I mentioned in my previous post that, at a previous career stop, I built open-source support into our IT lifecycle. Specifically, we used QGIS. The primary reason we made that choice is that we were a Mac shop. It’s true we could have run ArcGIS Pro inside Parallels, but I didn’t see the need to pay the “VM tax” in terms of performance to use software that was essentially a match in capability to QGIS, which ran natively on MacOS.
One of the nice things about QGIS and its open-source licensing is that we could simply add it to our Jamf profile for our geospatial analysts and install it remotely as part of the default setup via the MDM. It integrated seamlessly into our IT management without any worry about licensing. We had roughly five to seven analysts using it at any one time, with some working remotely (this was pre-COVID), so the ability to manage it via MDM was essential.
The other nice thing about QGIS is that it mostly ships with everything. It’s roughly equivalent to ArcGIS Pro with all of the optionally licensed extensions available, and it is easy to install additional QGIS extensions. But this post isn’t about QGIS versus ArcGIS Pro. There’s nothing wrong with Pro, but it simply didn’t work for our use case.
We were deriving significant value from QGIS and I saw the need to support it more directly. We didn’t really have the developer bandwidth to support it with coding, so financial support made the most sense. I could have walked in to my leadership, who were actually well-educated about the open-source community, with a number and I would have immediately been sent back out the door with direction to not return until I could justify the ask. That’s a fair position and how any well-run private sector organization operates. I was essentially proposing something unbudgeted. No matter how beneficial it may have been, it needed some rigor behind it.
I kept the math simple. I analyzed the amount of revenue we had generated from the products made in part with QGIS. Next, I determined the license tier and extensions required to accomplish the same tasks with ArcGIS Pro. I then looked up the ArcGIS pricing on GSA Advantage, which lists pricing for the Federal Government. GSA prices tend to have a discount applied when compared to normal commercial pricing, but the commercial prices of ArcGIS were not available and this was close enough for what I was doing.
When I was done, I simply took a third of the total pricing as our target donation. Next, I looked at the QGIS membership program and chose the tier that came closest to our target number, which I recall was Medium. (The prices have changed since then.) With all of that information, I pitched our CEO, who approved the plan, and we signed up for the membership. From that point forward, a medium membership was budgeted in every year, just like any other software purchase, until the company was sold.
So we supported the open-source software we used at roughly one-third of the cost of the commercial equivalent and banked the earned value. In subsequent years, we used that earned value to support efforts like Lutra Consulting’s QGIS installer for MacOS and the GDAL CRS barn raising. Lutra’s installer was much more MDM friendly than the one we had used previously and the we used GDAL extensively throughout our business. (Quick perusal of these efforts will show than none were an astronomical amount of money. A little goes a long way with open-source.)
Our approach to supporting open-source focused on value generation. Where we derived significant value from open-source, we found ways to give some of that value back to the projects. As a budget line item, it looked no different from a standard software purchase, but, in practice, there was no limit to our ability to install the software, either in terms of where or how, and derive additional value.
In the end, I find concerns about how to financially support open-source to be somewhat overblown. If you can determine how much value you earn from using it, you can figure how how much of that value to give back. If you can’t determine how much value you earn, then your issues are with yourself, not with open-source.
Yes, it’s easier if a vendor just gives you a price. Yes, it’s harder to calculate value on your own. And yes, doing so requires a pretty detailed understanding of your business and its value chains. But doing so makes you better, and it makes the open-source software you use better and more stable. It is an exercise that is well worth the effort.
Header image credit: Diomidis Spinellis, CC BY-SA 4.0 https://creativecommons.org/licenses/by-sa/4.0, via Wikimedia Commons
