Skip to main content

How Rhode Island DMV partnered with Experian to accelerate its data migration

Data modernizations are complex undertakings. As new regulations and customer demands require updated systems, public sector agencies across the country are working to modernize. 

Perhaps the most crucial, and most challenging, piece of modernizing is migrating data from legacy systems. Following a traditional data migration approach, many unforeseen obstacles can lead agencies to overrun original timelines and budgets. 

When the Rhode Island Division of Motor Vehicles (RI DMV) realized that they may run into some obstacles due to a lack of validation for their migration, they sought out a solution to help. Collaborating with Experian, the RI DMV was able to validate the accuracy of their data and the viability of their business rules prior to the migration, ensuring success.

Tune into this webinar to learn:

  • Why modernizations matter
  • Common obstacles organizations face
  • How to overcome challenges and accelerate the migration

Hello everyone, and thank you for joining us for today’s webinar, “Driven to succeed: How Rhode Island DMV partnered with Experian to accelerate its data migration.” Today’s webinar is presented by Experian, in conjunction with the Rhode Island Division of Motor Vehicles. My name is Jenna McAuliffe, and I am a Marketing Specialist here at Experian. I will be your moderator for today’s session.

At this time, I would like to recognize and give special thanks to our promotional partner: the American Association of Motor Vehicle Administrators. AAMVA is a nonprofit organization that serves North American motor vehicle and law enforcement agencies to accomplish their missions. AAMV helps develop model programs in modern vehicle motor vehicle administration, law enforcement, and Highway Safety Association also serves as an information clearing house in these areas and acts as the international spokesman for these interests. Thank you for helping to promote today's webinar.

To go through today's agenda. First, we'll review the importance of data modernizations. Then we'll provide some background on the Rhode Island Division of Motor Vehicles. Third, we’ll review How Experian helped. Next, how RI DMV is moving beyond the modernization, then reviewing the results, and finally going through the question and answers as a wrap up.

Next, we’ll review our featured speakers for today. We are joined by Walter "Bud" Craddock from the Rhode Island DMV. We're also joined by Chris Colen from the Rhode Island DMV, along with Experian’s own Basil Brown and Andrew Marr.

Walter, or Bud Craddock is the Administrator of the Rhode Island division of motor vehicles. He was appointed by Governor Gina Raimondo in 2015 and oversees the agency's day-to-day operations. He's also a member of the region one board of directors as the secure such as the secretary and the treasurer. His duties include regulation, enforcement of laws relating to the issuance, suspension and revocation of motor vehicle registration and drivers licenses, financial responsibility related to motor vehicle ownership and operation, vehicle safety and emission inspections, and other applicable state laws. Previously, he served as chairman of the Rhode Island Motor vehicle dealers license and hearing board and as chief of police for the city of Cranston. He holds associate's and bachelor's degrees from Bryant University, a master's degree from salvage Virginia University, and a juris doctorate degree from Roger Williams University. He is a member of the Rhode Island Bar, the Florida Bar, the Rhode Island Federal Bar, and the US Supreme Court Bar.

Chris Colen is an advisory board and principal lead for program and project management wouldn't silver tree services he served as the program executive for the Rhode Island Modernization Systems program. Chris has a track record of delivering major business transformation programs for more than 30 years, has leveraged his business knowledge and expertise and information technology to manage turnaround and the timely delivery of business critical projects and programs across a range of industries. Integrates the right mix of people, processes, and technology to solve complex challenges and drive change. With this background, he provided sound insight and leadership for Rhode Island's modernization. Chris earned an MBA from Southern Methodist University, has certifications including Amazon Web Services, started certified business professional and corporate council, Charter member and the Program Management Institute.

Basil Brown is a presales engineer and the technical lead of experience data management solutions. He brings more than 15 years of experience to his role at Experian were for more than two years. He has been helping customers to identify and implement software tools to fit their specific needs. He has built and designed a wide range of solution including data aggregation, data patterning and data quality components to solve challenges related to revenue assurance, fraud management, and business intelligence. Basil earned his bachelor's degree in computer science from the New York Institute of Technology.

And finally, we're joined by Andrew Marr, a senior account executive managing public sector customers and industry partners. He provides consultative support to state and local government agencies that are looking to leverage their data to solve business problems and informed decision making, working in public sector. Andrew understands that his customers are often dealing with limited resources and budget constraints. Despite these limitations, he enjoys working strategically to provide reliable solutions that are both cost effective and have a positive impact for United States citizens. He earned his bachelor's degree in business management from Babson College. Now we turn to a very timely issue about data modernizations. So we're here today to talk about data modernization. We know that in the past few years it's been a top of mind subject for many public sector agencies. So at this time it's my honor to turn it over to Andrew Marr.

