MMP7: Buy vs. Build, and why the two are not always mutually exclusive

Joseph Ferguson 
Developer, The Ohio State University


The audio for this podcast can be downloaded at http://2011.highedweb.org/presentations/MMP7.mp3


Announcer: This is one in a series of podcasts from the HighEd Web Conference in Austin 2011.

Joseph Ferguson: All right, so I'll give you a brief introduction of myself. My name is Joseph Ferguson. I've been working at the Ohio State University for a little more than 10 years. I started as a student and now I am a full-time developer, application developer with the Office of Student Life Technology Services. I've seen this department grow from a very small unit to now a larger team with a lot of moving pieces and processes that now help us achieve success.

So today we're going to be going over buy versus build and why the two are not mutually exclusive. Maybe a little awkward coming from an application developer but based on the history I've gone through with our department, I've learned a lot which hopefully can be helpful to anybody who's making decisions about buying and building. So, see if my little thing works. All right. I apologize for that.

 
01:02

So the first thing we're going to go over is identifying your organization. This is going to be kind of important because not everybody charges, not everybody builds for the time, not everybody has access to resources. So identifying where your organization sits in that will be extremely important. We're going to the pros and cons of buying and building. We'll talk about the environment at the Ohio State University and Office of Student Life. And then, we'll talk about some of the examples where we have built things and some of the examples in which we have purchased things.

And then, we'll talk about the many, many, many, many lessons that we've learned over that process. So hopefully, everybody would get something out of this regardless in which position you fall in. All right, so first part of identifying your organization is the, know what resources you have at your disposal. Some places have people, some places just have money, some places have infrastructure with servers, some places are really small and they have to grow. That's something that we actually went through ourselves.

01:57

So if you have money, you can get contractors. If you don't have money, sometimes you have to find students to do the work. All these are very important when making decisions about buying versus building. So another important thing is to think about the strategic initiatives or goals that your department or university has seen down, in some cases it comes from higher up and you have to find a way to get a strategy to make that look good so that everything you do looks good in the eyes of the people above you.

Sometimes that stuff goes university and campus wide but you also need to focus on what you do in your own office to make sure that happens. There are places where universities and colleges do not charge for the work that they do. It's important to think about that. Even if you're not charging for the work that you do, it's important to track the amount of time and the effort that's put forward because that will help when you are faced with decisions of purchasing and about how much money you're allowed to spend or what that's going to mean for your organization.

02:55

The next big thing is governance. And we've gone over that, I've gone over that a little bit but governance will help with prioritization which will help you decide which things don't have - you don't have the risk to actually lose or fail on. I either hope make decisions about choosing the standardized, something buying and buying a solution or just building yourself.

So everybody’s in some of these stages. Some of you may be buying, may be purchasing or you may have teams, hopefully you’ll get some information out of this. So those pros and cons, we’ll go over the pros first for building. As an application developer, I’d like to think that this is usually the best option where I’ve learned some lessons. First, again if you build, you always have the advantage of being able to manipulate these strategic goals and initiatives to best suit the organization.

03:46

So if you have an application that needs to meet certain needs, it needs to reflect the university in certain light to make sure that that can happen. Another advantage of building is customization. You can make it do exactly what the client wants to do. In our case at Ohio State University, we work with the, a group of clients and in some case it’s better for us to work direct with the client and get them exactly what they need. Sometimes that works best with the budget; another time that’s not the case.

But our control is really important. You want to be able to control the environment, the process, the path to upgrading, the time to upgrade. Sometimes that can be extremely useful if you’re building a solution. And the relationships, and I think this probably needs to be higher on this list but relationships are extremely important. I’ve been on the campus for more than 10 years now and I’ve worked with people who are students at the time. Again, the time I was a student and I’ve seen them move to other positions and have maintained those relationships.

04:44

Some of the success we have at the Ohio State University and the division of Student Life Technology Services is based on that concept of relationships. Those relationships also assist you with the user acceptance testing. That process is extremely important if you are going to customize to build a solution for your clients, make sure that, you can make sure that the client gets exactly what they need.

