 Good morning. Thank you for being here and your interest in Georgia's implementation of Drupal. Well, Georgia definitely made some news about five years ago when we decided to join the Drupal community and not just for a site, for a couple sites, but for this entire web enterprise platform. So it was pretty exciting right back then because we were embracing something that was just catching up in the federal space but none of the states had done that or even the local governments. When we did our research there were a few instances that we saw where Drupal was being implemented but nothing to the scale of what we were looking for. So it was pretty exciting for us as well. And back in the day we kind of had done some projections based on our solution that we were proposing internally within the organization and we had talked about it out in the open so clearly there was like a lot of expectations and last year or so I started getting a lot of calls from other organizations, other states, local governments saying, yeah, how's that going? Back in the day everything was promising rainbows and unicorns. I'm pretty sure technology project, come on, how's it going? So I figured this would be a good time to just kind of start talking about what we have been going through, what we have been doing, what we are planning on doing going forward. So again really thrilled to have this opportunity to come here and just kind of circle up and talk about what we have been doing and what our ROI has been going to Drupal. So I'm Nikhil Deshpande, I am the director for Georgia Interactive within the Georgia Technology Authority. So Georgia Technology Authority is like the technology arm within the state of Georgia and any other technology based consulting or technology providing that needs to be done within the state my agency does. But I'm not a Drupal developer or a Drupal person because back when we decided to kind of go Drupal I didn't know anything about it other than the fact that hey this is something that other organizations that are like us are using as a solution. I'm also not like a financial person, I'm not a financial analyst, I'm not a number cruncher even though part of my talk is going to be about the whole ROI aspect of it. But definitely what my job is is to find the right solution. So I'm a strategist, I'm a digital strategist and my background has been in user experience. So looking at what our users need and how we are serving those needs, it's my job to find the right solution. And this is kind of where I was put in this place to be the link between the solution group and the suits. So got to have to talk like both languages to connect the dots and make things happen. One of the primary things that we do in our team is have Georgia.gov up, live and relevant. And just to kind of give you an idea, I know that some of you have connections and ties to Georgia but Georgia is like a state of about 10.2 million residents. And obviously everything is centered around Atlanta being the capital as far as the state government is concerned. But Atlanta as a city is just about 445,000 people. But that's just the city of Atlanta. But the metro of Atlanta is about 5.2 million. So as you can tell right there, there's a lot of government that happens around that. Like there are different cities that form metro Atlanta and obviously there are different counties. If you guys don't know Georgia is like the second largest in the number of counties after Texas. We have 159 counties. That's a lot for the size of the state, right? So clearly there is a lot of confusion when it comes to how do I deal with the government? And that's where Georgia.gov comes in place. Like we don't want people to understand like if this is a county function, if this is a city function or a state function, you go here. If you are a resident of Georgia, you go here and we will give you at least the basic answer that you are looking for and where you need to go for this. So this is a content issue and this is not a technology issue but this is essentially what our vision was. And then that vision also was that based on this, because we are being like a centralized place to answer our constituents' questions, we also need to have a centralized place for our state agencies to host their web presence. So this is not new. This was put in place back in 2002 where we were all centralized IT for the state of Georgia. But only back then we had logins to multiple HTML sites and we used to update static content. And in 2002 we moved to a CMS system and that's when we started making sure everything was kind of structured and everything was a little cohesive but it was again 2002. It was not an open source CMS. It was a proprietary CMS. But the issue here is that these services that we offer are not mandated. I know that in some states and organizations if you have a centralized IT, all the rest of the organizations are required to use those. We are not. So we are essentially competing with anybody. We are competing with IBM. We are competing with Accenture. We are competing with the person sitting in Starbucks who can pull up a website and who can create a WordPress site. So it was fine. I mean, that was the premise. That was our challenge. So we had to find a solution that was scalable. That was elastic enough that we could be competitive enough. We could be competitive when it came to web publishing platforms that we created for our state agencies also for the professional services. Because agencies don't have, you know, UX people. They don't have content specialists and strategists. We do, but we work with them. We work with them in partnership to make sure their web presence is really up to speed. It's per the best industry standards. And as much as we act as a vendor and a partner with our state agencies, we can also have a centralized visibility into the overall presence. That also gives us a little bit of a governance aspect of what should we be doing. So we not so much from a dictating perspective, but we bring people together as a community and we work with them to kind of make sure, you know what, these are the best practices. Please don't use Flash. You know, please don't use like scrollers on your website anymore. No one really cares about like the count of how many visitors have you been, have you had on your website. But okay, it's important to you, they don't care. Take that, take that down, please. So anyway, things like these, you know, we work with them. So the structure for this presentation is going to be pretty simple, right? So it's just going to be why Georgia moved to Drupal. And if some of you have been in any of my previous talks, you might have seen a little bit part of this is because this was essentially the drive for us to move to Drupal. Then I'm going to be talking a little bit about like how we are using Drupal right now. And then, you know, essentially the promise of this talk is what has been the ROI for us using Drupal. So going about like why Georgia moved to Drupal, I'm just going to do a quick flashback of 2011, right? When we were in the middle of a technology crisis, right? I