45 Year Old Parking Lot Business Proves OpEx Models Are Broken

A family business started cleaning parking lots in 1981. It is still cleaning parking lots today. Across multiple states, through five recessions, the internet revolution, mobile computing, cloud platforms, and now generative AI, the core operation has not changed. No algorithm replaced the crew. No drone took the route. The business model survived because the work is physical, local, and stubbornly resistant to scale effects that compress costs everywhere else.

The Signal

A recent Business Insider profile of a family run parking lot cleaning company lays out what should be obvious but rarely shows up in facility budget models: some operating costs never deflate. While software costs per unit have cratered over four decades and manufacturing has automated aggressively, the cost of sending a person with equipment to a specific site to clean a specific surface has tracked labor inflation almost perfectly. It has not compressed. It will not compress.

This matters right now because industrial production is grinding upward. Federal Reserve data shows the Industrial Production Index moved from 97.10 in May 2024 to 98.67 by April 2026, a modest 1.6% increase. That flat trajectory tells you something important. American industry is expanding its physical footprint slowly and deliberately, driven by reshoring initiatives and distribution network buildouts. Every one of those new facilities will need exterior maintenance. Every one of those maintenance contracts will be priced against labor costs, fuel costs, and regional market dynamics. Not Moore's Law.

That flat but upward trajectory is the context for every decision below. Industrial capacity is growing. Physical footprints are expanding. And the labor intensive services required to maintain those footprints are not getting cheaper.

The Costs That Refuse to Compress

The parking lot cleaning business has survived 45 years without a fundamental change in unit economics. That is the data point facility operators need to internalize. When your procurement team models a new distribution center or manufacturing plant, they instinctively apply deflation assumptions to technology line items. They should. Software licensing per seat drops. Sensor costs halve every few years. Automation reduces per unit labor on the production floor.

But exterior maintenance, parking lot upkeep, landscaping, snow removal, and site cleaning sit in a completely different cost category. These are labor cost plus services. The plus is fuel, equipment depreciation, and insurance. None of those inputs deflate. A reasonable planning assumption is 2% to 3% of total facility operating costs allocated to services that require a human being on your property with physical equipment. That number has held for decades.

The decision operators face is simple but frequently botched: which cost categories in your OpEx model deserve deflation assumptions and which deserve inflation assumptions? The framework is physical presence. If the service requires someone to show up at your coordinates, touch your surfaces, and exercise judgment about conditions on the ground, model it at labor inflation plus 1%. If you model it flat, your ten year projections will be wrong by year three.

National Contracts vs. Local Operators

Large facility management companies have spent two decades consolidating service contracts. The pitch is compelling. One vendor, one invoice, one relationship across your entire portfolio. But the 45 year survival of a family run regional operator tells you something about where that model breaks.

Parking lot cleaning, exterior pressure washing, and site specific maintenance are intensely local. The operator needs to know the property. They need to know the traffic patterns, the drainage issues, the municipal codes for runoff. A national vendor subcontracts that work to a local crew anyway. You are paying a margin layer for coordination that a direct relationship would eliminate.

Directors of facilities managing multisite portfolios should run a straightforward comparison. Pull the per site cost of your nationally contracted exterior maintenance. Then get three bids from regional operators in each market. The gap is typically 15% to 25%. On a 20 facility network spending $400,000 annually on exterior maintenance, that is $60,000 to $100,000 in recoverable margin. Not transformational. But real. And it compounds every year because these costs inflate rather than deflate.

The framework for the decision: use national contracts for services that require standardized technology platforms or regulatory compliance expertise. Use regional operators for services that require physical presence, local knowledge, and site specific judgment. The parking lot does not care about your vendor's enterprise software.

Reshoring Math Needs a Maintenance Line

The reshoring conversation in American manufacturing has focused heavily on construction costs, incentive packages, workforce availability, and supply chain proximity. Those are the right top line considerations. But the second order costs of operating a new facility in a new market get underestimated consistently.

When industrial production sits at 98.67 on the index, up modestly from 97.10 two years ago, the expansion is real but measured. Companies are making careful bets on new capacity. The careful ones are modeling total cost of ownership over 10 to 15 year horizons. The less careful ones are modeling construction plus labor plus utilities and calling it done.

Exterior maintenance, site upkeep, parking lot management, and grounds keeping are operating costs that vary significantly by region. A parking lot cleaning contract in Phoenix operates on different economics than one in Cleveland. Different weather patterns, different wear cycles, different labor markets. These regional variations do not show up in national averages. They show up in your actual invoices 18 months after you cut the ribbon.

The framework for reshoring OpEx models: budget site maintenance as a regional variable, not a national constant. Survey local service providers during site selection, not after construction. If you cannot find three qualified exterior maintenance operators within 50 miles of your proposed site, add that to your risk register. It signals a thin local service economy that will price accordingly.

Automation Is Not Coming for the Parking Lot

The AI conversation in operations has been dominated by two narratives. The optimistic one says automation will compress costs across every function. The pessimistic one says it will eliminate jobs wholesale. Both narratives miss the enormous middle ground occupied by services like parking lot cleaning.

The work requires navigating irregular physical environments. It requires judgment calls about weather conditions, surface damage, debris types, and customer expectations. It requires loading equipment onto a truck and driving to a specific location. Robotic solutions exist in controlled indoor environments. Warehouse floors get cleaned by autonomous machines today. But an outdoor parking lot with curbs, drains, landscaping borders, parked cars, shopping carts, and weather damage is a different problem entirely.

For capital allocation decisions, this means facility operators should not defer maintenance investments waiting for an automation solution that is not on the horizon. The operator who invests in reliable service relationships today locks in better pricing and better responsiveness than the operator who waits three years hoping for a robotic alternative and then enters a tighter labor market.

The industrial production index has held in a narrow band between 95.4 and 98.7 over the past two years according to Federal Reserve data. That stability means the demand for facility services is steady and predictable. Service providers are not facing boom bust cycles. They are building stable businesses. The 45 year old parking lot company is proof that stability is the product, not disruption.

Forward Look

Every operator building a ten year facility plan should ask one uncomfortable question: which of my cost assumptions are fantasies? The parking lot that needed sweeping in 1981 still needs sweeping in 2026. The cost of sweeping it has tracked labor and fuel inflation for 45 years without interruption. If your model assumes otherwise for any service that requires a human being on your property, your model is lying to you.

This article is part of the Operational Leverage series on NeuralPress. New analysis published daily.