Thank you Jenna. Thank you again to an AAMVA and the great folks in Rhode Island, DMV for partnering with us. This has truly been a partnership that we're proud of. And to kickoff our Why Modernization Matters slide. It's really important to know that we're in a data-driven era and where information is instantly available with smartphones. Quite literally, this instant information is at our fingertips, whether consciously or subconsciously, we looked at data to make decisions. Public sector agencies are no different. They've done a phenomenal job in providing services to citizens with legacy systems and applications. Some systems dating back over 20 or 30 years old with new regulation and increasing expectations to deliver excellent service to their citizens. Public sector agencies are often faced with pressure to update and modernize these old and outdated systems. Legacy systems are ineffective for supporting the services that agencies look to provide for their constituents and processing exponentially larger volumes of data. Modernizing systems often serves as an important opportunity to offer more efficient services to their customers and a view of all DMV functions related to customers. We recently conducted a benchmark report on the public sector this year, and one of the things that we asked about was modernization plans. In the agencies that are surveyed, we found that public sector agencies will modernize a 41 percent of agencies have plans to modernize systems. In 2018 alone, the proportion of DMV planning to modernize is even higher at 60 percent planned for projects in 2018.

Whether your agency is currently planning a modernization or old tackle, one in the future, there are some common challenges to be aware of: multiple systems, outdated data, data quality and standardization issues, lack of resources, budget constraints on average in the public sector, agencies are dealing with 15 individual contact databases. Experience approach to effective data management highlighted in the diagram can help agencies be more effective with these outdated systems. This creates complications with modernizations as often one of the goals is to consolidate disparate systems in any database that is more than a few years old. You expect a certain amount of data to be outdated when you consider that some of these systems are decades old, bad data can quickly become an issue. In addition to outdated information. Data quality issues are common. Nearly one in three agencies, 32 percent in back, say data quality issues as a main challenge to their modernization, even more than data quality issues, a lack of skills and expertise and lack of budget pose, significant challenges to modernization efforts now that we've spoken to why modernization is so important in some of the common challenges. I'd like to ask Bud Craddock and Chris Colen and just share some information about the Rhode Island DMV and their goals for their modernization project.


Hi Andrew. Thank you. Good afternoon everyone. I'm again Bud Craddock here that we're an island DMV and just a little bit of background about us. The One island DMV is a centrally located agency with our headquarters in the city in Cranston, but we have six other locations serving approximately one point 500 million and a half and five 50,000 people here in the state of Rhode Island. Uh, we haven't bought a hundred 82 full time employees. And when I was appointed I was faced with taking over a project that was a struggling after being in development for approximately five years. One of the issues that was a concern is making sure that we moved forward not only with the development of the product, the process to get the system up and running, but we were faced with a getting a very large amount of data from 39 different databases into a one customer record system in the DMV. We're a full service agency. The last slide showed the number of trends that are the types of transactions that we handled. One of the things that was surprising is when we were putting the project together, we were doing approximately 240 to 250 separate types of transactions for the citizens of Rhode Island, so it was quite time consuming to bring this information over and I think I'd like to pass this to Chris to talk about how we actually were faced with bringing the data from the legacy into our new RIMS system that we put online last July.

Thanks Bud. And good afternoon everybody. This is Chris Colen, so, but sort of outline the. One of the major challenges that we had, which was we're moving to a single database, single customer design, and we were migrating 20 to 30 years of legacy data, um, and the legacy system had, I'm almost none if any validation. So over the course of 20 to 30 years, you can imagine the number of, of different, uh, permutations of people's information, um, and, and the multiple data stores that Bud is already talked to. So one of the key premises of, um, of DMV migrations is really the key outcomes. What are the key outcomes is really to get to that single customer single record. Um, and we chose to do this in a single database as well. So we moved from, as Bud mentioned, 40 plus different data stores to a single customer, a single database.

So the other things that we were being confronted with is, as you know, as real id compliance becomes even more and more critical. Um, you know, and the other items were that we really wanted to enhance our online and our, uh, our customer service, um, capabilities to reduce the inflow of people into the DMV. So, um, you know, our, our data quality and validation, um, was a very large, um, large hurdle for us to come over and, um, and that's why we, you know, we, we started looking at, um, you know, is there a better way we had, you know, we have two point 6 million legacy credential records and we had 11 point 7 million legacy vehicle records, license plates, titles, etc. So, you know, we, we sorta had to investigate is there a better way to get this done? And I'm already sort of talked to sort of the complexity of the databases, the data stores and you know, living through a very large period of have no validation. So, you know, we, we started looking at how we might specifically do this, what's some of the challenges that we had, um, and um, and then, you know, at that point we, um, but actually, um, had a, had established some connection with, with experience and um, and at that point we, um, we move forward.