Sometimes the cost to build can be lower if you’re building. Not all the time. We’ll also use some example of those. And on universities in colleges another huge advantages of students, as a student wants – I know that I was cheap labor and it was perfect, the perfect time to build as much as possible and save as much money as possible, especially because we had great students in our division.

05:37

Another big thing is data security. We’ll go over a few examples in which data security could be a disadvantage if you’re purchasing because the transmission of data always leaves holes for problems. So if the data stays in house, starts in house, stays in house, so a lot less risk because never – no, there’s never no risk what’s a lot less risk. So some of the cons, as we all know resources, if we decide to send all of our resources to go build things, we won’t have time to take on new projects and we’ll always be supporting the old projects, extremely important.

Next thing is support which is always huge. If you hire a team of developers or you contract a team of developers and that team is no longer there, then you need to find somebody who’s able to support those applications. And the applications that you have built still need to be supported. So that concept of the life cycle of a project going from one upfront cost and then 20% of the same cost over the course of the next four or five years, that needs to be taken into account when you decide that whether or not you're going to build.

06:39

And then you have to kind of manage the resources and realize that, that same 20% is going to continue to pile up for every single one of those applications as you accumulate things that you decide to build. Time to market, this is sometimes the biggest drawback to why we decide to purchase items. A department contacts the house, the Student Life Technology Services department and we decide that we don't have enough time to do it. They want it by autumn of 2012. We can get it to them by summer of 2014 based on our projections and our Gantt charts. It's important to make sure that those things are known going into the situation.

So another instance in which tracking for your time is extremely important and being able to forecast, support cost and the effort required. And again, support and replacement cost, replacement cost is huge. Everything has to be replaced at some point. So having a projection, the ability to project when something is going to need to be replaced is extremely important. So, and infrastructure costs, if you don't have your own servers, you don't have your own support team, you don't have your own server squad to put sites up, take them down, to run maintenance, this cost could extremely impact whether or not you decide to buy or build.

07:54 So you may look at a whole set of solution if you can't do it yourself based on infrastructure cost. So that's another thing that's extremely important to think about when making a decision. So now buying which is not really what I like to do all the time but I'll explain why. Time to market is usually one of those things that is so early. A lot of these vendors have solutions. Some of the vendors are right upstairs. A lot of these vendors have solutions. They meet you up and running or they claim to get you up and running in a matter of weeks. In some cases, it could take your department a matter of years to get the same type of product done.
08:31

Resources, if you're able to purchase a solution then that means your resources, the time, the effort that you get – may have to put forward to that project is now free for something that may be more important or maybe more critical to your organization. And standardization, in some cases if you're going to collaborate, if the departments you’re working with or your own department is going to collaborate with other universities, other campuses, other departments across campus, may not be a good idea to get something custom-built unless it's going to be open to the standardization so that you can share data.

That becomes extremely important when working with mobile applications when you have an open API, when you're sharing data feeds across the university, that's extremely important. And since technology is a moving target, buying could be extremely important. We don't have – we don't always have the resources to upgrade something every three years. There’s a security patch that needs to be done. There’s an upgrade that needs to be done. There's a new software. If Microsoft decides they're not going to support something after three years, this becomes extremely important to deciding how we're going to move. And if you purchase a solution, a lot of the times you get support built-in.

09:40

The concept of the community can be extremely important as well. Some of the items that we’ve purchased we were able to get access to previous clients to work with them based on getting information about how they migrate it to the solution, how they work with the vendor for support. And then cost, and as sad as it is for an application developer, sometimes it is cheaper to just go buy some something and you see that a lot with software as a service, you ‘ll see that a lot with host of solutions and the cloud solutions.

So next we'll talk about the cons, you don't always have the ability to control the strategic initiatives and how they’re looked at with the product that’s purchased. They can try to sell you and tell you that it's craft; it's a work within your strategic initiatives. You’d be give them an idea of what your department has planned. They can try to shoehorn it in there and explain it to you that it's working like that but it's not always a guarantee.

