What is our primary use case?
I'm an application developer with a fair amount of database background so I mostly use the tool to do physical modeling to support our application development. I'm a firm believer in not just adding columns to a table but to actually think about it, put together an Erwin model, and look at the relationships. I used to like to generate the model and generate changes all through the tool but being honest, one of my biggest frustrations with Erwin is that it's very difficult to forward engineer and keep things in sync. It used to be so easy and now it's very difficult. It's very frustrating to use this tool for that.
We use it for data modeling but they also do a lot of logical modeling and architecture, and we also use the naming standards capability to force corporate standards across the models.
How has it helped my organization?
It has improved my organization because using a data modeling tool is forcing us to come up with better models.
Its code generation ensures accurate engineering of data sources. It should generate correct code, so I can't say it cuts development time just because it's doing what it's supposed to be doing correctly.
What is most valuable?
I think the ability to depict the model in a graphical fashion, think about it, and keep things consistent is what's valuable about it. It's too easy when you're using other methods to not have naming consistent standards and column consistent definitions, et cetera.
This isn't specific to Erwin, it's specific to any data modeling tool but we also like:
- The ability to graphically depict how the relationships occur and the relationship lines.
- The fact that it migrates your foreign keys for you.
- The general principles of what a data modeling tool does.
Erwin does a lot of things well. It's just very frustrating in some areas that really should not be frustrating.
The people who don't use a data modeling tool but rather use spreadsheets or wing it typically have pretty poor data models. If you use a data modeling tool, the graphical nature of the data modeling tool forces you to think about relationships. It forces you to ask questions that you wouldn't ask if you were just creating tables and doing it off the top of your head. That's number one, in my opinion, from my own experience. The number one benefit of using a tool like Erwin, is that visual representation forces you to come up with a better model.
Its ability to generate database code from a model for a wide array of data sources is useful but we're 99% SQL Server, so the fact that it generates 60 other databases doesn't really help me too much. It doesn't support Postgres or Redshift which are the two other systems that we're using.
What needs improvement?
I find the UI very clunky and very difficult to use. The whole workflow of adding columns to a table could be so much easier. I get frustrated using it. Resizing dialog boxes, changing fonts, printing, scrolling around in the UI, etc is very clunky.
I've tried other tools. I've tried to get off of Erwin a few times. I always come back to it because every other tool has its own set of problems, and it seems like if I have to pick my poison, I stay with Erwin. But so many things that are clunky with it.
My biggest frustrations with the product is forward engineering and keeping things in sync. A lot of times I need to change a column definition and all I want to do forward engineer it over to the database, it used to be so easy to do that, way back in the early days with Erwin before CA bought it, and now it's almost impossible. It's very frustrating to do. I've spoken to Erwin about this in the past, and I can understand why they're doing some of the things they're doing, but I'm more of a casual user than a power user, and for me, it's so clunky. It's so much easier using Embarcadero to forward engineer changes to a database than it is using Erwin.
This product has been on the market for years and I'm amazed at some of the quirky things that I still have to deal with in this product. I wish rather than adding new features, Erwin would fix some of these usability issues.
For how long have I used the solution?
I have been using erwin Data Modeler for around ten years before it was owned by CA.
What do I think about the stability of the solution?
Other than the bugs, it doesn't crash on me, so I guess the stability is good.
What do I think about the scalability of the solution?
We have somewhere around 20 uses. I use it as a developer and the data architects use it as well.
We use the Mart model, we break them out into areas, and there are many models in each area. So we have around a couple hundred models.
How are customer service and technical support?
I haven't used their support in quite a while, so I'll say neutral.
Which solution did I use previously and why did I switch?
We previously used Toad and Embarcadero.
I've been using Erwin since it first came out, a long time ago. Back then it was a lot simpler to use and it was just so much easier. I think they tried to make it do everything for everybody and now it's very difficult to do some of the simplest tasks. It's very frustrating, and there are a lot of issues.
The forward engineering frustration I experience with Erwin is a thousand times easier in Embarcadero. If I want to just make a quick change to a column and forward engineer it to my database, it's a lot easier in other tools.
Some of the other tools were a lot better in the ease of use and stability of the UI but they also had their share of problems that are deal-breakers. For example, models won't print on one page. I keep coming back to Erwin. It was the lesser of the two evils. No product is perfect but I think Erwin tries to be everything to everybody, and sometimes when you do that, it's no good to anybody.
I don't use all the features, it's nice that they're there, but I wish the stuff that I did use was better usability-tested.
How was the initial setup?
I was not involved in the installation of this particular version. When we first started using erwin, we used to install it on our local machines, but now we're using the Mart model and it's installed on servers, so we have a group that maintains it. For years and years, it used to be that we all just installed it on our local machines and ran it that way.
It's a licensing thing. We have a concurrent license so by having it on a server, it's in one place, which is nice. That way, everyone's running the same version. Then, because we have concurrent licensing, if you have 30 people that need to use it, but people like me only use it once in a while, you don't have to buy me an expensive, dedicated license, so it's a lot cheaper to have a concurrent license for our company.
What was our ROI?
It's not necessarily erwin-specific, but by using a data modeling tool, it forces a better product, better application development, and better applications at our company. Using a tool like that is a must-have.
What's my experience with pricing, setup cost, and licensing?
I like the concurrent licensing. That's phenomenal. I think that was a big win for us.
What other advice do I have?
Sometimes you have an initial idea for a data model and when you try to design it in Erwin you realize that you were wrong in how you approached it. Erwin enforces consistency and accuracy. Quite often I learn something by looking at the generated code. It's not like I create table statements all day long. I don't do that generally. So when I use the tool, it generates the correct code in scripts for me which we will then hand off to the DBAs who run them.
I would rate it a six out of ten. It's frustrating. It could be so much better.
The problem is mostly usability. It has little quirks about the way the screen refreshes, things move around, and the workflow when you're creating columns and tables could be so much better.
I have a love-hate relationship. I've used this product for years. I've actually gone to training on it at Erwin, so I know what I'm doing with it. I wish they would make it easier to use. I would think if Microsoft bought it, this would be a totally different product.
Interestingly enough, Microsoft has tried to come out with data modeling tools a few times, and they are all bad. They're basically toys. You can't use them for anything real, which is surprising to me. You would have thought that they would have had a tool that could compete.
There are only a couple of big players out there that Erwin competes with. I looked at just about all of them, and I keep coming back to Erwin, but I hate it nonetheless. There's nothing better. There are certain tools that are better in certain areas but far worse in others, and so you pick your poison.