Thank you very much Chris and bud. And as Chris just alluded to, this is how it all began. Rhode island, DMV. Uh, we met bud at the AAMVA conference, as Jenna mentioned in the introduction and the is such a strategic partner not only to motor vehicle age and emcees, but also to industry partners. Uh, it is a great place to come together with the leaders from both industry and also from motor vehicle and law enforcement agencies and understand what are the challenges that they're facing. It's a great opportunity to learn and connect in 2016 at the end annual and conference. We met bud from the Rhode Island DMV and we presented a demonstration of our solution at the conference and seemed like a good fit to help Rhode Island DMV overcome some of the challenges they were facing after a conversation about the validation and data quality challenges that Rhode Island was experiencing with their monitor migration efforts. It seemed like we might be able to help initial discussions lead to further investigation on the Rhode Island part about how data management software would add value to the project and seeing the value that we could add with this tool.


Thank you, Andrew. I would echo what you just said that the AAMVA conference I think was a very important when I first met with you folks down there, I knew I projects for the monetization was going along appropriately with the development of the court program, but I knew a gating issue for a successful launch was getting that data out of the legacy into the new rim system. And with that as a backdrop when I met with Experian, I saw their product and I thought this might be something that would be beneficial to bring back to the states. And when I did get back to the materials that I took from the conference as well as, uh, what I learned the presentation I shared with Chris and the rest of the it leadership team and said, could this be beneficial? And at that poInt they looked at it and evaluated it and thought that there was definitely some worth to going a little bit further with the experience. And I would think that we were successful with our launch and not to any small, um, input from experiences that I think they were essential to us being successful.

I apologize,

I was gonna, I was gonna add, it was maybe some more from a historical perspective, but actually mentioned that the program had been developed for development for about five years and had, had numerous starts and stops and a couple things happened. I, um, you know, the team, our system integrator who was HP now DXC, I'm introduced an agile approach to um, to get a lot more alignment, business alignment and, and to actually begin to demonstrate the capability of the solution. Um, I joined, I joined the team in um, in august of 2016, as you may know, we went live in 2017. So when I, when I joined the program, the data, we already talked about the complexity, the number of databases, the volumes of data that we needed to migrate. And then the, um, you know, the going to the single customer single database design.

And then I started looking at the traditional approach to how we were. We were going to migrate the data and meeting with the team. And, you know, as you know, the typical way of sort of dividing the business rules, developing the ETL, testing the actual technical migration, and then I'm iterating time and time again through, through that migration, the team that was on the ground at that point, estimated that it would take them between two and three years to, to actually do that. Um, and, and it was sort of one of those things where, you know, obviously you're trying to figure out how to accelerate that. And bud came in and said, hey, I met these folks, um, and maybe we should take a look at this and it's sort of all came together. So, um, you know, we, we actually got together and talked about our challenges, talked about the complexities and um, and we went from there and that's where we met, um, sort of basil as part of the presales team and embezzled, just sort of rolled up the sleeves and jumped in. So let me, I'll turn it over to you can sort of describe the of the engagement, um, with, uh, with the Rhode Island DMV.

Thank you, Chris. So looking to protect the most judicious process for the migration, you need to ensure the accuracy of the work that was being done by the site now who had experience and particularly using the Pandora until we have what we believe is a unique way of approaching migrations. And I'll cover that on the next slide to give you some more detail, but fundamentally what we do is we ensure that quality and a picture of quality is taken at the beginning of the process and as then continued to through the process intuitively as the uh, quality of the data is actually lifted. So any point in the process, you've got a clear idea of what's good, what's bad, you're able to prioritize what needs to be fixed and what programs may need to be put into place or indeed what data needs to be dropped.

So parent experience, paired doors, capabilities with the SQL code in a covered up by the site actually helped to complement each other and to push the actual conversion over. Now in terms of the process, and we're going to first of all take a look at the data. What happens with a, um, a typical migration, as you can see there, the shell represents the different task groups, um, areas of work and what will normally happen is that migrations will happen. It will be based on samples of data rather than full volumes. What that actually means is when people are writing sequel code, they'll take a look at the stuff that they know about to such rather than thinking outside of the outside of that to look at stuff that they probably weren't aware of it and they become really quite challenged in terms of what the breadth of the data can actually hold as a norm for that.

