experience with designing websites and this sort of interaction that’s necessary there. So we had both individuals who are difficult personally to work with who didn’t have the professional experience and weren’t interested in hearing from us from the experience that we’ve had in that area.

Audience 3: So that your group didn’t take responsibility and haven’t felt the accountability for the content website or whatever website? 

Sean McMains: At that point we did have some responsibility for the design. We had worked to develop the initial design with the deign team in our area. At this point that is over too marketing. That is their responsibility and as far as content goes we provide advice but we don’t take responsibility for the ultimate content product there.

James Buratti: And that’s not a problem if they listen to us as far as information architecture and user experience goes but they’re like that was part of the struggle. Its like, “Well, I know this navigation is a little shorter than you’d like but if you put 40 items on there it’s not as effective but I’m sorry we need 40 items.” So when you have a group that was just like that is really used to getting what they want, we will probably very, very difficult to work with for them in that sense. We weren’t trying to be but again I call myself a ‘user advocate’ I’m going to push forward what I think the user is going to work best for them.

Sean McMains: And as far as the content management goes, one of our buy lines has been kind of to make the right thing to do, the easiest thing to do within our system that people can go crazy and do whatever madness they want but it’s going to be a little more work than just doing things more intelligently.

James Buratti: Sure.

Jeff Snider: So with the single page Photoshop file that they gave us, we didn’t have much else to go with. We were able to ask some questions but in our previous information we mentioned that we have five templates with five different colors. Now, we had one template but we still had a predominant color on the page. Our top banner image was predominantly Maroon which is one of our school colors and I got into my head to write a way to dynamically tint that image. They only gave us one but we needed five because of that previous deal so it turned out to be just as easy to write a dynamic way that would accept any color.

So Sean and I got going back and forth on this and put together a way for user to be able to just select a color off of a palette and it would tint all of the images and change all of the CSS for those highlight colors and basically just kind of re-tint the entire page.

James Buratti: It was so cool.

Audience 4: It’s really neat.

James Buratti: Now we call that our “Super Color Change Power” so the user could just go in and they had a flip of the swing change.

Jeff Snider: And they could do this at a site level, a sub site level, or an individual page so they have fair amount of flexibility.

James Buratti: So here we are seven months passed selecting and literally hitting the ground on this CMS and we’ve got our 50th website going out the door built by the users. If these are built by us because we felt sorry for the folks that have been in Vignette and we move their sites for them. We felt they shouldn’t have to like have double jeopardy and suffer. So we moved about eight of those but most of those are by the users so things are ticking along pretty quickly.

June 11th we launched the new home page so we turn around on that pretty quick. It’s about seven months from their design and enforcing them to do some user testing which they didn’t want to do. And we launched this and we’re kind of co-manage it. We got to do some fun stuff with some scripting and putting Flash and some other items in there and they were very happy with it and it went over pretty well at the users as well.

Sean McMains: Meantime, things have stabilized to the point where we can start doing monthly iterative releases. This is sort of an agile practice and we very much like the idea of being able to release regularly once a month so users know when to expect the changes and so that we’re able to roll those out on a very frequent basis and be responsive to what the user need is.

James Buratti: So here we are September 11th, 2007 and we basically push out the new template that marketing has provided. Now, remember prior to this they have some choices as far as which template. We kind of call this the ‘one ring approach’. There is one template whether you’re an academic department, a VP units, a non-academic department with all the sites are going to look the same so everything gets standardized. The beauty of the CMS is though we can just roll this out.

We did have about a month of transition where people could look at it, choose this template, or choose or stick with the old template but they knew this was coming. September 11th we threw the switch, all the sites rolled into the new template. Yes.

Audience 5: Do you say that half of your CMS? How do you deal with it?

James Buratti: Well, that’s for marketing to deal with. Wait again, the folks we have are mostly the ones that want to be in there and that’s somebody else’s battle. So here we are 11 month passed since the selection of the CMS and we have our 100th website get launched. So we are feeling pretty good about ourselves and how this is going.

