 Thank you, Dan. So What we're going to do is I'm going to tell you just a brief little bit about me I'm going to tell you about what I do at the University of Richmond and what the team I work with does What the driving forces for change were? What the change was that we accomplished and what the outcomes were? We have plenty of time for questions and conversations small intimate groups. So let's get going. So I'm the director of web services. I'm Responsible for the care and feeding of about 150 websites about 140 of them are branded University of Richmond We are fragmented into subdomains. So chemistry and Anthropology and HR and they all have their own sites Some sites are two or three pages Probably the average is a hundred the biggest sites probably six eight hundred pages. There's a lot of dynamic content on those pages though I am A geek at heart. I'm been a geek my whole life Used to do some chemistry projects in the garage and luckily I didn't blow anybody up including myself I'm a maker. I build stuff my drives my wife crazy because I don't put it up I just got projects everywhere But the most important thing in my life are my two daughters and they're involved in first robotics So they have some STEM interests. So she's goes to the University of Richmond. That's Jackie and that's Allison She's a rising 10th grader So we're going to talk about about benefits We're gonna talk a little bit about some heavy lifting stuff, but not too much because this is supposed to be the business track So what do we do at the University of Richmond in terms of the web team as I mentioned about 150 websites? We have all kinds of websites 140 of them are branded you are and have some common look and feel here We have Mobile ready pages every page we have on the branded sites are in fact mobile ready. We don't have two pages It's just a flip of the style sheet and off we go. So We hide some content, but not too much. It's all pretty much available there We integrate third-party apps. It's a routing system for our bus that goes downtown and comes back But we also build our own mapping systems I don't know that I haven't I have one picture of a mapping system. We have a content management system It's important in the context of what I'm talking about the content management system. We use is cascade by hand and hill It's a push content management system. It's not like Drupal or WordPress. You don't interact with it as a web user We just push pages so And we've built our own mobile app we went from zero to 200 miles an hour in three months. It's Android and iPhone It's got a lot of modules in it that takes content now most of those modules take content out of the system I'm getting ready to talk about and it's got an update wall and the update wall has These stories you can pick based upon your preference What kinds of stories come in based upon tags out of an XML database? So the wall update you see the one this was taken that 15 new updates And it also then works for our mobile pages so sometimes we downgrade to just taking you to a mobile page for For the expediency purposes. We learned an awful lot about this We've got a lot of phone calls people go who did your mobile app for you and we did it so we're very proud of the mobile app it is important in the context of What we've done that I'm talking about today But because we're enabling certain kinds of content both Mobile and web and it can go to other devices if we want it to so what are the driving forces for change? In the web and change at universities will some of these you'll probably know about some may surprise you first Is cost everybody knows about cost demographic changes and competitiveness so The cost of higher ed has been going sky high. You see it's double that of the rise in medical cost and And much higher than all the rest and I don't see it abating for a while and hopefully regulations don't come in and mess everything up, but For us We have to do things faster better cheaper. Also Demographics are changing. This is kind of hard to see Probably not surprising. This is you know, Caucasian white declining the highest curve up there is a Hispanic Rise in Hispanics applying to college. So now those change from their initial population. That's not overall weighted There's a huge tectonic shift going on in higher ed Fewer students are applying partly because of the change in the population curve Partly right now because of the economy Way more females than males. We're not fully understanding all the driving forces 57 percent at our university Of the freshman class or females So that's a pretty big shift. Some schools are seeing it more dramatic They're coming in savvy. They're connected. They're mobile Smart phones university richman's a fairly pricey school That we have some need-based students as well. They're all coming in with iPhones or Android. Most of them are iPhones or iPads They require better content. They don't want to be marketed to they don't care about marketing Everybody's marketed to 80 million times a day. They want to know what the experience is like And on average students are applying to 10 colleges now 10 When I applied to college, I think in which was eons ago I was like before dirt I think I applied to three places. So that's a big shift So what does this mean? Well, some schools are going out of business Some schools are not surviving. Here's two examples in the last couple years one in St. Paul in virginia St. Paul's Closed just like three four months ago We're going to see more especially look smaller liberal arts that are not particularly selective closing We have to do more with less. We have to do it faster better And we heard that call So we got a request in 2009 from the president's office Said reinvent our web. We want it to look good. We were now in the content management system We had taken a couple years to get everything in We want you to reinvent it And oh by the way, can you do it in 12 weeks and we don't have any money for you All right, so typical story right and we instead of freaking we said