You'll get to user acceptance testing. Some of the data will drop out, the migration will be rejected at that point and it's back to the drawing board to basically start again and go through and you ended up getting many iterations like that which not only pushes out the cost, certainly pushes out the intended goal, live date. Now, from an experience perspective, this is how we actually had basically the same chevrons that you've got there or again, the big differences though. The second chevron, a proactive American specifications and expedience platform we do for volume data analysis and um, and a migration strategies and mapping strategies within the platform so it acts as a sandbox within the migration, your integrations that you can see happen up at the top there which start from UAT and go backwards or basically all done within the task, the purple second group, but you can do as many iterations as needed to be done there until cover eventualities that are actually seeing the data. And then maybe you move on food you've effectively done in UAT there. So by the time you get to the user acceptance testing, you should be in the clear, when you've covered everything that you're going to encounter and you've really lifted the, um, the work of the rather the probability that your process is going to go ahead on time. You're de risk the project. Um, and certainly from what we've seen, pretty much every case, the migration projects coming on time and on budget. So with that, I'll pass back over to the bottom.

One of the things that was very important to me as the administrator or the DMV was the responsiveness that I saw from experience when we started to work with them. Um, they were very, very helpful. They were positive. They tried to make suggestions and streamline the process as much as possible. And I can't say enough about how when we hit stone walls or issues that were blocking progress, they were very responsive and jumped right in and roll their sleeves up, cut side by side. We'll either in person or with my staff here to make sure those obstacles were overcome as quickly and as easily as possible. And again, we were faced with a timeframe of starting a migration. Our actual go live in July of 2017. And as I said earlier, I think our success was in no small part a contributor into by the Experian process and the help that we got from the team that they had working with us, basil being they'll lead point of contact on it. Chris. So, and

I guess I would build on what Bud’s, Bud’s, comments. And really what I would highlight is to do this and you know, this, this is really hard work, um, and it requires the engagement and the participation by everybody. So, you know, we had a really outstanding team and it was a team effort. Um, you know, we had our, our systems integrator who HPC DXE experience played a critical role, obviously heavy engagement with the business because we're dealing with the data of the business. And we were um, you know, our success is based on the success of our ability to have this data ready to go on day one. And then of course the state of Rhode Island, Rhode Island's, um, department of information technology. So, you know, what I would say is that the real key was we had a team, we, it wasn't always a, you know, everything didn't go exactly to plan, nothing ever does.

But the true test of a, of a partner is when things really start to, um, you know, when things start to go, exactly to plan out. Everybody rallies around and comes together. and I, and I've got to say that a basil and his team have worked, um, many times around the clock and supporting us. We did, um, we did this in six months versus the two to three years. We did continuous iterations. We did mock cut overs, we did, um, to, to really, um, simulations of our go live. Um, and we were migrating data. I'm 100 percent throughout throughout that, through those processes. And actually when we finally did go live, we achieved a 100 percent data migration, data validation on that, uh, on that file, go live a little bit of luck. You probably would think a lot of hard work and um, you know, uh, it just really was what I would really stress it was the team, a tremendous amount of planning, a tremendous amount of dedication.

And just to follow up, this is Andrew on the outcomes of the Rhode Island DMV, Experian and system integrator. As Chris mentioned, this truly was a team effort and working with the Rhode Island DMV and their staff as well as the information technology group. It was, it was truly a pleasure. And it was a was a team effort where we were able to add a lot of value is through that reimagined approach of a migration process. We were able to speed up the process, give them the high degree of confidence in the record counts that were transferred to the new system, and as, as Chris mentioned, it doesn't all add all also helps to have a little bit of luck along the way, but what we're able to do finally it was validate the rules that system that the system integrator had developed in an automated way that saves both time and money and so that really talks about the modernization, the process that we had for the project specifically, and now we look towards beyond the modernization and experience, able to provide continued value through this data management tool.

In looking at the phases of data quality, since we have so much data, it can be very difficult trying to figure out how to use it properly, especially when you're trying to figure out what's wrong, what you're missing, where the gaps. This challenge is oftentimes that we don't understand the data well enough to know what's causing issues. We don't know what to look for. We might not even know how to start looking, and as a result, we continue to operate in a reactive process. We'll deal with that. Data agencies will accept and accuracies because that's the way that it's been. When we find a problem, we fix it, but don't necessarily understand why it's happening. We know why it might happen again and it will probably repeat itself and the process again, but how do we actually move forward and make progress? Experience approach to effective data management highlighted in the diagram can help agencies become more proactive.

