The Unexpected Beginning of a Transformation
When I took over the GIS Department for St. Joseph County, MI, in 2016, I assumed I understood the terrain. After nearly twenty years running a private GIS consulting company, I figured a rural county of 60,000 residents would have a modest, workmanlike operation. Something like a sturdy sedan that needed new tires, not a spacecraft held together by folklore and hope.
Instead, I found a department performing tasks simply because someone, at some unknowable point in history, had once performed them that way. No one could remember why. There was a faint sense that asking might disturb whatever spirits kept the printers running.
My staff had no formal GIS training. Their job descriptions owed more to union negotiations than geospatial logic. Workflows were less “workflow” and more “ritual.” And the data, well, the data had developed a personality of its own. Parcel lines drifted across the landscape with the sort of whimsical freedom usually reserved for migrating birds.
This wasn’t low-stakes work either. We supported Assessing, Equalization, the Drain Commissioner, and E911 Addressing among others. When your spatial foundation wobbles, tax rolls wobble with it. Emergency responders wobble as well, which is not a comforting thought when you are the one responsible for the wobbling.
To make matters more interesting, the entire system was locked inside the expensive and bloated ESRI vendor ecosystem. There was no transparency, little flexibility, and almost no alignment with what rural government GIS departments or residents actually needed.
After a short acclimation period, I realized the GIS itself wasn’t the root issue. The structure around it was. The department was doing GIS the way some families do holiday recipes. No one remembers where the ingredients came from, yet no one dares adjust them.
Stripping Things Down
With no real budget and an obligation to fix things anyway, I set three rules for myself.
- Simplify.
- Streamline.
- Automate, but only after the first two had been thoroughly honored.
So I questioned everything. Every map, every process, every strangely persistent task that seemed to exist mainly because it had survived the previous decade. I asked whether each action served residents or simply served tradition.
It turned out a great deal of it served tradition.
As the unnecessary work evaporated, something revealing happened. The workload shrank to its true size. We went from three full-time staff to two, and eventually to just me. Service improved throughout all of it. The department didn’t become efficient through cuts. It became efficient by ceasing to behave like a historical reenactment society for outdated workflows.
Other department heads looked at me with a mix of curiosity and mild suspicion. Local government rewards empire-building. More staff, more budget, more square footage. Voluntarily shrinking a department can feel like speaking heresy in the church vestibule.
Yet the practical work improved. The Drain Commissioner finally received drain assessment district boundaries that didn’t appear to be guessing. E911 received address data they could trust. And I discovered something quietly important: GIS had never been the real issue. The larger system of how information moved inside local government was creating the friction.
Fixing What Actually Mattered
Once the noise was stripped away, the true problem revealed itself as plainly as a crooked section line.
The parcel data was the weakest part of the entire operation.
So I did something most county GIS Directors never contemplate. I remapped the county parcels myself.
There wasn’t money for a vendor or staff, and the existing map appeared to have been constructed during a period when cartographic accuracy was considered optional. I set a simple rule. Every morning, over coffee, I would remap twenty parcels. No excuses. No skipping.
It didn’t feel heroic, which is probably what made it effective. Consistency is far less glamorous than promises of sweeping transformation, yet it actually works. Over several years, those quiet morning sessions repaired thousands of errors. I rebuilt PLSS corners and section lines so the county finally rested on geodetic truth instead of cartographic folklore.
The project is roughly 85% complete today (24,000+ parcels remapped). The rest is waiting on high-precision GPS section corners. Yet the impact arrived much earlier. As the data improved, long-standing problems dissolved. The daily calls about parcel lines slicing through living rooms became a thing of the past. Boundary disputes eased. The constant stream of nuisance calls slowed to a trickle.
The chaos had never been an unavoidable reality of GIS. It had been a side effect of bad data.
Clean data created a base for real improvement. Workflows could finally be trusted. Automation became safe. Public-facing tools felt solid beneath the fingertips.
Local government loves dashboards and apps. Yet none of it holds up if the data wavers.
Automation, Self-Service, and Quiet Reform
Once the data was stable, automation could finally carry its share of the load. I scripted updates, checks, exports, and routine tasks. We rebuilt the county website so residents could answer their own questions without phoning a human who was trying to finish a donut before the next crisis arrived.
Power users emerged. Realtors, assessors, surveyors, and title agents learned to navigate the system themselves. Dependency on our department staff softened. Confidence grew.
The End of GIS Data Sales
Then I did something dangerous. I reviewed our data sales program.
We were earning less from selling data than it cost to process the sales. To add a touch of irony, no one had ever purchased the full county GIS dataset. The pricing had been set so high that I suspect it was meant to dissuade anyone from discovering how unreliable the data had been.
So we dismantled the sales model entirely.
We replaced it with a simple, free download form. No hoops, no invoices, no timid tap-dancing around outdated fee schedules. Public data isn’t the harvest; it’s the soil. You don’t generate value by selling it off by the bucketful to a handful of buyers; you generate value by letting the entire community build and plant on top of it.
The Paradox of Success
By late 2022, the GIS department had become something unusual in local government. It was efficient, automated, and quietly reliable. A good portion of the work had become self-service.
This revealed an awkward truth. I could no longer justify a full-time GIS director’s workload.
The county still needed GIS, yet clean data plus streamlined workflows plus self-service tools equals a job that cannot reasonably fill forty hours a week. This led to a slightly philosophical moment about the profession. Many GIS departments spend a surprising amount of time producing outputs that resemble productivity but do not meaningfully change outcomes. Attractive maps that mostly exist to reassure conference attendees. Dashboards that sparkle yet illuminate nothing. Fluff.
Once all of that clears away, the real scope of GIS becomes visible. And it is far smaller than most structures built around it.
So I faced a decision. Invent tasks simply to remain busy, or apply the same principles that had transformed our GIS department to the rest of the county.
The universe made the decision for me soon enough, because, just then…..(BAM!)…..ChatGPT arrived. And as every AI thumbnail on Youtube says…”That Changed Everything”.
Author’s Note: The Road Ahead I’ll be following this up with the story of what we built on that clean foundation: the Digital Information Department. We didn’t stop there, though. The final piece of this series will cover the Digital Innovation Collaborative Exchange (DICE), a regional digital public utility we are currently building to bring this model to the rest of Southwest Michigan.