10:36

So another idea is a lack – another thing is that could be a con for purchasing is the lack of customization. It's not always a guarantee that you're going to have the ability to customize a solution and may cost you more in the long run. So some of the other things, our client expectations, it could be a huge problem. The client expects more size. Size is big. We’re a huge university. Sometimes we don't always fit well with an organization. Sometimes our IT department is four times as big as their IT department.

Some vendors have – just had one developer and we’ve had to work with those vendors and sadly we’ve had to realize that we could've just written it ourselves over time. And then salesman, salesman is support issues. These things are not always guaranteed, usually sold something, not guaranteed to get the same thing. Try to go through these as I’m slowing down a bit with all the explaining. But so one of the concepts is that you buy the standardized, talked about that a little bit and you build to compete.

11:32

I don't know if you guys are competing on campus, but if you are may be important to think about how you can compare. If there's a way to innovate that may be extremely useful to innovate when building. The trends as I said the software as a service becomes a trend. The costs are lowered and time to market is cut in some cases by 90%. Modular purchasing is another trend that we see. Purchasing, parts of a solution could be advantageous if you don't need the entire thing.

We on campus don't have our, we don't have blackboard for our course management on campus but we do have their ID card services. So modular purchasing could be useful. Hybrid solutions, I’ll go over that in a little bit and open source solutions, as most of you know, you get the concept of support with the community and you get a low cost. That's not always the best solution but that's another thing to think about.

12:29 As we talked about identify your organization, the pros, cons and cons of buying versus building, next we're going to talk about the environment at the Ohio State University, I forgot. So misprint out there; forgot the word of “the.” All right, so the first thing I like to explain is that developers are superheroes. I'm a developer, I consider myself a superhero. And before we left as – we left Columbus Ohio; we were in the midst of, the Mid-Ohio Comic Con. So one of the – it’s close to Halloween, so one of the ideas I have was, I wanted to give you my persona as a superhero.
  [Laughter]
  So this is what I'm thinking for myself maybe a little Batman flavor. I got a couple of ideas, one of – you guys are going to be my focus group. But this is just the reality is, if you leave it to your developers to decide what to do, I'm guessing they're going to tell you they want to build things. That's what they're there to do. We're there to build things. So think about that, I myself consider myself to be a superhero. The people I work with are superhero. The people that we purchase from are the villains in most cases.
13:25

So at the Ohio State University we have about 65,000 students. IT is kind of decentralized. We've got a couple of guys from Ohio State here at the conference today. We don't all work in the same office. I'm sure some of the larger campuses are like that as well. At Student Life, we have a diverse group of departments. We have housing, dining, food services, rec sports, wellness, the multiculturalists in there. We have tons of different units that we have to work together with.

We’re also a consumer, the data and the creator. So we have housing information, we share that with other departments and in return, luckily, I have actually listen to another session, I heard about promise about sharing data but we get a lot of data back which is very helpful for us. Hope this solve a lot of problems without having to pay other people which would be sad. We have six developers. We have about five student developers, about three project managers, a Q&A development. We have server teams, DBA, we have support center student staffed. They have about 20 or 30 students that are actually higher but there's always about eight or so that are available. That's extremely useful when you think about infrastructure.

14:31 So this is kind of built in to what we have and this will help you to determine why we made decisions that we made as I go over those. So next up, the second of option for my little costume, so you guys can help me pick. If not Batman maybe Robin but I don't know about the tights, the Batman costume had on jeans but well that number kind of changes because sometimes we find out that we still have access or we're still the owner of an application that we weren't aware of. We’re C sharp shot. I will let you guys kind of read over this because I don't want to hold those back too far area but we've grown into this. This is not where we started. We started, we had one website. Actually, we didn't even have that one website. It was hosted somewhere else locally and we kind of took that over.

So we started with just about nothing and now we've developed into this and it’s only taken us about, I don't know, nine years. So project governance has been extremely important over the last four or five years. That’s helped us decide what's most important and helps increase the transparency. That's extremely large when – becomes extremely useful when dealing with customers or clients that want to know information about what's going on and why their project has been pushed back, so one another thing to think about.