As Basil mentioned in his technical discussion, full volume profiling and analysis is extremely helpful and understanding just what you're missing, what's wrong with the data, and where the gaps, the ability to identify and improve data inaccuracy is within the tool, accelerates the general improvement process for standardizing data and enriching data, and if you automate your data quality processes so that are ongoing in the background, any issues or quickly flagged and you're able to have a more proactive approach into managing your data. For the Rhode Island DMV, there were three major steps that corresponded to these phases, migrating the data from a legacy system up against a deadline, profiling and resolving errors in the data and creating an automated process to provide control. Moving forward, and this really is, as we saw on the previous slide, a uh never ending process. We use this maturity curve that you see today because it helps us first characterize the state of which the department or business that we're working with currently operates. We discussed how the migration has helped the Rhode Island DMV move from a reactive state where the majority of organizations tend to be to a more proactive state and experience data management tools help to increase that proactivity with this ongoing relationship, we'll hope to continue to move and be in an optimized space moving forward. And with that I'd like to turn it over to basil to speak to a little bit of the technical things that we do after modernization has occurred and looking forward and ahead.

So typically with the tools that are Pandora experiment, uh, be able to give ongoing improvements to data quality and this is achieved via profiling and the creation of rules for ongoing monitoring or we like to term as controls in order for a, um, an any instance really the health and status of the data to actually be understood. So monitoring within Pandora to works in a couple of ways. Every time data is loaded or unloaded or refreshed. If, however you may want to call it, we automatically provide a full profile of that data. Now for most profiling tools that made me in about 17 different attributes, but I had to actually produces 227 different data attributes within the, uh, within the profile of any data that's loaded and there's zero configuration needed. Now that’s true with the box capabilities as such. now rules are different, but we provided easy interface for simple to highly complex business and data quality rules to actually be creative development tool and once again, once they're created, every time that data is actually refreshed, the tools themselves would be rescored.

And what that gives you a bit as well is the ability to measure trends for particular measures that you might put in place across time. If you've got a. If you've got programs which are actually costed money in order to improve the quality of data at different areas of the business. This can be actively monitored so that you can understand the return on investment that you actually get it or even understand what that quote that may need to make be able to or may need to be actually curtailed. As such, the final advantages that the tool sets and then two pronged really from the perspective of business users, they can actually be much more hands on with their day to actually access the data itself, wherever it may live with us and file form or in databases without needing to put request into it to actually get into their data themselves.

The business users quite typically are the ones who actually know the nuances in the data and actually understand the operational functionality that's contained within the data itself. What this needs to from an it perspective is that making the mailers, the guardians of the data or the systems certainly that the data resides in or may have been created, have all my past and float through and takes away the, uh, the other tasks which are actually normally piled on top of them simply because they didn't have the access to the data. As such, I'm. Pandora provides an interface which allows really technical sequel code to be created for a completely graphical and simple to use intuitive interface. And with that then I think I'm going to pass on to our last person, which is bad.

Thanks again, basil points that are worth noting that when we were going through our monetization, we recognized early on that the data migration was going to be one of the biggest problems and we had heard that from other DMVs that had modernize their system as well and that this could be a very long and arduous process. Uh, to give an example, when we looked at our legacy database records, we had over $22 million separate customer records in that system. That's where a state that I had mentioned earlier, I had only a little over a million people. Uh, we have since we're through the migration effort, that experience in Pandora helped and assisted on. We've taken those 22 million records and condense them down 2 million, 700,000 records now. So it's been a big, big benefit to cleaning up the system and making sure we have accurate records for all our residents and to go forward with the new system.

Uh, we had mentioned earlier about how am connects to the jurisdictions and the agencies with technology vendors or other people that can assist us in our mission to serve the public, but one of the other things that amazon does is they have working groups and as we were going live with our monetization, they had a, a group putting together a best practices document. We were privy to an early issue of that even though it wasn't ready for release to all DMVs across the country. And one of the points they made out is that data cleansing efforts will take several years to complete. It should never really end even after the new system is in place because it cannot be started too early in the process. Nobody agency ever be completely finished with data cleansing. The agency should plan for ongoing ordering and cleansing of a system data to ensure databases and records remain as accurate as possible. And as we go forward with our modern system now we're looking to continue that partnership with Experian and Pandora to help us keep this information that we brought over from the old system clean and usable for the foreseeable future.

Chris, anything you’d like to add on that? No, I think you've done a great job. And I think again, everybody realizes that, um, you know, one of the biggest mistakes that people make when they undertake a program of this magnitude to go through all of the effort to, um, to cleanse and migrate their data. But then, unless they have programs that continue and evolve, um, you know, to support the business, you'll just begin to revert back. Uh, I don't think we'll ever get to the situation if we, if we stopped everything today, we never get to the situation we, we, uh, in our legacy world because the new systems all have multiple validations, um, as part of them. But, um, you know, I think it's done a good job sort of talking about the results, um, And the, and the support that we've gotten not only from [inaudible] but from, from Pandora. And we actually leveraged the experience network and they put us in contact to other experienced customers, other, um, DMVs across, across the country. One specifically that comes to mind was Idaho. Um, so, um, you know, we also were able to, uh, to gain knowledge and insights from other customers and, and we, you know, the DMV and the state of Rhode Island is, um, you know, wants to repay that favor to other DMVs as we move forward.