mean, you know, the word, the rest of the world was celebrating the British royal wedding and, you know, there were like a lot of other milestones happening around. But for us, it was a little different, right? We were using a pretty old CMS system and, you know, our operation costs were like through the roof, right? Because this was set in place in 2002. So our infrastructure, which was almost like nine years old then or 10 years old, it was kind of crumbling, right? I mean, how long can servers just go without replacing parts? Our entire focus was just to make sure things don't break and things don't catch on fire. So there really was no strategy and that was hurting us. And in the return, it was hurting our customers. Because one thing up here is we have customers because we are centralized IT. We charge our agencies and we offer our services as a legit service offering that we can compete with anyone they can find on the street. So our customers were kind of unhappy. They were a little kind of concerned like where is this going? Because clearly their needs were changing as the time was passing and it affected my team, right? It affected everyone on the team. We were like, what are we doing here? Why aren't we just doing something different? So the cost aspect of it was essentially because just a couple years ago we tried to upgrade our CMS to the most recent version of what we were using and it was just bad. It was terrible. So we moved some sites up there and when we saw how bad it was, we were like, nope, we're staying back. We're not moving the rest of the sites to the new version. And then guess what? We had two versions to manage. I always equate this to Napoleon's march to Russia where they had no idea where they were going and by the time they reached there winter sucked in and everyone was just shivering. They were not used to that kind of winter. They were like, that's great for you here but let's go back. They started out with some half million soldiers. They went back with 10,000. So it was really crazy the way things were happening back then and also this was not an open source system. We had to pay licensing. We had to pay support every year and that was always just going up every year and that was beyond our control. We also were in a self-hosted model. We hosted in our own data center. We were part of a huge set of servers which had super specific requirements but they could not cater to their services just for our requirements. So we had to like check a bunch of boxes that were just not relevant for us. But we had to pay everything because that's just how the pricing model was. So clearly that was also a big pain point for us because costs were concerned. And of course like I said because we had two different CMSs we had dedicated support staff because the two versions even if it was the same CMS were so significantly different that we could not just upgrade our existing support staff to support two different versions so certainly our development staff had doubled. So clearly this was the cost aspect of it but also the infrastructure aspect. Like I said we had 27 servers which were 10 years old and we can even have a good experience with cars that are 10 years old without us having to fix things and everything sometimes like this comes down, this blows up, change that. We had 99% uptime. So for someone who doesn't really know uptime they're like 99% come on, what are you complaining about but that means you're like at an annual base like down for three days. Every day you're down for like 15 minutes that's just not acceptable. I mean for us, I mean for a state to have sites down 15 minutes a day that's just impossible to defend. And also everything was managed so when the server went down we had to open a ticket and then wait. So there was like this feeling of helplessness we want to do this, we want to change things and we're just not going to happen. And clearly the signs of a sinking ship started showing our customers were like, sorry we've got to go. Joe is going to build a site and he's just going to manage it. Yes he lives in the basement but he's still going to be managing our site. So we started losing revenue. The team was kind of now split between half people thinking we should do something different and the people who were invested in the technology that we were using, they were like, we can make this work, we can make this work, let's just stay the course. So overall there was this huge divide within the team and the morale and the engagement, it was split. So it was a little concerning there. So the next step was obviously to look for a CMS. So this is not going to work, we need to find something else. When we started looking at other CMSs, I mean, oh my God, that's like, what do you pick? Because right now we were so heads down on making sure our CMS doesn't blow up. There were so many other things that came up while we were just busy putting out fires. So we went to Gartner route and we've seeked advice and everyone pointed to like, well you should be using these because these are clearly magic quadrant. And I was like, wait a minute, what we are using is in a magic quadrant. It's like we need to do something different because granted that CMS might work, it's just not working for us. So we had to find something that worked for us. So at that point our strategy was just keep the lights on. They would just like make sure nothing blows up, keep the lights on because we had fell to this whole sunk cost fallacy. The leadership was all about, well we have invested so much in this CMS, make sure it works, just keep doing something. We don't want to invest in something else, just keep doing it, keep doing it. So essentially back then our strategy was Farmville. Back in the day it seemed like the right thing to do. It seemed like what else can you do? And no one was like actually open about, yeah I'm on Farmville, who can say here they've been on Farmville. No one raises their hand. It's like there's something that you are aware of that I've been here for so many hours, I've got so many crops to harvest. You keep going back, you know it's a time suck. But still you keep going back. So yeah, clearly that strategy was not to work. So looking at what the landscape looked like, it was just like all kinds of CMSs, all kinds of focus that we were looking for. But clearly we had a very specific filter system that we wanted to go and find because it was going to work for us. And the biggest draw for us was to find something that was not expensive, that was not Clujie to manage, but also something that was very well supported. So obviously the fact that I'm here we ended up using Drupal. And these were the four main filters that we applied. We had to find something which was robust enough to support our enterprise. We had to find something which was cost effective because we did not want to go back to the whole cost model that we felt like we were being held hostage to. And we wanted to make sure that this was user friendly. And we found Drupal extremely user friendly because again, our user base was not that sophisticated enough to handle