What is our primary use case?
I use two versions currently. The latest version that's just come out and the 2022 versions.
Currently, I'm using the solution to pull APC data using the data collection. I've got to transform the data massively to make it usable, so I use FME for huge amounts of transformation and to put it into ArcGIS Enterprise. FME is a ZTL as well, specifically the "L" bit. I also use FME for a lot of analysis where we want it quick and efficient on big datasets that ArcGIS Pro is unrealistic to do on.
I also use FME, where if somebody sends me a spreadsheet where, for example, we would need to get everything in capital letters, it's quicker for me to capitalize everything in FME Workbench than it is in Excel.
I can use FME for just about everything, from logging processes to looking at file structures, what's changed in file structures, what people are doing, or what they've updated. It's not just need-to-know. It's a massive piece of software that's also coded in Python. What FME can't do, I then write in Python.
I can make FME pull in ArcPy, which means that for every single thing that ArcGIS Pro can do, I can make FME. It's much bigger than the box.
What is most valuable?
When I do coding, I think about every single function. Some of these functions can be very elementary, like doing a substring or some capitalization. But FME removes all that coding because it's a transformer, so the time to develop an application to get to a point where you're producing results is decreased massively. It used to take weeks and months to develop software, and now I can use something like FME, and within one day, we get results. We can look at and validate data. We make minor subtle changes to the workbenches to improve it. We can share the workbenches. We don't have to use GitHub or anything else.
What needs improvement?
The one thing that always appears in the community is the ability to make really easy loops to loop through data efficiently. That needs to be added at some point. There must be a technical or commercial reason why it has not been done in FME yet, but many people ask for it, and it just doesn't appear. We can do it, but we have to create a child workbench to do it.
For how long have I used the solution?
I've used FME for more than five years.
What do I think about the stability of the solution?
I've used the 2022 and 2023 versions because of the application's stability, not the tools. What FME has done is they've bought out 2023 and update the backend code for the interface, and it's crap. The update was more about making the presentation pretty, similar to what they did with Office or ArcMap to ArcGIS Pro. I use 2022 when I want the interface to remain stable over a long period and 2023 when I want to see the updates and see if anything could have been changed.
The stability with the 2022 version is really good, and I rate it an eight out of ten. The stability with the 2023 version is probably a five out of ten. The stability depends on the version used.
What do I think about the scalability of the solution?
FME's scalability is fantastic. I rate it a ten out of ten. It's one of the most useful pieces of software I've ever encountered in my entire life.
More than ten users use it in my organization.
How was the initial setup?
FME workbench is dead easy to set up. It cannot be easier. Double-click on something and go, "Next, next, next." And then it's done.
Licensing is so straightforward if you've got a trial license or pay for it. You just type it in, and it authenticates on the server.
FME server, on the other hand, comes in two types: FME Cloud, with FME server inside, or FME server, which is on-premises. They are straightforward, provided you've got an IT department to set up the servers correctly in the first place.
FME Workbench took three minutes to deploy. FME server took a morning. FME Cloud is all automated. We just need to sign up, type in our details, and add a credit card, and It builds the server for us. The whole process takes under one hour.
What was our ROI?
We have seen an ROI, but it doesn't justify the price.
What's my experience with pricing, setup cost, and licensing?
The pricing is really bad. Last year, they rebranded the whole pricing structure. It used to be moderately priced at about £400 per user per year. Now they've changed the whole thing, and it's expensive, at ten times the price. The pricing is based on your company size, turnover, and all these different things to do with the company.
I had a conversation in London. I spoke to an FME guy from Canada at a conference and said it was horrendous. FME Server used to cost £10,000; now it can cost over £100,000. It's worth the money, but the rebrand is just killing them.
What other advice do I have?
I rate FME a seven out of ten because of the price.
The number of types of data you can save and read is growing massively. Similarly, the solution is improving at working with APIs, downloading data from remote sources much more efficiently. There are more transformers, but these are aimed at data types that don't apply to me.
Before choosing FME, immediately look into licensing and costs before you go down the road of trials and becoming dependent on the trial. When you finish that trial, you'll find yourself semi-dependent on software. Get the costs first because it's a shock.