wow, this is an opportunity So we had a crazy dream we wanted The ability to create Experience-based content quickly We wanted to support our factory model, which I'll talk about that dan calls it agility We want our users to find the content easily And we want to deliver targeted content to specific users So I like project names You know, I think project names help us get centered and focused So we found this name Artemis. Artemis is a greek goddess. She's the goddess of wildlife She's the goddess of the hunt And she's the goddess of childbirth. Okay Any cute kid? That's one of my co-workers young child. Okay, so i'm going to come back So wilderness we had no idea where we were headed We had no idea if we could reinvent the web in 12 weeks And that has nothing to do with the content on the back end that was kind of an add-on that we were trying to do We were on the hunt for a solution To those principles I showed you a minute ago And when we got in the middle of it, it felt like childbirth. He's a little older than a baby, but So that's how Artemis came about And his eyes are that blue So we already had a factory model. We were we're scrum based And anything that we do twice, we don't want it to be as hard the second time as the first time So we want to I talk about cranking the wheel. We want to crank the wheel And get productions going. We had done that with the content management system and really cranked out a lot of sites Um And we wanted whatever we were going to do with the content system to fit into the factory model Okay, so here is Um a drop down. It's pretty ugly Only only the admins see all of these because if you you have permissions You probably have permissions to six or ten of these items But that's in our content management system and those are various little pieces of content Types and whatever that you might want to access All right Everything comes out Initially everything was coming out of the content management system You'll see some pictures in a minute. Some things are coming out of third-party apis and other systems So, um, we had five guiding principles. We wanted to be flexible Um Integrated content. We wanted semantic joins. We wanted the content to be single source one of the problems in higher ed and I think it's probably a The keynote this afternoon was talking about 70 policy systems, right? We want a single source of content We wanted really clear wayfinding and navigation Very strong storytelling not marketing. So that was really important. We'd already started that process and we had a article system written in oracle That's we spent more time maintaining it than we spent getting content in it And we along the way we wanted to improve business processes It's always good to get some wins that are unexpected. That way you keep your job, right? So I had already listened to this floss weekly podcast Because I had a system that I built in oracle with the help of a contractor Content entry was model view controller. It was php app. It was horrible to maintain Um the xml was just stuffed into oracle as a blob It was not really searchable. Um And I wasn't and the people that I work with in marketing were saying well, wouldn't be nice if we had this content Tighten this content. I've been there and I was going home going oh my gosh I'll be working with the oracle until I have no hair. Oh, I have no hair So um and I listened to this podcast and like that's our solution if it only works Okay, so somewhere between that december and 2009 summer I found dan we talked a little bit about how we were going to do things and he helped us out some Got his training and all of that. That's um, but this this was my inspiration. So Exist eb How many people know what mark logic is? Don't know what mark logic is so a couple people so exist eb is open source xml database mark logic is a proprietary xml database It's standards based it has a rest api It was very attractive for real easy to get content out of out of it One of the really big things is it had a fast leucine full turps full full turps full text search ability And xml is treated as a first class object in it So we played with it a lot. It's pretty easy to set up. It's pretty easy to get started At the time, I don't know if it was xquery 3. I know it's 2 at the time, but it is now xquery 3 standards Okay So we played with it and we said, yeah, we think we can make this work All right. So here is a little bit of heavy lifting So the way the architecture is set up Is we have um on the right we have Exist db with the xquery service and xml documents and search indexes all that's just Right in we have to write the xquery it turns out that once you know what you're doing. It's pretty easy to do It's actually really it's not pretty easy. It's really easy to do We use apache web server someday. We'll probably move to engine x but Today we use apache. Um, that's all of our web content is served up by apache But within that we built off and proxy services xml to json service xml to html services and then we have other services like ldap and whatever So we proxy access to xml Through apache none of our users can get to exist db That way we can take full control over the look and feel because we're just getting xml out We're not rendering it as we pull it out. We pull it out and grab it into apache And at that point we put it into the dom model In the web pages right using php and so we get mobile And we get we get web and of course we have static web pages Okay, so think of this is now a mobile platform as a service or content as a service, right? Either one or both All right, so how do we get content in? Well, this is critical for the agility piece Okay, our content management system and our social screens in a minute our content management system has Um this thing called a data definition and all it really is is a form system You can just say add these kinds of form features on it name on Whatever and then we write a little bit of xsl t transform on the back end and take the data out of the forms We can generate it in whatever