15:51

We actually purchased the ticketing system which is used by our support center that helps us manage support calls. That really helps us get students involved in the process so that we can have a student pick up the phone and explain and go through a series of tasks to work through a problem. And then, that ticket could be sent to one of us and we can actually give it to one of our students so that we don't have to stop what we're doing to build. The better the students are, you have, the better that works. At one point, Student Life Technology Services had a great student and now he's in front of you today.

[Laughter]

All right, so size matters. Size matters in more than one way. If you have ton of stuff to support, you have to think about what support time means to your organization. It takes us somewhere around 50%, sometimes I see myself at about 80, 90% of managing support tasks. We have a large client list. That means a lot of backlog work. That means a lot of governance. And we also have those, the president of the university to decide that this one thing needs to be done as soon as possible because this is the next big thing and that has to be done. That's something I always think about.

16:55

Now, the build versus buys comes up all the time and based on the, some of the things we put in place, it’s helped us with creating rules for this, making sure that we have those relationships and making sure that we can get back and response to these things. So if you don't have a good relationship with your, with the other departments or organizations this could be critical to whether or not you're going to be in the right position to make a decision. Sometimes we get called after the purchases have been made and we're told that kind of just make something work and that becomes extremely difficult.

So next we'll go over some examples of build solutions. First we’re going to talk about our SMS. So we build our own SMS or CMS, we call this site management systems, some people call them content management systems. In this, we kind of stuff in content management, sometimes application management, user, profile management.

17:49

It's built internally. We've been building in-house for a long time. I know there are ton of people that are going to try to sell it to us but we've been building for this long. I don't know if we’d ever going to diverge from that. Our clients are extremely happy about it. Early on we had less options. It's grown into something that we've all been able to use. Our customers or our clients know it and understand it.

We’ve worked with them to get the best solutions may. And the changes are distributed to all of our clients. So if we come up with a good idea for how to solve this user control issue or to make sure that we can assign permissions, we can actually push that to all the other departments and groups to have it. So here's a couple of more images of it. We manage our central calendar system in it. We have user management.

18:37

As I said, this is very much against norm. Not everybody's doing this and not a lot of people are probably going to do in the future. It's kind of a commodity now, something you can pick up. You don't have to pay for it. So it's one of those things where we made a decision to build and that we made that at a time when most people didn't know what to do. There weren't a lot of options to purchase at that moment. And when there were options it didn't work with our architecture.

So we made a decision to build and the deployment time frame for, from the time that we need the site to the time that it is done is shortened, until the customization occurs but we do have the ability to do a lot of customization. We kind of share that code. The codes kind of starts off in this kind of vanilla form, we put it somewhere and then we make it work for the solution we have. All right, like I said as client approved, the developers are pretty much cross-trained. When we hire a new developer, that developer is kind of putting the little corner and said, “Go learn what we did here so you can recreate that because, and that system, they need to work just like that."

19:36

That was what contractors – although we can kind of just push them in there because we all understand exactly what we went through to get to that point. So hopefully that helps us kind of standardize our code across all the platforms. And now support is extremely useful. I'm on – I wouldn't consider vacation but I'm away from office right now, so if something goes wrong I don't have to be the one that has to fix it which is extremely important. Sometimes when you build very customized applications that could be a problem if you just have one guy responsible for that.

Some of the cons are technology freshness. Right now we're trying to debate between sticking with Web forms or moving NVC and that's something that is, becomes difficult when you're hiring new people. It's expensive to rebuild it. It's not just as easy as just saying, “Okay, we're just going to turn the switch and make one happen.” Because you still have to fit this within all the other projects that are going on. And when you have all these other projects going on that are actually bringing in revenue to the organization, this is the one that doesn't immediately turn those results.

20:33

The life cycle alters our versioning. What that really means is that if we fix a problem with the way images are done or we had a new feature in one place, if it's been three years since we've replaced it on another site, it may be something that we either have to then decide to put in there as a support costs or we decide that it’s not going to go until the end of that life cycle when we replace that application.