CMSs that required them to be at a certain technical level. And obviously we wanted to find something that had a huge government presence. And Drupal fit the bill for everything. But there was just one issue with that. It was open source. And back in 2011, still at the state level, people were not using open source at the scale we were trying to implement this. So we had to do a lot of education about open source and people started asking questions like, come on, I'm an open source. How good is it? I know everyone can just do whatever with it. It's open. How secure is it? But wait a minute, it's free? So no money at all? So everything that we are spending, everything is saved? So there has to be certain education done. So we did all of that. We answered all of those questions. We put together a business case, in which we kind of did a pretty decent job of explaining that yes, open source free, not like free beer, it's more like a free speech. But honestly, if any of you are in the position to justify this, I don't think, I mean this is like way past, so I don't think we should be having this conversation. But the best analogy that I've found for open source was from Scott McNally. He said open source is free like a puppy is free. Yes, it's free. You can just go out on the street, pick it up. Then what? It's not like you just come in, just put it on your mantle and it's there. There is a lot of nurturing and care that needs to go in it to make sure it grows up to be a dog for what you want to walk and what you want to have in your house. So it has to have that environment that you build to nurture open source. So now that we are on open source, I'm going to talk about how are we using Drupal. So we did bring that puppy in our house. So the next thing we wanted to make sure that we nurture that puppy. It's not just enough to go and procure open source. So we built a platform. We built a platform that we started treating it as a product. We made sure that the platform was single code base so we don't have to duplicate effort every time something needed to be changed or upgraded. But we also had a multi-site and a multi-database structure in place. As much as every agency who was on that platform were happy to be on a cohesive platform, they still wanted that wall in between them. I don't want them to look at my things. You're going to be publishing that tomorrow anyways. But still, today, no, that's my thing. That's fine. These are some of the challenges that as a centralized IT, all of you must be facing if you're in that position. We also made it theme-based. So at least there was like a starting point for people to relate themselves to. The themes constituted our different look and feels. So it was all about, all right, I'm the governor's office. I want something red, white, and blue. But hey, I am community health. I want something soft. I want something friendly. But I'm GBI. I don't want something soft and friendly, right? I'm the Bureau of Investigation. I want something that's absolutely official, something that is a little bit of scary looking. So we had a bunch of themes that we offered to our agencies because we were moving 55 websites at the same time. We had a timeframe of just 10 months to make that happen. Because our infrastructure was honestly so bad, our licensing was expiring, and we had no appetite or the budget to renew it for another exorbitant amount of money for just next year because we couldn't move those sites fast enough. So we had to build something that could just consume that content and build a site fast enough and let our old infrastructure just die out. But also while doing that, we were really concerned about are we going to be responsive to the changing landscape of technology? Are we going to be accessible enough for all of the users that we were not so far? I mean, it was really shameful that we had like a M Dot, like a mobile site for our rest of the websites that just showed our site map because our old technologies did not accommodate what the changes we were trying to do to make our older sites responsive. So we had to make sure this new platform accommodated everything. So even before actually we started making any technology decisions, the first thing that we did was we had to change our strategy. We had to look at it from a very different perspective because the beauty of Drupal that we saw was that it was actually a solution to fit our strategy and we didn't have to fit our strategy to a solution. So our strategy always has been users first content first and mobile first, right? But in three different ways, not everyone can be first. But in our planning stage, we always had users first because anything that we are doing, we have to be aware that it's end of the day for our users. Anything that we are designing, we have to make sure that we are designing for the content. We are not designing for the whims and fancies of the leadership of the commissioners or the directors, right? Just because somebody likes blue or somebody likes red, that's great. You know, paint your rooms, bedrooms or living rooms with that color. But for your website, if your users are expecting something that they are looking for a certain content, I mean, yesterday the keynote that Sarah said was like so right, right? So many of our Georgians, users come to our sites in so many different mindsets, so many different situations that at that point, if we started being cute, that's the last thing they want. All they want is where they can go to the nearest office to seek the service that they are looking for. So we had to take a very user first content for a strategy, but at the same time, while we were doing that, we saw our technology landscape changing. So Mobile First was imperative for us to do that. But at the same time, we didn't know what to do. This was 2012. This was 2011, actually, because we were part of the whole planning strategy, the whole part of what do we do from here? And the responsive design article had just been published a few months ago. So we were still in a kind of research mode of what do we do to make our sites responsive. But then we realized, yes, technology only can drive this solution, but unless we change our business model, unless I change my entire team to support this, it's not going to work. So earlier, when my team was just pretty much divided into people who were managing content, one version of the CMS, second version of the CMS, now it was more of a very overall model of like a 360 model, where our team had two arms. One was the product, the other was outreach, because there's no value in having an awesome product if there is no outreach for the community and if they don't know how to use it. And like I said, our users are really not that sophisticated except for a very handy few. Most of our agencies don't even have dedicated content managers. Content management is like a third or fourth job for maybe like some admin or somebody else who just happens to know something about computers more than the others. And that's just how it has been. So it was our goal to make sure we educate them about this is