format we want. We put it into xml and we We write it to a file Okay, and then that becomes our producer endpoint for a middleware system So this is the middle piece here this kind of greenish color Is a middleware system that we wrote it's in it's in java, but it's really simple it picks up files At the simplest it picks up files and does an hgtp put to three exist db instances one for qa and two for production But it can optionally transform stuff so we can go look something up in oracle Or we can change something morph it into whatever We now have built a version service where we can write it to a version s3 bucket and say versions We haven't put that into production yet, but we're soon going to do our university legal policies in that And then that becomes a consumer endpoint and gets hgtp put into the database Okay, we also then built we have a third party A little java problem that goes out to third party apps and grab stuff from rest We go to twitter we grab twitter feeds we go to the athletic The athletic sites are run by another company We go get rss feeds for scores and news and all that We we go to a lot of different systems. There's some other third party systems we go to With it and then we also have a topic manager the topic manager actually runs in exist So it's an app even though it's over here. This is for data entry And because it runs in exist it you can save natively right into exist And we don't need to Do an hgtp put This is the top of a data definition So I apologize, but the small resolution makes it hard to see here. We have a tiny mce like a wizzy wig And then we have other fields that stuff goes out into those those data items go into the xml It's really easy to create We have 150 users across campus that know how to put data in them There's an unstyled version of the topic system. We haven't styled it yet and dan's very embarrassed. He helped build this It's not styled yet. We will style it soon But that goes right into the database. That's for us. That's a little knowledge management system for us to maintain topics Okay, so that's it on the heavy lifting. Okay, so now we're going to talk a little bit about So what does this all mean for us? Okay Why did we do it? We talked about the driving forces. So now What did we achieve and why do we think those are benefits? All right, so one of the things that we have in the system are bios and Biographical information we started with faculty now. We have faculty and some staff Bios show up as a bio a full bio over there for Dr. Salisbury They show up in the directory So you can get to his bio in the directory They show up In an expert guide if they're labeled as experts. That's a place for the news media to go find Who's the expert in intellectual property? Who's the expert in whatever topic? Those are the ones that will take a phone call from the news. Nobody else will It shows up under faculty staff listing. That's how that got opened up over there And here's an article this is the experience stories that really drove the whole initiative So this is how Serving in the peace court continues to affect you are Professors decade later right and we have a semantic join that we find out who which Faculty members are involved and really in some cases other people are involved They show up as related content on the ride This is kind of our inspiration was people who bought x also bought y That's kind of what that is Um without doing it in a more sophisticated way Yes Yes, yes, it it's it well it's extracted Right, so we say we we know This is tagged with a unique idea for the professors Um, and then we know how to get bios snippets out for those professors So we don't do it as a join we do it as two steps But there's a block for the related content and it's like four lines of php All right, plus a global kind of a global library that renders xml other questions about that Okay, we also have faceted search. We have a google search appliance which unfortunately google's getting rid of We have to replace it soon For just searching our web pages, but we also pull people back from the directory Which is also run from the content system from existdb And from this is other stuff that's in the existdb. So this is now faceted search Which has been very useful for us folks love that All right, so now we've got an experience where people can just navigate to what they're interested in And that's what those kind of lines show over there Um, mr. Nielsen of the nielsen rating said users want to construct their own experiences by piercing together content from multiple sources emphasizing their desires in the current moment I'm sure you all know that because you probably engaged in that as you surf and do whatever Um, I don't know how many is anybody in here involved in higher ed Okay, so go to websites pick a university go to your alumni site, whatever most of them are really hard to navigate And I know we have places that are hard to navigate. We're constantly trying to make it better and better You go to sites especially public schools. You go to sites Chemistry will look different than biology, which will I mean, they're all different. They have different styles or different groups I'm lucky. I work at a very small school. We're 3,500 fte's full-time students equivalent of full-time students and We have a president says work together That's very unusual in higher ed. So he you know and we were working together before he showed up But just made it easier when he got there All right Here's a very different look So let me check the time We have this place called the center for civic engagement And their job is to find partnerships with nonprofits in richmond virginia, which is where we're located And connect students with those nonprofits so they can get service work learn about the community learn about service work And where the folks that have the need get the need satisfied all free labor, right? And traditionally students would go to the center for civic engagement and say I kind of want to read to elementary students Or I want