How to track spare parts across a store and technician vans, record consumption per job, set reorder points, and find the jobs that are quietly losing money on parts.
A retailer's stock sits on shelves and gets counted. A service business's stock is scattered. Some is in the store, some in six technicians' vans, some already fitted at a customer site but not yet billed, and some returned as defective and waiting to go back to the supplier. All of it is your money, and only the first is easy to see.
This affects service and maintenance businesses of almost any size — air conditioning, water purifiers, lifts, pest control, electrical, plumbing, medical equipment — and the symptoms are consistent. A job is delayed because a part everyone assumed was in the van was not. A quote loses money because the parts actually used were never costed. The physical count never matches the register, and the difference is written off without anyone being able to explain it.
Service businesses rarely have an inventory problem in the warehouse sense. They have a location and consumption problem: parts leave the store, and what happened next is reconstructed from memory at month-end. This article covers how to fix that with the fewest moving parts, and what becomes measurable once you do.
The single change that resolves most of this: stop treating a part as consumed when it leaves the store. Treat it as transferred — from the store to that technician's van — and consumed only when it is fitted on a job.
One extra step: the technician confirms which parts were used when closing the job. Done on the phone they are already holding, at the site, it takes seconds. Done on a paper form typed up two days later, it will not happen reliably — and unreliable consumption data is worse than none, because people make decisions on it.
Serial-level tracking is genuinely useful and genuinely expensive in staff effort. Apply it selectively.
Getting this boundary wrong hurts in both directions. Serial-tracking consumables means the tracking gets skipped and your data becomes fiction. Not serial-tracking a warranty part means a warranty claim you cannot support when the manufacturer asks which unit failed.
A water purifier service company runs five technicians. Each carries filters, membranes, a few pumps and assorted fittings.
Before the change, the store issued parts each Monday, marked them consumed, and reconciled at month-end. Two problems recurred. Technicians would report a membrane out of stock while another van carried three, because nobody could see van stock. And the company quoted a flat rate for one service type without knowing that it consumed a pump often enough to make the rate unprofitable.
After the change, Monday's issue is a transfer to each van rather than a consumption. Closing a job records the parts fitted against that job. Within two months, two things surfaced: the flat-rate service type was running roughly 40% above its assumed parts cost and was repriced, and one van consistently showed a shortfall on fittings, which turned out to be a genuine recording habit rather than loss — the technician was fitting from a personal stock bought locally when the van ran out. That was fixable once it was visible.
For each part worth stocking you need a number that triggers a purchase before you run out:
Reorder level = average weekly usage × supplier lead time in weeks, plus a buffer.
A filter you use twelve of a week, from a supplier who takes two weeks, needs a reorder level around 30 rather than 24 — the buffer absorbs a bad week or a late delivery. Take lead time from your own purchase history, not from what the supplier claims. Revisit levels after a season change: cooling parts in April and heating parts in November do not behave like they do the rest of the year.
Once consumption is recorded per job, four figures become available that were not before, and these are the ones that change decisions.
The root cause of most mismatches, and the one change that fixes the largest share of them.
It feels rigorous and produces unreliable data, because the step gets skipped when the technician is on a roof in the rain.
Parts get attributed to the wrong job, or to no job. Per-job cost then looks precise and is not, which is worse than obviously missing data.
A hundred well-set levels beat a thousand guessed ones. Guessed levels generate purchase noise that people learn to ignore.
By the time an annual count finds a discrepancy, the cause is eight months old and untraceable. Cycle counting finds it while somebody still remembers.
Van transfers, job consumption, purchases and the customer invoice all touch the same parts. Split across systems, somebody has to reconcile them, and the data is stale between reconciliations.
The practical requirement is that a technician closing a job on their phone updates stock and the invoice in the same action. VeloCrew records job-level parts consumption against the same inventory ledger that the store and purchases use, and closes the job through the same field service workflow — so van stock, per-job cost and the customer's bill come from one set of numbers rather than three that need matching. Businesses that also run maintenance contracts will find the wider picture in our guide to field service management software in India.
Make one change first: parts are transferred to a van, not consumed, and consumed only against a job. Then choose your tracking level deliberately, set reorder points on the parts that actually stop work, and cycle count instead of running an annual event. Within two months you will know which job types are priced wrong, which is usually worth more than everything else on this list combined.
Usually because parts are marked as consumed when they leave the store, not when they are fitted. Anything sitting in a technician's van is then invisible: the register says it is gone, the shelf says it is gone, and it is actually in a vehicle. Treating each van as its own stock location removes most of the gap.
No. Track serials on warranty items, anything a customer could claim against, and anything expensive enough that one going missing matters. Use batch tracking for anything with a shelf life, and quantity only for consumables. Serial-tracking washers and cable ties means the tracking gets skipped and the data becomes unreliable.
Average weekly usage multiplied by supplier lead time in weeks, plus a buffer for a bad week or a late delivery. Take lead time from your own purchase history rather than what the supplier says it is, and set levels only for parts that stop a job when missing.
Record parts consumption against the job rather than against a date. Once that exists, parts cost per job type becomes a report, and most businesses discover at least one job type they have been quoting at a loss for years.
They do it reliably when it takes seconds and happens while they are still at the site, as part of closing the job. They do not do it reliably when it is a separate form filled in two days later. The timing matters more than the tool.
Cycle counting works better than one annual count for service businesses. Count high-value and fast-moving parts monthly, everything else quarterly, and van stock whenever the technician is at the store anyway. Fifteen minutes a week catches problems while the cause is still traceable.
Explore how VeloCrew can help you manage employees, field operations, attendance, tasks, HR and business workflows from one login — built for Indian operations teams. Every account starts on a 14-day Pro trial.
Start free ›Prefer a walkthrough first? Request a demo for your business.