important, this is what we're putting together, but this is how you use it. So just build it and they'll come, it's not going to work for us. Just talk to them and keep telling them about what we not just have built but what we are also going to be building and to take this further, we need their input. We want to make sure we build this together. It's not a centralized model where we just are physically detached from them. It's a centralized model because we can all reach out to them and we can work with them. So a lot of the stuff that we are doing, our agencies have a huge input in that. So clearly now instead of just saying, hey, you know what, this is just a technology that we are giving you, this is more of a SAS model. This is more of a service. And this was more of a solution for them where you don't have to worry about anything. You don't even need to know HTML. That was the beauty of Drupal which made it so user friendly. Part of my team was doing research and development. Part of it was doing education and training and then part of it was obviously doing marketing and making sure that the word gets around. What we did was obviously used a distro, a Drupal distro open public and we then started making that into a product that was essentially catered for the state of Georgia. So that means it has like all the sophisticated product milestones that comes with developing a product. It has like all the development sprints in it. It has all the bug fixes that are constantly done. So we don't wait until something catches on fire and agencies start opening tickets saying, well, this is not working anymore. We no problems ahead of time. We had like now an internal set of people who are just dedicated to doing testing and focused on features. What are we doing? Where are we going? Not enough to keep the lights on. It's just like a very small piece of what we should be doing. Going forward, we need to make sure, we need to feed and maintain our system and that's kind of like where we started moving towards. So we did a monthly platform release schedule. We had like a very hard date every month and anything that we can accommodate in it, we will push it at that time. So that also kept us on our toes as far as meeting the schedule, trying to push and accommodate as many changes in it. We also had like a bunch of other testing tools like we had manual testing, AB testing, we had automated visual regression testing, which otherwise we would not have caught. Little things, right? There are some things that just happens. There's so many ghosting machine scenarios like sometimes when we push something, all the dates for our tourney general suddenly switch to the date of the push. All the dates for the opinions, right? So imagine like, obviously everything is so formal and official up there and suddenly an opinion which was like 10 years ago starts showing today's date, like they would flip out on that. But things like the visual regression testing tools immediately will catch that otherwise we have a blind spot for those things. So we needed to make sure like our testing effort was sophisticated and also complete for catching these things. And obviously there was a certain aspect of manual testing. If there are any regression issues we need to go and identify before we push those out. And then obviously the module of upgrades and code audits, right? Even though you keep chugging along at some point, like every year all of us go for a physical just to kind of have that overall view of how our health's doing. We need to have that occasionally. So we started doing that as well. So this is essentially just to kind of maintain it to make sure nothing goes below of what we already have said as a starting part. But we also wanted new tricks, right? The puppy needs to learn new tricks. It's not going to be fun otherwise. So this is where our wish list came in. This is what we wanted to do. This is what our agency customers wanted to do. And we started making the wish list. Like what do you want? Not everything that they wanted is something that could happen but still we want to make sure that they tell us everything that they want. And once they do that then there is a certain way of us to make those into enhancements. If we need to go reach out for new modules if we need to create content types so be it. But we had a screening process in that and we communicated that process with them, right? So going back to our overall strategy if it's putting the users first in whatever the wish list item is that's a good thing. Then it moves further on. Does it lessen the load for our content managers? Because at this point we have like a little under 500 content managers across our enterprise that are using this platform on a day-to-day basis. If it helps them that's huge value, right? I mean overall, does it provide value or are they just trying to scratch the itch of just like two or three people? Just because they want it. If that's the case then yes, sorry. We can't do that. It has to have an overall effect. And also is it future-focused? Because things are going to change, right? Just because something's needed today doesn't mean we invest so much time into developing it. So we have to make sure that it stays relevant even in the coming years. What that happened obviously is it helped us then put it on a product roadmap. So everything that we did the wish list, the prioritization, the scheduling of whatever made it through that we made sure we communicated. We added that onto our roadmap which really kind of helped us now where we are standing looking back it helped us with the over-critical path. Because if you just take a quick look at where we started, right? In 2011 we were designing that platform. We were designing something for yes, this is my ideal of where we want to go. But also we were coming from such a dated environment and it was such a huge shift. We also didn't want to make it into like a night and day change where the training efforts are just going to be overwhelming. So in 2012 we completed the migration. We had one entire year to stabilize the platform. But that's when we were planning to make something happen which we did the next year. We made the entire platform responsive. Within a period of months we changed all of our sites to a responsive design model. Back then I guess we were pretty close to about 70 sites. Because in 2012 we just launched Georgia.gov as a responsive because that was a new technology back then. We wanted to see how it works. There were certain compromises we would have to make making something responsive. So we knew there was some learning effort behind that. But once we did all of that once we had our lessons learned we started applying it towards the rest of the platform. Last year we did a huge push for accessibility. We did have some level of accessibility baked in but then we also made sure that we want to be absolutely accessible as a state. We are not required to. We are not required to be accessible but that does not mean we should not be. So we partnered with Georgia Tech. They have a dedicated accessibility division within their