If you were talking to someone whose organization is considering erwin Evolve, what would you say?
How would you rate it and why? Any other tips or advice?
Think about the overall processes and governance of your company. If you're trying to do it as a business management system, just getting to consensus on the content may be the most challenging aspect. You have to plan for enough time to get build the content before you can actually put it into Casewise. Also, think through the object model and the relationships. Try to get that done correctly upfront. It's better to put enough design and thought upfront and it will pay dividends down the line. We would like to use the solution's collaborative web modeling capability, but we took a more document centric approach to start with. We would like to build process maps to be able to visually describe our processes in a future phase.
Erwin EA Agile is very easy to use. It has great web-sharing of architecture to many different stakeholders, so that you can socialize your architecture. One of the things I was most interested in, in the 2018 version we just deployed, is the pan and zoom capability. I haven't tried it yet because we just got it running yesterday. We have not used a lot of the solution's integration capabilities with other tools in our system, yet. What we have done is a lot of Excel import and export, and that works fantastically. I've done some experimentation with the Visio import and export, and that is helpful. What we haven't done yet is direct connections with other modeling tools or our other systems of record, but it's something we are looking into. With the Excel importing, the "up to date" part is the challenge. If we had a real-time integration, we could keep things up to date for whatever kinds of change points we had. With Excel, it is more that you have to export from one system then import it to another, so it's better for data that doesn't change that often. In terms of the comprehensiveness of the data when importing from Excel, it's fine. You can get a lot of data from Excel and bring it in. As an example, well over a year ago we were going through a major merger. People were collecting information in Excel about all the interfaces between all the systems, theirs and ours, and how we were going to reconcile them. I was able to take that spreadsheet, that had 250 integrations in it, and rapidly, in less than a half an hour, bring them into the model and start showing diagrams that visualize that data. That was huge.
Start small. Even if you are thinking big, start small. Use the first object and, after that, make it evolve. The biggest lesson I have learned from using this product is that because the product is really easy to configure and it's easy to create new things with it, sometimes you are expecting more from your contributors than they can do. The tool allows you to publish what you have. It's cool for that. We are using it to document what we are transforming. We have decided to document our process applications and interfaces, which is a longer project. We are documenting the project as nodes or objects and we are documenting what we are creating. This works very well but the result is that we have thousands of objects to manipulate. So if we want to do an analysis, a statement, it's easy. But if we want to do a big analysis, EA Agile helps but it doesn't do it entirely for you. It's something that can help someone who has the mindset for this type of work but it won't do all the work for you. The solution's collaborative web modeling capability for live, remote collaboration on a model works, but people in our organization are not ready to use it a lot. In the beginning, we thought that a lot of people would use it. But after a few months of use of the application, we figured out that you have to be an architect to take advantage of it. Someone who is not making maps or architectural things every day will not contribute. It's not due to the tool, it's due to the mindset. In terms of the solution's integration capabilities with other tools in our system, there are pro and cons. The positive is that it's really easy to integrate because there is an API, but it's quite new. So it works, but not that much. There are some tools inside the solution to manage import and export of data and that works. But sometimes, it can be really massive and generate a lot of errors. You can integrate it with almost everything, but you have to watch what you are doing because when too much is integrated it's really hard to manage your errors. So you can integrate, but at some level you will need another tool. It depends on what you are doing, but for example, it's easy to import a list. You can integrate it with Active Directory. We were the first customers to who have that functionality and it works really well. But to integrate with one of our providers took some time because we needed to align our vocabulary. Now we are integrating with our documentation library but we always do the same thing. We first try to do it manually and then automate it. Trying to automate it first is really not easy. We looked at integrating the solution with the erwin Data Modeler and Intelligence suite but, at this moment we have not moved forward with that project. Right now, we're making a data catalog first and that is something we can do with the solution we have already bought. We have some 500 attributes and eight or ten project objects. That has been a big job for our business. We're taking things step-by-step. We may integrate those other modules in the future. The ability of the solution's web platform to tell you which insight has been accessed the most helps us focus on what matters to our colleagues in business change activities, but not natively. We use another web analytics solution on it, so we definitely know who is looking at what. That helps us a lot because, having created a lot of objects, we want to know if an object is being used or not. And with the help of the editor, we track changes made by users. We have been able to define what we want to track. But it's not a native feature. To get the functionality you have to so some configuration. The solution has back-office and front-office rights management. In the beginning, you have to be strategic about it to keep it simple. Otherwise, you can end up configuring something that is really big but not usable. After that, you define what you want to be automated inside. Because the tool is really customizable, you have to define some strict rules and learn to use them. That's pretty difficult. What is difficult, really, is to define what you want because the tool is really open; maybe too open. We have about 30 people using it, overall, but it's not always 30 people at the same time. In the beginning, it started like that but now, effectively, every day we have less than 10 users who use it every day. We about 800 users, unique IDs, in total. We have something like 50 users on the site every day. Among them are project managers, service managers, and a lot of technical people.
Start with the basics and get a fundamental understanding of how the suite is configured. There is a modeling suite, an object analysis suite — the object window — and the analytical tools. I love the web deployment, the Evolve capability. That is really nicely done. The ability to collaborate across geographically distributed installations seems to be working well for us. I wish it had more API-type interface exchanges with other tools that made it much easier to transfer things to and from something like System Architect, and not to have to go through a lot of man-hours to recreate what we needed to recreate. But otherwise, it has done the job for us and continues to do so. We have five users, with three of them being power users. And for the maintenance and administration of the tool, it's me and another gentleman in West Virginia who mostly deal with things like that. We don't have any plans to extend usage beyond those people. If anything goes wacky on the server, we just get erwin people involved. I give it a seven out of 10 because there are that I'd love to see it do more easily.
We didn't have the perfect design on the paper before starting to implement it. We just started to implement it, then tried to use it. It is better to try and hope than wait and miss out. Adoption is not connected to any tool with enterprise architecture. The capacity of people to feel the benefit of enterprise activities will be from going through education and a commitment on behalf of key people. I see a willingness in the company to add new features. Though, I don't feel they are pushing the solution hard enough. I would rate the solution as an eight (out of 10).
Start small. Start with the projects that you can get value out of it. Don't try to boil the ocean or to map the whole world. Say to yourself, "I'm creating a map," like Google does while they're driving around in their street-view cars. Figure out what your initial goal is and use the tool to gain that goal. If you have success there, go on to another one. It's an approach you could apply almost anywhere. The biggest thing I've learned from using the solution is that it's really hard to document the data usage of your company. That's one area that we're pretty weak in. At least in our company, it's really hard to get a handle on all the data elements that people use — both customers and employees: How that data is used, where it is stored, what the liability is for it, and how long we have to keep it, etc. erwin has a data governance tool but, when I looked at that, it seemed we'd need a handful of people just to do that alone. Early on, I thought that I could get a handle on the data much more quickly than I have. I would rate erwin at close to nine out of 10. It's not a 10 because it still has little janky things that are wrong with it. It's not intuitive and it takes a while to get value out of it. A 10 out of 10 is when you get value right away. Using erwin, if you work at it, you get a lot of value. For other people who are going into it, it might be that their bosses or their budgets aren't going to tolerate a month-long or even year-long ramp-up to value.
We are currently considering the public cloud deployment model using AWS. It's a good product. I would rate it eight out of ten.
My advice to others would be to have this solution on your shortlist if you are looking for a data modeling tool. I will rate it an eight out of ten. I would like to some additional features, like data lineage, included in the next release to make it a ten out of ten.