So another problem with that is if we find a bug, sometimes that bug is everywhere. And we’re working on finding solutions for that problem right now but that can be a big issue. So as another example, we are a blackboard campus when it comes to our ID cards. Previously, we had a third-party vendor that we contracted to work to actually get money and put the money on the buck ID, return information about a balance. All of the stuff was done through this, through another company.

21:24

That was extremely expensive. It became more expensive as time went on. We had some turmoil with that vendor and we decided that we wanted to kind of rebuild it. Well, what this meant for us is, that we had to kind of make decisions about how we were going to get this done because we didn't just have the technology or the experience to work with the blackboard servers so we need to find a way to do that.

What we ended up having to do was get out of this contract which was very helpful, but plan to get out of the contract and then plan to buy some things. So we had to purchase licenses for new servers, those new servers were going to cost us some money. That money needs to be included when you compare those costs.

 

Contractors. We actually had to hire contractors to do this. They helped us with documentation and they helped us with testing, it helped us when that person was there but there’s also a lot of risk associated with that. So actually, I think what ended up happening is the cost to put together the solution for the first year equated to the end of support cost and the transaction totals for the service we are the paying for one year.

22:27

So it became, it actually became worthwhile when comparing. And now we’re working on, we're actually working on creating a solution for getting rid of that licensing fee right now. I think we're paying $15,000 a year just to have those service sit there and make sure they still work. So as I said, the cost was too high, time to market was not critical because we could end our contract at any given time.

Data security was a bit improved because we are working internally and we weren’t sending data back and forth. That was extremely important around the time of our assisting and conversion. Then we have this, service-oriented architecture became extremely important when we worked with the mobile application for OSU, for the OSU mobile application. And across campus it will be useful if there are other departments that want to work with the solution.

23:12

The con is the uptime. It’s all up to us. And we can’t get mad at them when it doesn't work because it's sitting in our closet. Licensee fees are there when trying to get rid of those now. And documentation and support became an issue at times but we’ve kind of worked through that. Actually, we ended up bringing back the contractor just explain to us, so asked what he did before he left and that helped us get through some of the hurdles.

So this is option three for the superhero costume. I don't know if anybody got in late but working on Halloween ideas. So I’m being a Wolverine. I don't know how the dreadlocks are going to work with the little Wolverine ears but hopefully we can put something together that is functional. You guys will have to help me pick what goes on next.

23:53

So we’d gone over the build solutions. Next, we’re going to go through the buy solutions. And this is where it's going to get kind of into the nitty-gritty. I’ve been told that because of NDAs, because the agreements we've made we’re still in contract with I think all of these people. I won't disclose any names and there are no screenshots because I don't want our legal department to call me. And now that it's being recorded, and I make sure I don't do anything wrong.

So what it comes down to is our – we installed PeopleSoft on campus. PeopleSoft requires to find a solution for the housing management system. Our housing and mill plants are all controlled in the same office. The system we replaced and I think was more than 10 years old. Cost of ownership had risen. Support costs started going through the roof. As we wanted to put more stuff in place, we then find the developers or find somebody to work with the technology.

24:45

Our client, the housing, I don’t know – actually, right now the university residents and dining services division, they’re our largest client. They make the most money. They have the largest pocket of all. They could have purchased any solution they wanted to. So we actually went to the process of trying to find a way to improve the workflow which is extremely important in their situation, all these moving parts and they have a staff of about six to handle all that information.

Now at times, it's really just a staff of one person. This one person that actually sits down in the room, puts out all the room, the housing agreements and tries to figure out who's going to work best with somebody who’s messy. We have 10,000 beds on campus. This is the task that should not be handled by one person but that's what's happening. That was what's happening in our office and actually could still be going on, she’d be getting away with that still.