to work with rebuilding homes in in a you know in a bad neighborhood or whatever And they would help them find the right partnership Um, this replaces that so if you look in the middle it says issue Students look up the issue they are they care about Um, they click on the issue They see the location for that in the city. They see the vendor Um, here's more information on the vendor and then hopefully they pick a place. This is revolutionized And this sounds for us in the technology business probably like well, that's pretty easy Yeah, it is actually pretty easy, but in higher ed and whatever thinks about these things, okay This is revolutionized their business Students walk in going I want to work on this project So now that group can do more with less And and they're not cutting staff. They're just doing more work in the community It is totally revolutionized How they do business and all of that content comes out of exist tv And then of course we did the math stuff and why do we do it? Well, this is why we do it You know the outcome is for the students Okay, so Really faculty don't always understand this But our constituency is not students. It's not staff. It's not faculty It's prospective students and their parents And a few guidance counselors in high schools and whatever That's who we do this for Because without prospects, we don't get students and without students. We don't graduate All right, so here's a quote. This is typical I'm not going to read it. You can read it. There's a typical quote I walk across campus, you know all the time to meetings or whatever I run into people They always look lost because parking lots are never labeled, right And they're like, well, what do you do here? And I'm like, I do the web And they're like, we're here because of the web I mean, I've heard that like 10 times in the last two years We're here because your website was so attractive. It was so easy to use or whatever Just like this All right, we're also we've gotten rid of at least 400 person hours in admissions The application rate has gone up partly because of the web partly because of other factors And the prospects get to do more self-service and we're not done yet So we have a crm replacement going on right now for admissions and we'll be tightly integrated with them in about six months Okay We stopped printing catalogs we started putting catalogs into xml Before exist db But we shoved them into oracle That that worked but it was problematic in terms of effort All of our catalogs and most of our schedules are now in this back end And we're not printing 20 000 or so variety of print publications a year We're saving trees. We're saving money We're not printing a annual directory anymore of people. That's also coming out of this system So so we're saving trees and that's a lovely tree near the school not on campus, but all right Now agility time is money, right? So How do we get The agility well we start with a data definition And here is a rundown data definitions take a couple minutes to a couple hours to create The middleware system takes five minutes to add The endpoints in if we have to add a new transform Most of the transforms we can just reuse and reconfigure But if we have to add a new one, it's usually about 10 hours or less of programming The x queries anywhere from 30 minutes to four hours most of it's in the 30 minute range that's programming About 30 minutes to code it to shove it in the DOM object And then putting it in the pages is not relevant because we'd have to do that anyway So in a day we can add a new content type or in two days Okay, usually the problem is where's the content coming from some groups got it created Um, so people come to us and they're like, oh, we know this is really hard and we always play dumb Yeah, this is really hard. It's so much work And then we're done with it by the time they've Got back to their desk and we've got it enabled and then we just wait for them Okay We have 56 56 Integrated content types I think there's another one that's missing up here. I think there's 57 now Corsi valves study abroad press releases reporting positions Um admission blog system we have we got rid of wordpress For admissions and student bloggers that blog for the benefit of the university. It's all in Existdb With the content management system So how do we not work well when we release koi pond one we're now at koi pond three That's the name of like weird names right koi pond Long story. I'm not going to go into it the bounce rate dropped pretty precipitously when we roll it out We added a bunch more content types tweak the css and a couple other tweaks and stuff that we like when it went live We're like, oh, I hate that. Don't you do that like you probably when your projects are done You like hate the project All right, you find all the weak spots and everybody else is like, well, we really love this But we hated it so we fixed that there a few months later and look at it plummeted So it's running at about 45 percent bounce rate That's across everything Rolled up All right, so here this is dramatic So if we look at this the the light blue is the in the middle bar is The either the minutes on site or the pages you touch If you've touched the calendar somewhere in your visit and the calendar events are done in this system So 170 and 243 percent increase This is working from an engagement standpoint if it wasn't we'd probably rip it out and do something else If you touched a feature article time on site goes up because you're reading the article probably more than anything You're probably linking to a couple other articles Page views goes up 173 percent there This is how we know it works Okay Folks don't complain about being marketed to it's actually kind of hard to find what were ranked For us news and we'll be reporting all that What they do find out and what every prospect wants to know is am I going to fit in that's what they all want to know They already kind of know the sat scores that are needed and all that am I going to fit in So the stories the feature articles