Thank you, Chris. Truly appreciate the partnership and that's something that we as an experience to be a catalyst alongside AAMVA to grow those relationships and connect between motor vehicle agencies that are going through similar projects. Uh, as bud mentioned, that emphasis, demonetization working group has truly done a phenomenal job in putting together best practices that states can make a repeatable process off of these modernization processes. And, uh, for data migration efforts moving forward, we'll be able to help enrich that community and help share best practices alongside both of those organizations. A would that I believe that concludes the formal presentation and I'll turn it over to Jenna and a to move into a question. And answer.

Perfect. Thank you, Andrew. I'm so everyone on the line, if you have any questions, please do type them in and we'll try to get across as many as we possibly can. um, I've seen a couple come through already so we'll get started with those. and like I said, please do keep a entering your questions as we go through the answers here. Um, so the first question that we got is we are preparing to undergo a modernization. What are your top tips for getting started and how can we make sure to include all of our data and all of our systems?

Is that too? Do you want me to answer first? Sure. Basil. I think you're probably a good person. So I can't emphasize enough the initial upfront, um, things which are known book based about the setup of the project itself. When you actually starting to move into stuff with the data, you've very first thing has got to be a full analysis of all your data and by that what we actually, he made his profile every single table which holds data rich, which you're going to move forward with profile certainly all of the attributes of the pairs of attributes that make for keys within the data that you know you will be moving forward. Understand just what the uniqueness, the completeness of that data actually is, and then start to step further back away from the data and defined the rules which will help you to understand data that you will have to take over as opposed to data that you will not be able to take over.

Ensure that those business rules can be actually enforced on through the data itself and the state that actually is there and all of the records that you're taking over. Please, please, please try and uplift the quality of any data attribute that you have in that. By that I mean, um, things like addresses, things that they spell it on, the formats of social security numbers. The validity of date of births that you have in the system, all of that stuff. Once you've assessed and understood what we'll be going over won't be. you should then have programs in place to actually uplift the quality of that data. If you want to have a successful migration with minimum, if not, no pain at all.

I think it's probably important to ask Bud and Chris if they could add their perspective from a jurisdictional perspective on to answer that question as well.

Chris, you want to go first?

Sure. I'll jump out there. I think did a very nice job covering a lot of, a lot of the points. I think, you know, um, you know, sort of from my perspective, the mapping, the data mapping, um, was what is critical, um, you know, finding we're finding where everything is and all the difference in our case where all the data resided. Um, and then we actually developed, um, around 53 data objects, um, that we were going to migrate from our, from our legacy to our, our modernized solution. So, um, you know, basil talks about the mapping, he said talked about defining the rules, um, you know, the, the analytics are extremely important and then, you know, and maybe this is a little bit down the road, but the, the trial and error and the continuous improvement, you can't, um, you know, you sort of can practice your migrations and you can't do that enough and um, you know, and I think that that would, that's what I would add. I don't know. But if you have anything you want to say,

Sure. I will talk about the technology aspect of it from my perspective, from an operational standpoint is identifying individuals who have good business knowledge as you have points of lead on the project. so you have continuity. So you have somebody that knows the business operations, that can guide the development of the software itself that you're going to be using for the modernization and make sure you have a strong governance behind it so that, that person, when decisions have to be made, can come to the governance council, uh, to get the proper decisions of where to go if they're struggling with where we should be at the end of the day. Went on any particular point in time but brought out two. Excellent, excellent to the.

And I would, I would, um, I would also add is, um, when I would I engage with a, with a program, when I looked for business participation, I go to the business and say, who are the people that are going to cause you the most pain in your day to day operations if you take them out. And those are the people we want to be the thought leaders around the program. And it's very hard to do, but if you want to be successful, I, um, I think that that is truly key. One of the other things I throw in as well as we found that we engaged all staff on a regular basis, bringing them in and you're doing showcases as we could develop the solution so they could see and we had staff input and making suggestions that were very beneficial to the overall development. and also it helped that one we did go live. They all felt ownership in the project. So there was less reluctance to be switching over. And that was one of the, I mean, one of the true benefits from going to the agile process. We were able to actually show the system as it, it continued to evolve and, and, and actually put feedback into that process.