organization that just does this. They have disabled users working for them. Part of it is not just relying on tools but also having real humans interact with your sites and then takeaways and then we made those fixes back then. And then of course we do have things on our roadmap. We do have like now it's been a while everyone is asking for where the design is going now. We need a new design. So we are actually instead of doing a very specific focus design effort we are opening up over design solutions to making it more pattern based. Making it more team based where earlier our teams were really locked into a certain look and feel. Now our teams are opening up where we could actually choose different look and feel as far as colors and fonts are concerned. So we are looking at a UI redesign. We are looking at different content types. We are looking at what we can do with data visualization because now that the rest of the platform is stabilized agencies are asking for like oh this is great what can we do now. We have a bunch of data that is sitting in PDF files it's sitting in Excel files. We would love to do something else or also reaching out to them saying hey why do you have all these Excel files you can just map it out. You can just make it such that users can interact. They can actually apply filters and seek the data that they are looking for specifically. So when we got the needed nods by agencies we know we have a community now we are looking for ways to do that going forward. Also this is like user facing websites that we are using Drupal for. Now that agencies are having such great experience with that they are like hey can we change the internet to Drupal because we are using this amazing product that everybody uses but I don't think anybody likes it. So we want to change it. We want to make sure we like this part of our web experience. We want to do the same thing for our internal experience. So there are a lot of things that are happening and I'm just kind of glad that we are able to accommodate that. So our next challenge was like I said the old CMS that we were using it was hosted in our own servers on data center using our own servers but with Drupal should we be going by the same model or do we do something else? So the next challenge for us was where do you house your puppy? Where do you put that dog house? Should be in, should be out should there be a little flexibility? So there were a few hosting decisions and this is where organizations have very different philosophies about hosting. If everything that they are used to is no no no this is all PII you know needs to be in our data center you know the first thing I had to say was like nothing we do is PII there's no reason why we should be still hosted in our data center and everything something has a hiccup I open a ticket and wait for hours for someone to respond just to say okay I saw your ticket I'm going to find someone who can work on your ticket and next day somebody who is found to work on my ticket emails me saying yeah I got your ticket I'm going to put it in my queue right yeah it just I don't think so so we definitely were aware that we needed a better uptime than 99% right 15 minutes a day it's just not acceptable there was also the security aspect but that's okay right I mean you know all of our sites are either FISMA moderate or FISMA low and there's like nothing up there that requires me to have like a tier 4 SSA compliant or you know Fedran compliant data center if it is awesome but if it's not I mean it's not like you know that's something that's going to stop us so back when we started doing this there was a little internal fight but I'm kind of happy to say that we were able to just take everything out of our data centers and put it up on the cloud and that just changed everything that aspect of it literally changed everything for us it freed up a lot of my time actually gave me some time over the weekends to spend with my family but end of the day who does it right my team is 10 people my team is 10 people we have 500 or so content managers we have about 80 sites we can do everything we can do everything so back when we were looking for Drupal as a solution we also started looking for you know we need vendors we need vendors to help us do this because Drupal is not our expertise we need someone who does this for a living and if you're an organization in the same situation again where if you're thinking if you can have internal resources support this or if you have you know vendors who can do this for you give this give this a deep thought and do your analysis over and over again and also factor in all the risks that come with both situations both scenarios in our case you know we were lucky enough to find like a really good set of vendors our RFP we had multiple responses but the response that won was a collaboration between phase 2 who was our primary vendor they had partnered with media and media current we did our research I'm not going to lie we did our research we looked up everything that they did and not just this set of vendors everybody else who applied for our RFP we did our research we made sure you know their working styles their inner working styles the other projects that they are worked on kind of you know were aligned with the way we like to work in our RFP we were open to ideas but we were very specific about what our requirements were after everything was done right one of the conversation pieces that stuck with our team and actually I was reminded like a couple weeks ago about that was we were told that you know reading your RFP we could tell that you're not just looking for the IBMs and the Accentures of the world because essentially that's where some of the other vendors are like well you know this is clearly not something we could compete against because you're not looking you don't maybe know what you're looking for and if it comes to somebody else telling you what you're looking for and then saying I can do this for you you know we know how those projects end up so you had to be pretty specific if you are in a position to write your RFPs and references are crucial because there are a lot of things that you would want to know about and then you would want to know about the vendors that otherwise would have been some it's not good or bad but it just it's just a fit right because as much as I say these are vendors because they responded to our RFP they are our partners they are not someone who we just like toss our requirements over the fence and then they go develop and you know purchase back the solution now it's a collaborative effort when we talk to them a lot they are the extended part of our team so going forward you know the model the product side of it the product side of it exclusively worked with Phase 2 and Acquia because Phase 2 does all of our day to day software updates and maintenance and then we do over hosting with Acquia again our servers hosting internally a lot of headache at least in our situation because a government organization right I mean we are pretty vulnerable to anything not just that the servers were clunky or whatever I