And then March 1st 2008, someone in marketing discovers Jeff’s Super Color Change Power is in the system. It comes across a website that I think it’s blue, it was like a business site and blue is business and just goes running into his boss sounds like, “They can change the color of their site,” and goes running out and then we later get a great description and he comes in and says, “Its worse than I thought.”

So we have not hidden this, we had put it on our release notes it was out there. We also had an advertised it. To put it on a wrestling terms, we got it like off the third rope, a big smack down on this. We had to take it out [Laughter] so everybody went back to maroon and gold. Users had mixed reaction.

We had basically the evil empire versus the good of the cause. For us it’s like it’s still a cupcake but the cupcake with all the coloring was a lot more fun but it’s still a cupcake so nobody is going to turn down a good cupcake. 17 months so we’re into this year, April 4th we launch our 150th website is out the door remember these are being done by the users. We have this launch process we’re really rolling with it and you can address this one.

Sean McMains: Meantime the CMS is finally getting to the point where it’s mature enough that we can actually take some of the people who have been dedicated to it and start repurposing them, reusing them in other projects. So now we have some of the engineers who are originally hired on to do CMS work helping out with the learning management system with an event calendar that we’re working on putting up around campus as well.

James Buratti: And on the flipside, my group is ramping up. I hire a full-time support person for support and I hire a graduate student that is shared with some other systems so he’s ramping down and I’m able to use my Life Cycle and say, “Look, I have 600 people trained and I have 150 websites, can I have some staff?” So we’re ramping up on our site while they’re able to put their resources elsewhere.

Sean McMains: Is this with me? All right, so [Laughter] in that time we have also been adding new features that have generally been well received. One of them is called “Link Checker,” and this is basically a spider that goes out every night and prepares reports for all the site owners on what links are broken on their website. What ones are pointing to area of 500 pages all that sort of thing to help maintain the quality often page hopefully without people even discovering that there are broken links? We also added a slide show builder that lets people just take digital photos right out of their camera and upload them to their site.

And it rotates and rescales and does all the necessary stuff to make a nice little DHTML fading, scriptaculous-y slide show thing so that’s a lot of fun. One of the things that people really like is being able to stick in paragraphs into our system where the data is maintained automatically for them. So one of those pulls data from our HR System so you can do a departmental directory have that constantly updated, always up to date based on the latest data that’s in our HRF Systems. We’ve also added a Document Management System which gives people a central repository for things like Word and PDF document that they may want to reference from multiple pages in their site or embed for multiple sites.

And we’ve added more sophisticated forms so people can do like registration forms on their sites or more information forms then they build it with various elements and have them sent to people and have a confirmation email and all that sort of stuff. And we’ve added a bunch more stuff that’s all in our release notes page if you’re really having a hard time sleeping some night you can check out.

James Buratti: So this entire time my group, we’ve been surveying our customers and we previously about six months in we have a 95% would recommend it to others. As of two weeks ago we surveyed everybody, 97% of our users would recommend the system to other people. We asked them what is most important about this system, “Ease of use.” Not even close to anything else for the end user, ease of use.

What’s the number one hate the common looking fields which is why we did have that theme in here and why we want to show the process that went to that? There is a problem when you move from one site to another. There is not necessarily that visual cue to know that I have just moved from the college of whatever to the department of whatever, they look very similar. But yet common looking fields are also an item to love that people like so it’s one of those that goes back and forth. Oh Jeff, something to say on that?

Jeff Snider: So the statistics, we gather some of them today, some of them just before we left. We’ve got 183 live sites and how long have been doing this now? 16 months? 18 months?

James Buratti: 18 months probably.

Jeff Snider: Yeah. We’ve got 60 still going in development that aren’t launched yet that changes all the time of course. Almost 17,000 pages on our campus are in this system, and of those developments those are totaling about 3,000 pages, 750 users that doesn’t mean that they’re always all using it but this is how many people we have registered in the system and who can edit pages if they want to. And of all of our traffic across the campus, we’re pushing about five million page views a month out of this system.