Um, I, I think just to reiterate the, um, the, the, the points that were made by Chris and Bud, it is absolutely imperative that the ownership actually lies with you. The organization that's, um, that's the subject of the one for all those points that they just, that they just highlighted. And you'll do remember that you complete whatever tools actually into the program and process. You can be whatever color consultants in from the outside who beat the target system that you move in so that the data is always yours and the nuances and rules around that data has been developed over time within your own organizations and you truly are the experts with that and it's your knowledge that actually makes these programs succeed.

Great. thank you all for your great answers. To that question. We'll move on to another question that we received as a business user at my agency. We have to make a lot of requests to it in order to gather information and to write SQL scripts. You mentioned that Pandora was able to help ease this process for Rhode Island's modernization, but is it able to help in other ways?

Yeah, so I, I think certainly those two points that's, um, that, that you just made, there are probably lots of bella being able to let business users actually get in and get the heads, does he actually get the data themselves? But I think the primary thing when we talk about what I talk about that particular fact is, um, you know, business users who understand that though the day I can literally scrolled down through tables of data and notice little nuances and little that are out that you know, to the rest of us being absolutely nothing. So once they get It rolling back background to that experience that leads within the, um, within the organizations making the data accessible food to pet door, first of all, allowing them to get their hands on it, highlight stuff that is wrong. And then actually from a Pandora perspective, you're actually able to prototype out fixes and changes to the data. You're able to use reference data for verification. These are all really, really simply achieved. And I'm, like I said, you can do this in an intuitive manner, which means that as you prioritize and progress through the changes that you have to apply to the data in order to make it fit for purpose, you really can't chart map and understand exactly how that progress is going.

Oh, you guys. Yup. Sorry, basil. We were gonna. We were gonna add

the, um, so to your comments, to build on your comments, was there really two points? First of all, um, how did the business users were able to come up to speed extraordinarily quickly? Um, when we did, when basil came and he did his first morning and training our business users were doing queries within the first two to three hours and getting and getting resolved. The other point that bud mentioned, and I think we interrupted you, but we were in a bud, was pointing out that, you know, in, in the past we had to have the it organization or people who were familiar with writing sequel code do all these queries and with, with Pandora we were, um, we were able to do that with, you know, business rules, business based queries. Um, and that, that's what really accelerated our timeline because as, as I mentioned, the original estimates, um, you know, we're two to three years just because of the amount of SQL code that we had to write to vote. We were doing that to validate the ETL. Um, we're actually running a parallel process which was, that was the original design construct, um, and um, and then we abandoned that. So, um, and that, that was really the two points where you don't need people to write sequel code, um, and that the business users can, can drive this with, um, with the support of, um, of the Experian team.

Great. Thank you. Excellent. Excellent points made there. Um, we've got another question that seems like it would be fitting for you bought and Chris to answer. uh, so did Rhode Island have [inaudible] files and slash or mainframe data that needed to be migrated?

We had mainframe, uh, data that was migrated. Most of it was, most of it resided on the mainframe. Some of it, there were flat files there. They were all different kinds of data stores. We just had sort of amalgamation of, of different data stores and embezzle you may have better visibility into that because you got into quite a bit more of the detail.

Yeah, I know. Certainly the data that came out, I think, yeah, some legacy Cobalt systems. I'm not so sure Whether they were have output, but they suddenly had cobalt output which was coming out through that stuff and that certainly not uncommon for the bar in the world as such, but yeah, there were certainly some of that in the background that was coming through that to be dealt with.

Great. Another question that we got is about specifically for using technology, so basil, I would guess this will probably go to you first and then maybe we'll kick it over to Chris. If you are migrating a system which is a canned product or a vendor product that doesn't have any foreign keys built in to the database or internal documentation for the product and you don't have expertise on that product, how are you able to profile and identify the specific data issues?

Okay. Let's deal with that. I think that in two different levels, first of all, if you're talking about not having documentation up and not have been foreign keys, which are actually defined what you need to understand that, uh, two things. First of all, the structure that you go into because that they'll certainly have the phone and keys and nicus between these different relational data sets defined within it that gives you a clue as to exactly what it is you're either looking for or that you're going to have to build or identify. Now, our tool, Pandora will certainly allow you to take values for what you suspect may be any kind of a key field as such, and compare that to what else is in the other data sets that you've brought in. That's one way. Certainly will be there to pare down. We'll confirm that you can actually make these links between the different bits of data.