25:37

So they went with the purchased product. Although I won’t say their name, will let you know that it’s an international company. It's a large company. They have some web interfaces and a desktop application. It's highly customizable, their strings, there’s tons of stuff that you can do with this application, had an ideal amount of training and setup included in the cost. I think the total cost of ownership for three years was planned to match I believe three years of support for the original, for maintaining the original system. But one of the constraints was that we had to get off the original system because of the change for SIS system, for our student information system.

So time to market was great. This freed up a lot of resources or we thought it would free up a lot of resources. We did hire some new people for this and increased standardization in area where it was extremely important. We had a client and vendor relation that actually needed some work but it was good because they started from the beginning on great terms.

26:35

The vendor would actually fly in from whatever international country and show up and train our people or explain to the department what need to be done. Getting set up was perfect. We had access to the data. This is something that I can't, I cannot emphasize it enough. Having access to the data means that you have a copy of it. If you have some access to the data that’s in that system, it could be extremely powerful for your department and for your group. So we’ll talk about that a little bit more.

One of the cons though was that time to market meant that it would be up before our SIS conversion came out so that meant that we’d also have to do some migration and moving information from the system that was put in place. The scope was incomplete, there's no way that he could have determined how many different pieces and moving parts this organization had, there’s reporting, there’s exception, monitoring, the shortcuts that we had to make. So we had to do all these custom additions.

27:31

We’ve really built an application that sit right beside that application. On some ways it’s really good because it gives us control to kind of ease the burden on the client and hopefully help out that relationship between the vendor and the client. We put something in the middle that helps us keep them out of the system and can help us kind of put in some monitoring place to avoid big problems.

Another thing that was a con at the beginning but became, actually they become a pro because they actually moved it over, was data access campus wide. We built some extremely valuable relationships through this application. This application allowed to get access and to assist with working with financial data across campus. We are now sending the residence hall addresses to the rest of university for students that are on campus.

28:20

And it kind of actually helped us build an infrastructure and a backbone for all the information that we had access to. Because now as we’re building these larger projects, we now have more data than we ever could have imagined having. When we had one website and we only had those 10 pages to worry about, I don't know if we would ever imagine we had access to so much data. This data became extremely valuable later on.

So next example we’re going to talk about – actually, does anybody have questions? I’m really going to try to fly it through but if you have any questions, feel free to just raise your hand. Next is we’re going to talk about is the recreational sports management system. So we have rec sports on campus. I'm sure a lot of campus and universities have rec sports. Remember management was extremely complicated, that new facilities that were going up, they wanted a solution that was going to assist them with this process. They had people, they're hiring people. They had enough money to pay for just about anything.

29:14

But I don’t think we were – we didn't know, we weren't going to get involved and be able to get an application up in time for the opening of these facilities. They're on an extremely difficult time schedule and this had to be done. So our client, recreational sports was one of – it's again, one of the largest schools for any vendor. The Ohio State University happens to fall into that almost all the time. We're a huge, huge, a huge client.

So the cost for us to build like I said was too great, so we ended up going with the purchased product. Industry leader, I won’t say their names. If you have any questions about that you can ask me, we can talk about that. But this was the most or the best fit. And I say best fit because it didn't have everything we needed.

30:00

Resources, it was great for resources because we couldn't build that. We're going have to hire an entirely new staff just to build that. They're going have to have Rexfort's IT staff just to put this up in any kind of fashion at that time.

An increased standardization, it help with member management a lot which is extremely important in dealing with a bunch of people and all those people into a system. And an improved operations rows in offices. We saw this, I think, it was kind of also tied to the facilities but there was just a bunch of people that work for Rexfort before and now with the system, kind of, delineated things and put people in certain positions. These people had, you know, more defined roles which is extremely important for their organization and helped us all a lot.

And access data, again, as I said before, having access to data, it proved to be extremely important in this situation because we relate to this relationship. They're trying to purchase this. We had some involvement during the upgrade that was more important but our involvement, I wish it would have been better.

 

31:00

That's all to say that. The situation that the application that was purchased was actually going to be sunset. You know, it was on its last leg and it need to be replaced. That was kind of an issue for us because that meant that we had to upgrade to a new platform.