James Buratti: You put the wrong way.

Sean McMains: No I didn’t.

James Buratti: That was today. Now we’re shooting into the future.

Sean McMains: All right, so the future for us we’re looking at trying to release all of our GATO work under an open source license so that other people can benefit from it as well. We’d hope to have that up and available today but I haven’t been able to pull together all the things we need to, but we hope to have that up before too long. In your handouts, there’s a link to our open source at Texas State site which is just a track installation. It’s pretty regional right now but will be adding more content to it as time goes on.

So if you’re interested in having to look at it, that’s the place to go and tell them what they want.

James Buratti: And on your seats, like I said we had fun with this. You have your GATO temporary tattoo this is our graduation diploma before they go to our classes so people show off. And then we set up basically, this is our test environment so if you’d like to go in there for the next two weeks you can jump in to our test environment. We have about 200 or 300 accounts created each one of these is unique. You can go in, play around in our test environment you can’t break it.

There’s if you want documentation you can go to our website which is in the ‘Do your notes folder’ and actually download our training manual. But you can go in and play around and like I said we’re trying to get to a point hopefully like I said Sean made a Christmas present, December 25th, but you can go download what is GATO. You would pretty much have everything that we provide for in our system. That’s it.

Audience: Does that include the Link Checker and the Slide Show Builder and most things are built?

James Buratti: He was asking if that includes the Link Checker and the Slide Show Builder some of those things. I want to let you guys handle which of those that would and wouldn’t.

Sean McMains: Link Checker is not…

Audience 6: The Forms Builder.

Sean McMains: The Forms Builder yes. The Slide Show Builder kind of, it relies on external portal imagery sizer module that Jeff wrote I don’t think we’d had any problem with the listing as well. And then the Link Checker actually runs as an external process so it looks from the outside the system just like a user would so it’s something that we probably will release but isn’t part of the Core GATO thing yet.

James Buratti: OK, we’re getting really close to running out of time but we do have a break after this. OK, we’re good. OK, questions.

Audience 7: A few questions. One is you hutted that you’re keeping all of users and is it one system and you manage all of your users under one system? Second question is, does the system put the pages together on a fly or is it one of those deployment systems where it pushes the pages out statically?

Audience 8: Can you repeat the question?

Sean McMains: Yeah, absolutely.

Audience 9: The podcast of that online?

Sean McMains: There were two questions. One is whether all of the users are under one system and that is yes, there are all in one box. The second was are the pages rendered statically out to HTML files or are they generated dynamically on the fly. The answer to that one is that they are generated dynamically which is really useful because again we have department directory information that gets updated obviously people change where their pages are located so the links get regenerated automatically on the fly as well.

James Buratti: However it is kind of a little bit both ways and that we have an editing environment that people publish their material once they’re done out to the public environment, so there is an opportunity to check it before it goes out.

Audience 6: Two questions. First is you guys had a law in and reports.

Sean McMains: We do.

[Laughter]

Audience 10: What would you do if you could go back if you could do a project differently, is it a project management to Vignette but what would you guys do if you have to do it all over again?

Sean McMains: If we could go back in time, what would we do differently is the question. I don’t know if you guys want to speak to that at all.

Jeff Snider: Well, I think the first thing that what we would do is when all this started because we took a hard look at what we needed to do and decided that we didn’t have the skill in house to be able to manage of this size. And I think that has been proven clearly wrong and that’s what started out the process of hiring the project manager they got this all wrong in the first place.

James Buratti: Yeah, I would agree. I was actually on that selection committee I didn’t have much clout and like I said we hired the best of three bad candidates and we didn’t have fate that we had the expertise in house to pull this off when we had pulled off some pretty good stuff prior to that. So I don’t know I could have changed that. I’m not one to be quiet and ask anybody I work with so I probably complain as much as I could but yeah.

Audience 11: Sorry, specifically about Magnolia the three things. What were the three things that drew you towards that?