You can then go ahead and actually try the full blown joints themselves as you shape the days, molded to the, uh, to the outgoing structure. And I think once again, just to reiterate something I've said a couple of times now during the course of this, you know, your ability to actually profile and understand your complete this, a film that you may have, that any of these fields are combinations of fields is something that tool set actually does, um, straight out to the box and those are the kind of attributes and functions that you really need to be able to get into the data. Not necessarily you're not gonna write any sql code at all, but certainly understanding that this particular column of this table for this old legacy system had, is actually going to be my key driver identifier when I take this across it into my new system on the other side.

Chris, do you have anything to add there?

Nope, I think basil covered it.

Okay, great. Um, we have another question up. Please do get them in if you have any more. Otherwise, this may be our final question. So just please do enter them if you have any left. Um, we have more than 40 legacy data systems with important information. how is it possible to consolidate our records and Improve the data for our modernization? Basil, we’ll send that to you again. Actually, maybe we can ask for Chris and Bud to provide their input at a high level and then Basil if wants to elaborate technically. That'd be great.

Well, you know, I think I really would have to ask for his, his knowledge around that because um, you know, once we identified the, um, you know, the data, the data stores, it was really the, the creation of the, um, the targeted, uh, you know, sort of, uh, data, uh, attributes that we're looking for. Um, and then the mapping. I mean, you know, it's all, it's a lot of hard work, um, and you know, as you identify your targeted data objects, um, and, and um, you know, how your new system is going to operate, um, you know, it's, it's what basil described earlier about the full analysis and the profiling. Um, and then, you know, looking, looking for the attributes that align, um, and then you know, that that mapping, which is, um, there's two phases to that mapping and I think bud and basil, all of us have pointed that out.

The first stage is the business side of it, you know, just how, you know from the business rules, what are the business rules that we're going to use, um, you know, uh, to accomplish the mapping and then it's the, you know, that it's a very complex technical, um, job to, um, to develop, uh, the, the etl is going to take all that data from those all the disparate data stores, databases and, and bring it across. So, you know, you probably have a more eloquent technical way to describe that, but it's just a lot of hard work as, as we've said. Um, and it's um, and it's a lot of trial and error. Um, and, um, it gets it. It's very frustrating. I don't know if we, if we sort of painted, I hope we didn't pay too rosy a picture because we definitely had some very tremendous challenges trying to take the volumes, have bud highlighted over the period of time that, that all that data was created and bring it into, um, bring it into a single database and it's, it's all about trial and error and, um, and not, you know, not giving up.

And then the other thing I would say is you also have to set your, um, your targets for what, what is the sort of the minimum viable qualities that you're willing to accept that the business can run with because you can get into a situation where, you know, to get to that 100 percent solution, you know, it's the old 80 slash 20 kind of a rule. So, um, yeah, that, that's how I would answer that question from a, from a program and a, a business perspective.

one thing I'd add is that start early, one of the issues that faced us is we were so focused on the core functionality of the monetization system that we're putting in. We kind of overlooked the data migration for a long time and then when we realized that that's when we started the sweat, that's when we were a little worried about a project and that's what caused me to be intrigued by Experian and Pandora when I went to the conference that I saw this as a possible aid to getting us to that finish line. So definitely start early.

Great. Thank you. I'll just rephrase the question one more time just so that basil can give them more technical answer and we haven't forgotten what was specifically being asked. So we have more than 40 legacy database systems with important data. How can we consolidate the record and improve that data for the modernization is the question.

So I'm actually not going to give it 10:00 so, but I'm actually going to just call out one thing that we haven't mentioned. The tool, as we all know, that we have an understanding on a very, very high level, you know, earlier on, Chris mentioned about going in and I tend to find the people in the business that if they lost them, it would be the most painful, um, I think, um, and uh, Tim, I'm going to, I'm going call out Laurie Mcdaniels over at Rhode Island. The work but she did in terms of applying her knowledge to the data sets that you guys actually have, the clarity that she had. She was able to build it into, build it out with business rules. Actually that was the side of applying the business knowledge that actually helped to smooth out or corral the number of data sets that were actually needed to be pulled over. So, um, you know, that that's a non-technical officer, but that certainly calling out something that was very, very important to the success of this project.

Thank you. And thank you Bud and Chris for joining us and for the great answers that you were able to provide to the questions that our, our attendees have asked today. Thank you to basil and thank you to Andrew as well for the great knowledge you've provided. Um, I will just a couple of items here. The recording will be made available for the Webex as soon as possible, so do look for that in a follow-up email and as well as the slides which you can anticipate getting sometime next week. Um, so with that, we'll conclude today's webinar. Thank you all for attending. Again, a thank you to AAMVA for being our promotional partner in this, and a very special thank you to the Rhode Island Division of Motor Vehicles for their tremendous partnership not only in the project but also in making this webinar success. So thank you all for attending today.