Most digitization stories are sold as efficiency plays: fewer spreadsheets, faster admin, happier staff. This fleet management case study is about something more interesting. A mid-size car rental company digitized its operations expecting tidier paperwork, and instead discovered that a fifth of its fleet was barely earning while still consuming full maintenance budget.

The company, a car rental firm in the Jabodetabek area with around 85 vehicles, ran the way most rental businesses their size do: bookings in a shared Excel file, vehicle assignments decided by whoever answered the phone, maintenance on a fixed calendar schedule, and a monthly recap that took the admin three days to assemble and that nobody fully trusted.

The owner's original complaint was mundane: double bookings were embarrassing him in front of corporate clients. What the project actually delivered was a diagnosis of where his money was quietly leaking. That is the real lesson here: digitization is a diagnostic instrument before it is an optimization tool.

The Starting Point: Data That Existed but Could Not Speak

Nothing about this company was badly run by industry standards. The data existed. Bookings were written down, service records lived in a folder of nota from the bengkel, and drivers reported odometer readings, sometimes.

But the data could not be interrogated. Ask "which vehicles earned the least last quarter relative to what they cost us?" and the honest answer was two weeks of manual cross-referencing that nobody would ever do. So the question was never asked, and the fleet was managed by feel: buy more Avanzas when weekends sell out, service everything on schedule, retire cars when they start feeling problematic.

Management by feel is not stupid. It is what rational people do when getting the real numbers costs more than the decision seems worth. The point of the system was to drop that cost to near zero.

What We Actually Built

Deliberately boring technology, sequenced over about four months:

  1. A central booking system replacing the Excel file: availability calendar, conflict prevention, customer records, contract status. This alone killed the double-booking problem in week one.
  2. A digital vehicle logbook: every unit's rental history, odometer readings captured at checkout and return, and every service record with cost, entered by the admin from the bengkel nota.
  3. A simple utilization view: for each vehicle, days rented versus days available, revenue earned, and maintenance cost, per month and rolling twelve months.

Total investment landed around Rp180 juta including a year of support, plus real internal effort: about six weeks of the admin team backfilling six months of historical records. That backfill felt like drudgery and turned out to be the most valuable part, because it meant the utilization view had history from day one.

No IoT trackers, no AI, no mobile app in phase one. When a business is running on Excel, the leap to structured shared data is where nearly all the value is. Everything fancier can wait until the basics are earning.

The Utilization Report Nobody Expected

Two months in, with historical data loaded, we generated the first fleet utilization ranking. The top of the list surprised no one: the workhorse MPVs ran near 80% utilization and printed money.

The bottom of the list was the expensive discovery. Seventeen vehicles, about 20% of the fleet, showed utilization under 30%. A few were understandable: two units effectively reserved for one corporate client's sporadic needs, a luxury sedan kept for image. But the rest were simply the wrong cars: an over-supply of a model customers no longer preferred, units parked at a pool location with weak demand, and three cars that spent more days waiting on recurring repairs than earning.

Here is the multiplier that made it hurt: maintenance ran on the calendar, not on usage. Every vehicle got scheduled service as if it were working full-time. The idle fifth of the fleet was consuming close to a full share of the roughly Rp40 juta monthly maintenance spend while producing a small fraction of the revenue. Add insurance, tax, depreciation, and parking, and the owner calculated that the bottom seventeen vehicles were costing him around Rp55 juta a month against perhaps Rp20 juta of revenue.

That leak had existed for years. No one was negligent. It was simply invisible at Excel resolution.

What the Company Did With the Diagnosis

The decisions, once the numbers existed, were almost easy:

  • Sold nine vehicles over the following quarter, the chronically idle and the chronically broken, freeing roughly Rp950 juta in capital, part of which funded two more units of the high-demand model.
  • Moved maintenance from calendar-based to usage-based using odometer data the system now captured reliably. Low-usage vehicles moved to longer service intervals within manufacturer guidelines. Maintenance spend dropped about 18% with no change in breakdown rates.
  • Rebalanced the pools, shifting units from the weak location to the strong one, which the booking data made obvious within a month.
  • Set a standing rule: any vehicle under 35% utilization for three consecutive months triggers a keep-or-sell review. The report runs itself; the discipline is a ten-minute monthly meeting.

Twelve months later, fleet size was down about 8% while revenue was up 11%, and the owner described the change simply: "Sekarang saya tahu mobil mana yang kerja dan mana yang cuma numpang parkir."

The Transferable Lessons

You may not run a rental fleet, but the pattern transfers to any asset-heavy or inventory-heavy business:

  • Digitize to diagnose first, optimize second. The ROI case for the project was "fewer double bookings". The actual ROI came from a question nobody had thought to put in the proposal. Structured data reliably surfaces problems you did not know to look for.
  • Utilization is the number feel-based management gets most wrong. Humans remember the busy days and the problem assets, not the quiet middle of the distribution where money leaks.
  • Fixed-schedule costs deserve suspicion. Anything you pay "per asset per month" regardless of usage, maintenance, licenses, subscriptions, deserves a usage-based look. The same logic applies to idle servers, as I wrote in Understanding Your Cloud Bill Before It Understands You.
  • Backfilling history is worth the pain. A system that starts empty takes a year to become interesting. Six weeks of data entry bought this company an instant twelve-month view.
  • Boring beats clever. The entire result came from a booking calendar, a logbook, and one honest report.

The Takeaway

This fleet management case study reduces to one sentence: the company did not have a technology problem, it had a visibility problem, and technology was simply the cheapest way to buy visibility. The Rp180 juta system paid for itself several times over in the first year, mostly through decisions the owner made himself once he could finally see.

If your business runs on assets or stock and your honest answer to "which of these earn and which just cost?" is a feeling rather than a report, you likely have the same leak at a different scale. If you want an experienced technical partner to help you find it, that is exactly the kind of engagement I take on, see /partner. Start with visibility. The optimization tends to take care of itself.