mean you know we constantly get d-dost if there is something that happens in Georgia that attracts international attention sure what's the first target georgia.gov go take it down or point it to something else and you know have a message up there about like why we disagree with whatever you are doing so we had to make sure we were pretty immunized from you know that kind of threats and that kind of issues so every time later that we saw that we were we had a chance of getting d-dost all it took was a ticket open to Acquia saying hey there is a chance this might happen and immediately they were like alright we keep an eye on it and in many cases they said we had to double your you know infrastructure to kind of make sure that we could accommodate that level of traffic before we could just make sure we block those IPs it was a matter of 20 minutes and I can tell you how much of a value that adds to it just the fact that you open a ticket and both your vendors jump on a call to make sure nothing goes down and everything stays up and running it's amazing it was like a totally different experience from what I was used to in my older model so because this was all taken care of my team now could focus on other things right we could be making sure that we could be focusing on training we could be focusing on the governance aspect of what we were doing the knowledge base building of a knowledge base every time we had any enhancements rolled into our product we had release notes that informed our agencies okay this is happening remember you ask for this it's done go check it out if you need training we have a quarterly training sign up for it we'll teach you how to use it if there is something that's big enough we would schedule something special for it go sign up for it we do webinars we do training like on campus as I would call it or online we have office hours where if someone is struggling with one little thing they don't have to sit through the entire training they just come sit one on one with us and we train them twice a year we also do an event called Gov Talks it's kind of tailored after the TED Talks model there we just focus on certain aspect and then we have like 15-20 minutes talks about that one thing so we've had Gov Talks about mobile social media data usability content all these issues which are not really technology but just best practices these are like industry issues where everyone is going and we don't want to be as a government reactive there's no reason why it is you know government to be reactive to any others when a lot of the other aspects and a lot of the other pieces of government really are like ahead of technology you know compared to the private sector so we make sure like everyone comes in and they understand about all the industry best practices what the standards are we keep working on updating our state standards and once we do that for example like a few years ago we updated our accessibility standards saying like everyone has to be accessible if you are part of the state government and you know that's led for the entire platform to be accessible we talk to them about what's happening with our platform our training issues right so amazing stuff now that we can actually focus on because we don't have to worry about servers being down or you know a backup didn't happen so the last part of this presentation is alright so what's the ROI right what's the ROI here so the success criteria obviously for an ROI is did it save you any money what how did it how did it affect like going forward as far as your business is concerned how did the conversion happen and then you know what is the efficiency here honestly when we did our business case all we focused was on the cost savings like this is how much it's gonna cost us back then we said it was gonna cost us five million dollars over the next five years and that's a great message for the suits right so like wait a minute five million dollars alright let's talk so that was our key to at least get the conversation going but to be very honest with you when it comes to even cost savings it is not apples to apples it's apples to oranges because everything changed once we moved to Drupal right we could splice everything that we are doing and like one little piece of what we do now is what we did before but everything else that we are able to do now because we just had this technology to take care of everything that we used to be able to have nightmares and sleepless nights so just to kind of give you a glimpse right of what our 2012 operation model was that between our support staff over hosting cost licensing and professional services my total operational budget was one point five million dollars every year we were told it was gonna have a ten percent increase as far as the licensing cost was gonna go also sometimes maybe for our hosting and support I had four dedicated Java developers making sure my two versions of the old CMS don't go down so my support staff actually as you can see was the highest and then professional services is something that we had to get in case something goes wrong which was not covered under their regular support which almost nothing was so every time something happened we were like yeah that's not actually a support thing we're gonna have to change this because the implementation blah blah blah blah blah professional services money that was actually something that I had put towards future facing issues was getting used up and just keeping the lights on jumping back to 2015 my cost of operations the biggest difference was it was a little over half a million the support staff was significantly low because we have just one Drupal developer on site everything else is our professional services which happens with phase two we have a certain set of hours that we go through every month to make sure that accommodates our enhancements our features and if anything that goes beyond that clearly has to come from agencies as a request and in that case our agencies pay for it but zero licensing costs how awesome is that right I mean obviously everyone here knows that as part of the open source but this is amazing and next to each other crazy just crazy and this is just the cost saving aspect of it just the cost saving aspect of it because essentially it was a huge reduction it was about 82% reduction than the hosting costs 82% we just didn't need the rest of the peanut butter that came with you know the rest of the services that were offered to the 5000 servers that are hosted at the data center my 27 servers just need some basic upkeep and that's all I need I don't need to pay for everything else but it was just not the model that I could choose so my hosting cost dipped 82% no licensing fees one developer right so I mean honestly this really started proving like we might save more than 5 million dollars in the next 5 years but here's also the thing right this the old model would be at an incremental cost because we were definitely paying 10% more every year for licensing and support we were also paying more for hosting so if we just kept professional services way they were if we