Light Fleet Management

Years after I left MISOM, MST Global bought the company. Part of what made it worth buying was a fleet-management product called FARA, which was, to my knowledge, the first mobile-based fleet-management tool in mining. The first screens of FARA were wireframes I drew as an intern who had never built an app. Designed, shipped, sold. My part was the first verb.

MISOM's bread and butter was helping mines integrate their data and surface it as useful information, and a huge share of that data poured out of fleet management, the dispatch brain of a mine, the software that tracks every truck and shovel, assigns the next load, and logs every haul cycle. The systems that did this were built for the largest mines in the world and priced to match. Standing one up on even a 20-truck fleet ran millions of dollars in hardware and software. Below that, the alternative was a spreadsheet your dispatcher typed into. Between those two poles sat a whole category of mines that had never been served at all.

I knew exactly what the spreadsheet end looked like, because I was often the somebody who built it. For mines that couldn't afford a real deployment, part of my job was the poor man's version, Excel tools where a dispatcher could rapidly key in cycle times and truck assignments as they came in over the radio. They worked. They were also a person imitating software, and everyone knew it.

Then cheap phones and tablets started showing up everywhere, mine sites included, and the first wave of IoT hardware around them could suddenly handle things like geofencing, replicating chunks of what the big systems did without the server room or the ruggedized hardware bolted into every cab. MISOM's founder made the call. Build light-vehicle fleet management, right-sized for that messy middle. Our Excel-era tools were useful, but they ran on human fingers. This was the automation step.

The catch was that every developer at MISOM was a backend engineer or an IT expert. They could build nearly anything, and nobody among them could picture what a dispatcher should actually see. The standard move would have been to spec it from the database out and let the screens fall where they may. Instead the founder handed the design to me, the intern. Not because I could code apps. I couldn't. But my whole job had been translating between frontline operations staff and the engineers, and by then I'd sat in dispatch shacks at seven mines across four companies, mapping how their data actually moved. I knew what the person on the other side of the screen did all day.

Designing software I couldn't build meant finding a way to make the design real anyway. I found a wireframing tool that let me mock up each screen and wire the mockups into a rudimentary clickable app. No code behind it, just screens linked together so a person could tap through the workflows. I spent almost all of those screens on the input side. Dashboards were a known quantity; MISOM had been turning mine data into displays for years. The unknowns lived at the point of entry. How does an operator log an activity? How do they fix a mistake? What happens when the network drops? What happens when three people share one device? Does the thing need user accounts at all? Those edge cases were the design. It was my first crack at a native app interface, though I'd been building UIs in Excel for years.

Then the internal team clicked through it, and the debate began. "This is what we thought it was." "This is how the workflows would run." "This screen wouldn't survive a real shift." "Why do it this way and not that way?" It was a bit deflating at the time, but I realize now that this is exactly why you make wireframes. They're cheap, and the first design never survives combat.

Shortly after, I left for an engineering job at Rio Tinto's Kennecott mine. MISOM built from that starting point and shipped it as FARA, and years later FARA rode along in the MST Global acquisition.

I've designed and built far more sophisticated software since, and I keep coming back to this project anyway, because it's the cleanest proof I have of where the value in design lives. Not in whether it's pretty or sexy. What I cared about then is what I care about now, reducing clicks and cognitive burden. A dispatcher juggling a radio doesn't admire an interface, they count the taps between them and the next truck. I still design that way. My crop planning tool earns its keep for one reason, it's so stinking fast to interact with. I still think this is the most overlooked part of "good design."