Going us to the new platform require migration. Well, we're also going through an SIS conversation. So we we're taking, I'll just call sensitive data in replacing it with less sensitive data. And in our university that was an extremely high pressure situation. So because member management was a moving target and we thought we have an easy time importing users into the system, we had to really build a solution to help us put in place 65,000 students every single quarter.

Or on a daily basis, if somebody dropped the class, they didn't pay, they got out of the university, they still have an ID card. It still would have worked off at the desk. We have to find a way to make sure that it didn't happen.

 

32:06

So, we had to do a lot because member management was not done like it was supposed to.

Another thing is missed deadlines. We were sold that things would work as certain way that, I think, it was the age group in council. It was council in campus. So you have children, you want to see your children to the summer camp, this new product will have that have built in. All the version on the old product actually work.

There was a time when we were holding both in place and the old ones still have a lot of sensitive information on it. The new one didn't but those two still need to work together. So, we actually ended up building like a shopping cart application so that we could put together, have access to the new information and still be able to work wise through the old environment.

So, what actually, you know, into the patternings of our resources kind of went up. We had to maintain some of the systems for a little bit longer.

 

33:00

We had to work with this group and we really, kind of, sat next to them for a while just to understand exactly what's happening. Yes.

[Pause]

Well, we have pitt bull for project manager. At that time for this application, there was a pit bull for project manager. It think it has saved us some money but really it added to the headaches. It has done now but I think they missed the deadline by almost a year.

I mean, it seem like it took nine months for us just to maintain the bold systems. And it seem like every quarter because we're still on quarters, we have to migrate users from the new system into the old system. And then we had this cross-walk situation because they had sensitive material in some places. Then we just said, we're not going to hold to sensitive materials anymore.

Some of these agreements, you know, it broke down a lot of situations in our own office. Our infrastructure team did not want that data on our service anymore.

 

34:00

Didn't want anywhere near us. Then we had to kind of, you know, maintain that really hoax into allowing us into to do that. That answer your question? Because they didn't work any faster. I don't feel like they work any faster. It got done but it still took a very long time.

[Pause]

Yeah. The big problem with this Con is that you're hostage to them after you have signed that contract. So, you're a hostage to whatever they told you what they're going to do. And as soon as you start to build things to work with that system, you're still a hostage because now you've got a client that used to. Not only this awesome system but this awesome system you've been on top of the system.

And once that awesome system is gone, they don't do something right. They expect both of those. So the next product has to be even better. In some cases, you put yourself in a situation where going end up building or have been required to upgrade things as they upgrade.

 

35:00

So, yeah. You really have to get very close to them. Sally doesn't always work the best.

So, next is conduct management system. I am going to really fly to this because I want to make sure there's some time for questions. We did this internally.

Originally, we started with the housing management system for student conduct. The desire on campus was to centralize this all across the university. And the university with 65,000 students, there are four or five conduct offices. There's university police. So we tried to put all these people in the same room, look at products that took a very long time.

But we ended up purchasing from a smaller company. That was extremely versatile. They seem to be up with the technology. They're remotely hosted, was less expensive annually than what have been for us to recreate the situation for all these other departments on campus and include the changes that the housing office wanted. And it kind of created a situation where housing was kind of driving.

 

36:00

Since it was housing's application first, housing had the keys and they kind of could have gone in the direction. So we ended up purchasing this. It would also meant is that we had to send data and transmit data to outside party.

Resources were free, that was very valuable. There's better set of features than what we are able to create. It would have taken us a long time to get that up and running. The product we actually ended up doing background checks.

I think there was a visit of their offices. We met with their team. They flew out. We need to be talked to them, they filled out paperwork. They had to agree to certain things we have. On our campus now, minimum computing security standards, that's very useful.

If you guys don't have that. Hoping everybody does. If you're going into buying versus building decisions, that becomes very valuable. And training was included. Training was very valuable because we have, I think, 300 RAs on campus.

 

37:00

There's a large number of RAs on campus. I can't sit in a room and train all of them at the same time. Because with house, they was such a large client. They're going to do a lot. They've bent over backwards to make sure we were happy.