did not give a single dollar raised to our developers and hope they didn't quit this is what we would be looking for every year last year that cost would be almost under 2 million dollars just to operate our content management system but the next aspect of our ROI is not just the cost but also what happened with the conversion in 2012 we have 56 sites and we were losing customers the moment we implemented Drupal by the time we hit that stable state in our critical path we started getting contacted by agencies left and right because now they started seeing what we did with that they wanted a piece of that they wanted to be in there they saw how futuristic it was in 2015 we had 77 sites and this is just a site count a site is a site I wish I could say that if every site was exactly the same but it's not in the last couple years we have added a department of revenue department of labor we are working with department of driver services between these 3 sites I don't think there is any Georgian on this Europe baby that you don't interact with these 3 sites at some level you do interact with that so now we are overall touch point with every constituent of Georgia is humongous just to kind of give you a usage ROI other than just the site count in 2012 you know 57 million page views in 2015 it's 107 million page views our total sessions went up almost twice our unique users went up almost twice but our mobile and tablet traffic this is not just because of Drupal it's just the overall climate of how mobile traffic is exploding but still I can imagine having about 18 million or almost 19 million mobile users all of those 19 million mobile users would have a terrible experience because we are not able to cater to that and we are actually looking at some mobile users that are exclusively mobile that don't own a laptop or a desktop if they need to they have to go to the nearest public library because all they can do is afford a phone and a data plan it's not useful to them if they are looking for government services it's a big fail on us so really happy that this is happening now but these are the tangible costs these are the costs that we can actually quantify there are some other things too efficiency our old CMS, believe it or not it would take you 15 or 20 minutes to do anything per content push adding content or editing content because the workflow was so convoluted you had to click like a dozen of buttons go back, change a certain state push it again it goes in a queue make sure if something that needs to be published like right now you go and have like a secondary effort of making sure it publishes right then with Drupal under a minute it's email right it's gone so easy I can imagine standing here and telling you if everyone saved 19 minutes every time they wanted to do something with it I think we have some 50 or 60,000 content items that get pushed every month the amount of cost savings as far as just efficiency is concerned it's amazing it is just amazing but there were some hidden cost savings remember the servers I talked about that were like clunkers 10 years old so out of those 27 servers 3 servers died so we had to get new servers those 3 servers were 10 years old so back when we had bought them our licensing cost was based on the 10 year old server model which was just one CPU the new servers were 4 CPUs 4 our old systems licensing model if applied to those 3 new servers actually was coming up to $740,000 for just those 3 servers because the new servers now had 16 cores so per core license was just through the roof I can't even imagine justifying doing that for all of those 27 servers and thankfully sometimes bad processes help you when we had applied or rather requested 3 servers it took us 16 months to get them by then we had moved on to Drupal so by the time I got my email saying those servers arrived I was like thank you we don't need them find someone else to use them so $740,000? nah we didn't have to spend those so anyway concluding the value definitely outweighed our cost the value was this tremendous value of moving to Drupal just was tremendous as far as the cost was concerned but also you know if you are doing this regardless of the technology it applies to anything not just Drupal just to seek the value in what you are trying to do and not don't be focused on cost because there are so many cost effective and so many low cost solutions that otherwise you still be back to square one we stayed user centric user future focused and that helps us a lot and Drupal being Drupal and so flexible it really helped us succeed with that strategy and last but not the least take chances take chances because back in the day when we were proposing this solution I was told to my face that if this fails I am going to be out of a job and honestly I was not a Drupal expert but I just felt like if so many people have not been successful with this there is no reason why this will fail so it was a true chance and I had a great team that supported me I am just happy to be here talking to you but there is this huge team behind here huge team of ten people they are toiling making sure this is all up and successful back at the ranch it's amazing that they supported this vision back then we were trying to make this change but there is nothing wrong in taking chances making sure things work and then we invested in our puppy and then eventually when it became a big dog everyone was like singing praises like how do you do this and then how could we do this blah blah blah so anyway if you are interested in any more information about this project about what you just saw on our website interactive.georger.gov feel free to go there there is a bunch of blog posts that we write every team member that you know as part of this project has something or the other contribution as far as knowledge base knowledge transfer our training materials up there all over gov talks videos and everything up there if you think this has been any beneficial for you I would really request you all to just tweet to the Georgia Gov team I'm really lucky to be here talking about this but they are the ones who actually make it happen so it will be awesome awesome if you just kind of tell them that this project has helped you even a little bit in helping you in your strategy or your choices that you are making so by all means do let them know that this is going well thank you and if you have any questions yes so one thing I kind of forgot to do while I was doing the whole thing is I have a team member here with me Kendra Skeen she is our director of product and you know the half of the team that I showed you about product and outreach she leads product so right now Kendra is focused on and we are looking at obviously the D8 features and what's happening clearly it is not one side right it's a platform so there is a lot of research that we are going to be doing as far as moving to D8 we are also working closely with Phase 2 to make sure what their plans are as far as open public is concerned and I don't think we are going to be like the early adopters but at this point for D8 we would want to wait until things