are about the experience and from that they can read it and say yeah, I can relate to this or I can't Mobile users Most of them are coming back multiple times a day on the left so look at the bar up there and The number of visits they have in a month are Some are coming back You know 50 to 100 times 200 times or more So we know the mobile is working as well most of the mobile content comes out of the same system All right, and mobile users say on The site about 15 minutes, which is amazing My poor eyes they don't work so much on them Okay, why exists well it's schemalists it's xml You know you all probably most of you probably hear about scale performance scale performance t value Which one should we be in but it's all about scale and performance 100 million users or whatever You know For us It's really about ease of use. It's about agility and content discovery The engagement numbers you just saw is a you know They're finding the content and they're getting engaged with the content. We like exist first because it's open source It's a perfect fit for us if we were running, you know 100 gigabytes of data We'd probably need more logic or something but for us it's perfect our data is about 100 megabytes We probably have 20 to 30 thousand documents in it now The full text is lucine, which is now called solar is it Okay, all right, so And it works the search works good The search is uniform across everything we do other than the google search appliance But most people use the faceted search. It's easy and it's flexible Um Dan came in and trained us. We had to do a mental shift because it's functional programming. We're procedural programmers php. Whatever Um And it took a couple months for us to get our head wrapped around xml's first class object instead of treating it like text um, but once we did that we Do really really well the guy that does most of the x-query programming touches it In a given week, maybe zero Maybe two hours some weeks may be eight But most he's probably averaging about an hour and a half or two hours x-query work a week He just does it boom boom boom He did a little side project for us something totally unrelated to this something outside the university and he built four x-query modules in like 90 minutes And they work no bugs Just work it's easy So we like it We're big fans and it runs it just keeps running Um And that's it That's my wordly slide at the end. So questions Oh, yeah, yeah, yeah Oh, oh geez, um I wish I had a picture So, um Check the time Okay, we have plenty of time so uh the digital libraries group Is that we have two kind of digital Information groups that are not public web We have the digital scholarship lab, which is about digitizing things and analyzing it from a humanities standpoint They just got an eight hundred thousand dollar grant from the melin foundation to build an american atlas From the early 1700s through the late 1800s All maps and it's Some of it will be in existdb the maps themselves will be in s3 We also have a digital library group. They digitize stuff and make it accessible Is the easy thing to say about it Their first big project they came to us to help style and then to teach them how to use exist Then we built some visualizations in it virginia 151 years ago 152 years ago Voted to secede from the union And at the beginning of the civil war Somebody got all that information together and in the early 1960s some editors published a book We got the rights to use the text from the books Copyright permission. It was all digitized and turned into tei Encoded documents and we built a discovery system on top of it for the president And they're learning new things about why the vote actually took place It's not always obvious and it's got density maps and it's got all kinds of visualizations. So Collections.richman.edu will redirect to it if your interest is really cool We're now working with them on We have the largest collection of japanese War crime tribunal tribunal documents in the world. They're now digitized, but they're not online yet That will be in exist tb They also have the digital dispatch, which is the local newspapers called the times dispatch It was originally called the dispatch And they have I don't know 100 years of dispatch Publications of the local newspaper and they're all tei encoded And they're that's a current project. They're going to be put online with a discovery layer on top of it. So And exist just makes that easy. We don't even think about the exist piece anymore The hard part is just getting all that content, you know and encoded in tei and making it look good The designers want to make it really look good So is that Yep, I'm really glad you asked that question because I just told you didn't even think about going over there So and it runs on a different server just for isolation purposes. Yes, sir Um exist is um is an open source xml database. So it's it runs as In most cases you run it as a service. We run relinux based. Um, but we also run it on the mac for You know for development purposes, so it's written in java It's got an api and it's got it's wrapped in jetty. If you all know jetty jetty's like Kind of a embedded Tom cat ish like a java server page, right? So it's wrapped in jetty and so it makes web stuff with it really easy. So Okay Other questions Yeah, you can boost search you I mean you can customize it so That you know, for example, if some if some phrase is found in a title You can say that's more important. Um, then if it was in the body You can do you can actually do proximity searches. We don't do too much of that We love it. Um, it's a full text search engine The only time we have problems with it if is with the directory of somebody's got a name that's in the stop list So we have a person whose last name is will and will is in the stop list We have to remove will from the stop list Um, otherwise they they never find will in the directory, but minor details, right? That's why we're paid You know to figure that stuff out. So All right, any other questions? No All right, thank you all