Sean McMains: The three things that drew as to Magnolia was the question and I touched on briefly the ease of use for the end user is quite the biggest deal both on our surveys in our own personal experience. Ease of engineering and customization was also a big one and then having a manageable administration was a big deal for us too both in terms of fewer production machines and less demanding hardware requirements and being able to setup that instances without taking a whole day to do things like that.

Audience 12: You said the number one dislike was that people have this thing of where they are with their departments, have you looked at ways investing that kind of solutions or is there any?

Sean McMains: So the question is have we looked at various ways to help people have a sense of location within the various sites around the university and we have done some of that. The colors was a nice cue and unfortunately we lost that but we do have things like bread crumbs in there. We have the site title boldly at the top of each page. There’s probably more we could do in that area but we tried to address that need to some degree and give people some sense of geography there. Does that help?

Audience 13: How many people are supporting Magnolia right now?

James Buratti: How many people are supporting Magnolia was the question.

Jeff Snider: You’re talking about us or with the corporate?

Audience 14: IT folks.

Jeff Snider: OK, with the IT people.

James Buratti: On my side of the house which is training and answering tickets and now phones which we just started our first whole year and half with support was just via email. There are one full time and then a student worker and then two people that I would say half dedicated to the project in their shared time, and then my time whatever I threw in there I would say like a quarter of my time so maybe half.

Jeff Snider: Nothing here.

Audience 15: Is it curiosity when you guys were looking at different Content Management Systems? Did you consider the open source Dripple?

Sean McMains: Did we consider Dripple when we’re looking at the open source management systems was the question. We did look at it briefly. It wasn’t something that we inclined too much towards because it seemed a little bit more community building system than just a content system which is a little bit closer to what we’re looking for but also because we didn’t have much PHP expertise in the house. And we’re a little bit reluctant to dive to the deep end of that pool without knowing a little bit more whether there were sharks in there.

Audience 16: A questions back a little while ago, as far as the Magnolia release, have you had any security scares as far as site defections and things like that?

Jeff Snider: The question is that we have any security scares like site defacements and like. No, we haven’t had anything like that come up. We had people who actually accomplish something that they wish they shouldn’t have but nothing as an outside force. Also something that’s built in to the history cache that we mentioned earlier is the ability to kind of rollback time.

So if we did have defacement we could just tell, “OK, rollback to slightly before that it actually happened and we’ll be able to undo it. We have clean up work to do after that of course but it would get off the main site.

Sean McMains: After the cross site scripting session this morning I’m anxious to go back and look at a few things.

Audience 17: Besides Magnolia, if Magnolia hasn't existed, do you have any other runners up?

Sean McMains: The question is after Magnolia, if it hadn’t existed with the runners up that we really like and I want to say they were. I don’t remember what they are I think we actually looked pretty seriously at paper pins product. Is that the one that was done in cold fusion? I’d had a really nice interface from what we have seen but cold fusion again was not our area of expertise and doesn’t seem like as many people’s area of expertise anymore. But that was one that we were drawn to. There are a couple of other ones that would have been pretty sound contenders as well.

Audience 18: What was the implementation for Magnolia?

Sean McMains: The technology stock? Technology stock at Magnolia is basically Java, its JSP built on top of bunch of circlets that runs in a standard container. We’re running Tomcat on ours. It uses four storages, it uses JCR which is the Java Content Repository also JSR-170.

A lot of the Content Management Systems are moving over towards using that as the underpinning. It’s basically an object database API and it can be built on top of bunch of different ways of storing the data. You can put it on files on your file system. You can put it on relational databases. You can do all kinds of things but at the application level you’re insulated from that because you’re just going for the JCR APIs.

James Buratti: I think we’ve got to end it up. We’re going to be here through the week so

Audience 14: Yeah, I think if we still have questions feel free to come up and talk to them one on one. Thank you gentlemen very much.

James Buratti: Thank you.

[Applause]

Announcer: For more presentations from the 2008 HighEdWeb Conference visit HighEdWeb.org/2008 or sign up for our podcast and feed at HighEdWeb.org/podcast.xml

[End of Music]