Some of the Cons and I think one of those come up, first the size of the issue. The limits of their system were kind of theoretical. We're sending 65,000 students to them now. We want to send more students. Still we want to send students from last quarter and students for the next quarter riding a few departments. And they need that information.

We do not really know what their limits are but we will have to find out together and that could be very risky. Internal client issues, I think, everybody here knows what that means. Some client thinks that they're more important. They think that their business practices are better. They don't like the other department because they're too similar and they don't want to be merged. The vendor actually ended up being over the less flexible than we would want. For me, it matters but it only really matters to when it actually impacts how much work I have to do extra.

 

38:02

That came to at a point where we have to build basically another application to find information for these students before the information got employed. We've since retired that and they're using the system the way it was sold. SLA is extremely important. You need an agreement or you need an MOU. You need to know what the client needs to do if there's a problem because in a situation we have third party vendor, the client may contact you first.

Our only responsibility is making sure of the data feed what got sensitive. We send bulk data. If there's somebody missing in that bulk data, then there was a problem for us. Authentication is a big problem. We were told that they would have Shibboleth thought indication on our campus within the first three months. It still hasn't happened. We have had the product for two and half years now. That's a big problem. Alright.

So overview. We'll go over lessons learned. I am going to fly through these as quickly as possible. Alright.

 

39:00

Understand your client. Understand your organization is probably more important. If you are doing this for yourself, understand exactly what your organization is capable of and then understand what your client wants and needs. Be involve in scopes and requirements. Expect operational delays and complications. That is always included.

Make sure there's a pause for data that's coming and data going out. If somebody decides to purchase something or if you decide to build something, make sure there's a way to make sure that information is available to you or other departments after it has completed. If some of these is decommissioned, you still want to get that data. And then make sure that the strategic initiatives are observed.

This is important if you're trying to sell something to the upper management. You need them to believe in the system. You need them to believe in you building it or they need to believe you buying it. You need to use what they want to get exactly what you want. And your clients should understand that already because they deal with this in their own organizations.

 

39:58

Size matters. For the housing university, the size has always been an issue for us. There's 65,000 students, not every solution is going to work. Get involve as early as possible. I have said that before. You don't want to just have something dropped off from your lap. Communicate with the vendors if you have the ability to do so. Talk to them. Work with them. As an IT group, they should understand you have questions and concerns. Prepare for support. This means, even if its not your problem, you have to be prepared.

We have an authentication system on campus, Shibboleth. We use that for most of our authentication on our applications. They call our office all the time to tell us that Shibboleth is not working. That means nothing to me because I don't control Shibboleth but you have to be ready for that. That's going to mean something for your bottom line.

Maintain security. Keep that idea up and then again explore hybrids and this means you may purchase something but be ready and prepared to have a plan in place to design and develop something to work with it. I'll go through this fast as well but I'll go backwards.

 

41:00

Be prepared. That's important, as I said before. Service orientation. If you can make it what we've done on campus is kind of built. Actually what we've done in student's life is we're starting to build this architecture that allows us to get all these information from all these various sources and deliver it whatever ways we see fit. So that means web services. That means XML feeds. That means bulk data feeds to these vendors.

We are trying to be as prepared as possible so that we can have that available. I, as a developer, want to build everything but I know that it's not always the option. But if we need to build something to work with something else, we want that sub authority in place. And make sure you have data infrastructure in place. We have done that. It improves availability.

Make you're not deleting data before you pull new data and get to the source as close to the source of the data as possible. And this, I believe, management sees us. As a developer, I think, we're seen more as jokers on campus and in many ways, I think, it's the truth because our goal is to make our jobs look better and to make us look cool and to do fun things but in some cases, they just going to get things fixed or up and running.

 

42:12

Does anybody have any questions? That end was really fast. I'll try to answer your questions as quickly as possible.

[Applause]

Thank you very much.

[Cross-talk]

Batman? Batman? Alright. Batman it is then. Thank you guys.