mature things settle down a bit but going forward I do want to make sure we are not like a very late adopter as well we do want to leverage the awesomeness that D8 is bringing but at this point there is no date at this point on the road map any other questions yes sure so I mean honestly Phase 2 had a lot to do with that but we selected Omega and then based on that we created sub-teams which kind of differed in the look and feel aspect of it but everything was driven from the base team Omega can you develop your own base team correct and also when we made everything responsive we made Omega responsive and Phase 2 was actually awesome enough to kind of get the person Jake who did the Omega team make it responsive for us sub-teams so we started off with 12 and Kendra helped me now we have like 16 16 right now because then DNR joined us and we didn't have anything friendly for environment so like alright we'll make you an environment friendly team somebody else had a question up here maybe answered them already any other questions yes go ahead Brian what about multisite strategy how many different sites run on how many code bases what's the process at a high level for vetting a new potential customer a new agency for a new multisite what does it look like to actually get them off the ground okay sure I can start off with this and maybe Kendra I feel pretty chimed in here but come in here and talk we still have a single code base so if any new requirement that comes in right now we try to obviously make sure that we are having the same model unless it's a completely different environment which is right now we are working on this project for a school report card and we are using a Drupal base solution called Decan and for that we have like a completely different code base now so that is not even sharing over same hosting servers but there is something specific that is needed we would develop that as a content type to give you an example we host the Georgia Bureau of Investigation and you know they have like these found remains of people missing persons listing blah blah we had to like create like very specific views content types for them but that just applied to that site that is not across the emergency content type that shows up on every site but we have to handpick certain agencies who can actually control that so if there is like another you know Snowmageddon in Atlanta then you know Jima or the governor's office or us can actually go in and add like an alert that shows up on like all these sites at the same time so there are certain aspects that we make sure that certain sites have access to but also none of it is like so different enough that it requires its own code base we are still managing everything together, do you have anything to add to that we try not to but sometimes you know the commissioners wants to look different right so yeah alright you get your theme but that's what we are doing now is we are going forward with more of an open theme approach which Kendra named it Prince not Prince Ali but Aladdin because it opens your world from what they are used to is they are not bound to colors and fonts they are not bound to like a certain look and feel they can pick and choose and we help them pick and choose so regardless of like the sub theme that we are using we can still have different sites look different using the same sub theme or like a certain sub theme all the sites created still have the same look and feel as far as the colors and fonts are concerned but just the modules that they are using and the placement is different that's what kind of makes them individuals no flake if you can there are I guess different expectations from agencies if they just pick and choose one of our existing themes then you know it's easy then we just build it we train them they start moving their own content this is something we really stress on our agency is you own content we are not really going to be we can help you with best practices of content but you own your content you need to really understand the whole content value so they move their content we help them with analytics we help them with strategy but you know once they move everything we do our internal testing we manage domains and everything and then you know we decide there have been instances where we have launched somebody within a month or within like our fastest was like almost two weeks once right I mean it was like a really small site for like a commission or something like that but there have been ten months almost a year now with the one of the recent ones that we are getting ready to launch because content has been like the most piece that we struggle with or if they need new functionality as you know what it was saying locations so if they are looking for locations you know we are not going to build it for them then we start going back to every other agency that has that need and we find out like what their needs are and then we kind of like put them together as far as like these are like the most common needs so our MVP really needs to have that in there then we can make it very specific to maybe just driver services or this particular agency but yes it can I mean it's literally like can take like a month or a year because content is so easy when we start the project and when we actually get to the content phase we are like oh who is going to be doing this so yeah any other questions yeah you had a question yeah so it actually did include the growth of the platform but in the previous life that consumed just keeping the lights on because the licensing and support model was such that any issue that we had certainly was like enhancement to the product so we had to spend a lot of our growth money into fixing issues but with the new model it is definitely you know a big part of our growth budget yeah we just have one Drupal developer and sometimes we kind of just divvy up you know the work between what she can do what the professional services need to do if any change actually affects the entire platform then we definitely have phase 2 in to make sure we have like the overall view of how it's going to affect but if it's like for example you know an agency wants a chat module to go up that is something we can easily do internally so we do that and before we actually push it we might request phase 2 to just take a look at like just make sure it doesn't break anything else and you know that's done then so our Drupal developers like 100% dedicated in making sure she's focused on enhancements which she can handle but some of the enhancements that are beyond her go to phase 2 but phase 2 is definitely kind of focused on the maintenance and the patches and the upgrades correct yes that will not be part of this operational budget that I would pitch it as like a key milestone within you know we need to kind of upgrade this and just the way we went through the initial migration that kind of costed us a little more than what our operational budget was but you just kind of do that within like one year and then you're back to the stable state well thank you everyone I really appreciate you all showing up