I spend most of my time talking to OEM program leaders who have done everything right on paper. Connectivity running, data flowing, devices on machines across the installed base. And the conversation always arrives at the same place: why isn’t this doing more for the business?
The honest answer is simple. Connectivity isn’t just something you do for your customers. It’s the data foundation for your own business.
Most programs were built the other way around: designed to satisfy rental customers, not drive OEM commercial outcomes. That decision set the ceiling most programs are now hitting.
Every machine an OEM ships is already generating data from the first hour in the field. Loads, fault patterns, duty cycles. For many OEMs, that data never makes it back to the people who could use it.
Aftermarket teams work from historical averages. Engineering waits for warranty claims. Dealer networks react to customer calls instead of machine condition.
None of this is inevitable. The data already exists. BCG’s 2025 research makes the commercial case clearly: aftermarket service margins run significantly higher than new equipment sales. The OEMs capturing that margin have a program built for their business, not just their customers.
Getting there starts with a decision most programs haven’t made. Connect every machine by default, not just when a customer asks. Fragmented coverage produces fragmented insight. An OEM with 60% of its installed base connected is working with 60% of the picture.
Full-fleet connectivity changes that. Complete coverage reveals which components fail early and where machines run beyond design assumptions. It shows where warranty exposure is building before claims arrive. None of that surfaces reliably from a partially connected fleet.
Connectivity creates coverage. Coverage creates pattern visibility. Operationalizing those patterns is what turns machine data into revenue, cost control, and uptime.
The question worth asking is simple: what is this data actually for? Most OEM programs were built to answer that question on behalf of the customer. The shift is in answering it for the business.
When field data flows back to engineering, real problems surface. One OEM with 10,000 machines found them consistently over-specified for actual duty cycles. Closing that gap generated a validated €2.5M efficiency gain across the fleet.
Real-world usage data makes that possible. It shows how machines actually behave, not how engineers modeled them. OEMs who build this feedback loop catch design issues before they become warranty campaigns.
When it flows to service and aftermarket, dealer networks stop reacting and start anticipating.
They know which machines need attention before a customer calls and dispatch with the right parts. Demand signals from connected machines replace the historical averages aftermarket teams have always had to work from.
Some OEMs decide to build the intelligence layer themselves. I call it the do-it-yourself stack trap. They stand up cloud environments, build BI dashboards, start layering in AI tools. And it works, for a while.
The challenge isn’t the first version. It’s maintaining it: security certifications, compliance, API versioning, data normalization across machine generations. In my experience, teams often hit the wall around 18 months in. Stretched thin keeping what they built running.
The OEMs moving fastest focus on what they are best at: the machine, the application, the customer. An operating data platform platform handles the infrastructure underneath. The speed difference is real. One OEM deployed over 10,000 customer portals in two weeks after standardizing on a platform approach.
Once the foundation is in place, the next challenge is access: getting data to the people who need it. Model Context Protocol (MCP) is the open standard making that practical at scale. It connects familiar AI tools directly to live machine data. No exports, no custom integrations, no report cycle.
A service team can ask which assets are approaching a maintenance threshold. The answer comes in seconds.
Most OEMs already have enough machine data. The real question is whether that data exists primarily for customers or also, for the business itself.
The OEMs finding the answer are redesigning their programs around commercial goals: aftermarket revenue, engineering margins, machine sales. Each of those goals already has a data answer. The machines have been telling them for years.
Paul Wilson leads OEM sales across the Americas at Trackunit. He has spent a decade working with construction equipment manufacturers to build connectivity programs that go beyond basic telematics. Before Trackunit, Paul ran the IIoT division at ZTR and held senior sales leadership roles at TELUS Business and Info-Tech Research Group.
