The maintenance backlog compounds because deferred preventive work turns into reactive reconstruction at roughly 18× the cost. A continuous-data operating model, with a weekly condition refresh from existing fleet vehicles, changes the economics of prioritisation and moves engineering from reactive to proactive. Three sketches of what that looks like for a small, medium and large municipality are below.
Municipal road engineers do not, as a rule, lack expertise. What they lack is current data. Most municipalities operate on a three- to five-year inspection cycle: good enough for budget submissions, useless for monthly prioritisation. By the time a condition survey lands on an engineer's desk, it is, on average, two years stale. The backlog compounds.
This piece is for teams that already know they want to move to continuous data collection but are trying to picture what the operating model actually looks like at their scale. Three sketches below: small, medium and large.
Why the three-year cycle loses
Pavement deterioration is non-linear. A road that scores PCI 65 at inspection can drop to PCI 45 inside 18 months under heavy rainfall, freeze-thaw cycles or a sudden change in traffic (new subdivision, new school run, new industrial operator). The engineering consequence is that the budget ranking your team developed against the 2023 survey is, by 2025, the ranking of a network that no longer exists in the state it was measured.
The financial consequence is sharper. Preventive crack-sealing at PCI 70 costs roughly R50/m². The same segment at PCI 40 requires mill-and-overlay at R300–R450/m². At PCI 20 it is structural reconstruction at R800–R1,200/m². The three-year cycle systematically lets segments slip from the R50 zone into the R450 zone before anyone notices.
What changes with continuous data
Continuous data doesn't mean real-time. It means weekly. A municipality that runs smartphone capture on its refuse fleet, water-services fleet and traffic-enforcement vehicles picks up 1,000 to 10,000 km of fresh condition data every week without a single dedicated survey trip.
The operating change is subtle but important. Monthly engineering meetings move from "what should we do next quarter?" to "here is what changed in the last 30 days and what should we do before the next rain event?" Maintenance planning becomes a rolling function rather than a quarterly exercise.
Sketch 1: a small municipality (under 500 km)
Small municipalities usually have a single engineer responsible for roads, plus a contracted maintenance crew. The constraint is not expertise, it is attention. A continuous-data setup for this scale looks like:
- Smartphone capture on 3 to 5 fleet vehicles, running normal routes.
- Weekly condition refresh covers the entire 500 km network.
- A single dashboard the engineer checks on Monday morning.
- Alerts for any segment that dropped more than 10 PCI points in the previous week.
- Work orders drafted from high-severity detections; the engineer reviews and approves.
Annual cost: typically R70,000–R350,000, depending on which modules are enabled. Compared to a single specialist drive-survey every three years at R700,000–R2.6 million for the same network, the continuous model is cheaper and produces 150× the data.
Sketch 2: a medium municipality (500 to 3,000 km)
At medium scale there is usually a roads department with 3 to 8 engineers, a call-centre pothole-complaints system, and an existing asset-management platform (PMS). The integration conversation becomes important.
- Smartphone capture on 10 to 25 fleet vehicles across multiple operational departments.
- Weekly network-wide refresh plus daily refresh on priority corridors.
- API integration writes condition scores and detections straight into the existing PMS.
- Complaint-to-detection matching: incoming citizen complaints are matched against AI detections within 100 m, so engineers see both views together.
- Monthly budget-allocation report generated from deterioration forecasts.
Annual cost at medium scale is typically R350,000–R2.1 million. The saving over legacy survey approaches is usually 50 to 70%. The harder-to-price benefit is operational: most medium municipalities report that their citizen-complaint backlog drops by 30 to 50% inside six months, because engineers are now finding and queuing defects faster than residents report them.
Sketch 3: a large metro (3,000+ km)
Metros have the hardest political environment: visible failures, competing departments, historical procurement contracts, and usually a legacy pavement-management system that nobody wants to replace. The playbook here is not "rip and replace." It is "feed the existing PMS better data."
- Smartphone capture on 50 to 200+ fleet vehicles across multiple departments and contractors.
- Daily network-wide refresh plus alerting on priority corridors.
- Direct API feeds into the existing PMS, GIS and work-order systems.
- Ward-level condition dashboards for councillors and ratepayer communications.
- Multi-year deterioration modelling against capital-budget submissions.
Large metros typically run R2.1 to 14 million per year in platform costs. That number looks big until you compare it to the maintenance budget the data is informing, often R500 million and up, and the avoided cost of deferred preventive work at that scale, which compounds fast.
What it is not
It is not a replacement for engineering judgement. The platform produces evidence; engineers still make calls. It is not a replacement for structural investigation: when a segment needs sub-base work, you still need ground-penetrating radar. And it is not a replacement for your PMS. Modern platforms feed the PMS, they don't replace it.
What continuous data is is the operating-layer shift municipal engineering has been trying to make for a decade. The technology is finally cheap enough and accurate enough to run at municipal scale. The only remaining constraint is the three-year procurement cycle, and that is a political problem rather than an engineering one.