The short answer: small-crew inventory is not a warehouse problem. It is the twenty or so items you always run out of. A mid-job supplier run costs far more than the part, once you count the drive and the lost momentum. Counting those few items before you load out takes about ten seconds and removes most of those trips.
The job stops at 10:40 a.m. for a four-dollar part.
You are under a sink, three-quarters through a straightforward install, and the half-inch fitting you were certain was in the third bin is not in the third bin. It was used on Thursday, by someone, on some job, and nobody said anything. So now the morning has a hole in it: twenty minutes to the supplier, ten in the aisle and the checkout line, twenty back. Call it fifty minutes and a burned stretch of your best working hours, for a part that costs less than the coffee you grab on the way.
Every crew has this story on repeat, and most have made peace with it as weather: annoying, unavoidable, part of the trade. It is not weather. It is an information problem (nobody knows what is on the truck), and information problems have cheap fixes.
What does a mid-job supplier run actually cost?
Price the run honestly. Fifty minutes of a $95-an-hour tradesperson is about $79. If a helper is standing semi-idle back at the job, add more. Fuel and the van's wear, a few dollars. The four-dollar fitting costs a twentieth of the labor it interrupted.
Then add the costs that do not show up as minutes:
The client watched you leave. Mid-job supplier runs read as disorganization, even when the client says nothing. It is the small-scale version of the crew that does not show up, a crack in the impression of competence you are otherwise earning.
The schedule ripples. The fifty-minute hole lands somewhere: the afternoon stop starts late, or the job needs a second visit that was never priced.
The panic buy costs more. The part you grab at the nearest supplier at retail, in a hurry, is rarely the part at the price you would have chosen while restocking calmly. Do that twice a week and it quietly shaves margin all year.
One run like this a week, a conservative rate for most small crews, is on the order of $4,000 a year in labor alone, before the ripples. That is the budget you are working with when deciding whether ten seconds of logging discipline is worth it.
The material quantity calculator runs four quick take-offs from measurements you already have.
Small-crew inventory is not warehouse inventory
Most attempts die the same way: someone decides to "get serious about inventory," builds a spreadsheet of four hundred SKUs, and abandons it within a month. Maintaining it is a part-time job nobody has.
That failure teaches the wrong lesson. The lesson is not that inventory tracking is for big companies. It is that tracking everything is for big companies. A small crew needs to track the short list of items that actually stop jobs:
The stoppers. The twenty or thirty items you have run out of mid-job in the last year. Fittings, wire, fasteners, the specific caulk, blades, filters, the breakers you use weekly. You already know this list; your last dozen supplier runs wrote it for you.
The expensive-to-forget. Items where over-buying stings: the $180 specialty part you bought twice because nobody knew one was already in the shop.
The billable consumables. Materials that go on invoices, where "how many did we use on this job?" is a money question, not just a stock question.
Everything else (the mixed box of odd screws, the one-off specialty order) is not worth a ledger entry. Deciding what not to track is what makes the system survivable. Twenty well-tracked items beat four hundred abandoned ones, always.
Three moves: Receive, Use, Adjust
Zeus keeps inventory inside the Price Book, the same list of items you already quote and invoice from, so there is no second catalog to maintain. Each tracked item carries an on-hand count, and that count only ever changes through a movement. Three kinds carry nearly all the traffic, and they map to things that already happen in your week:
Receive. Stock came in. You picked up two boxes of fittings at the supplier: receive 40. Takes seconds, ideally in the parking lot while the receipt is still in your hand (scan that too, but that is another article's story).
Use. Stock went out, onto a job. Six fittings into the Hendersons' install: use 6, against that job. This is the one that requires an actual habit, and the one that pays twice: the count stays true, and the job's costs stay true, because materials used feed what the job actually cost you.
Adjust. Reality disagreed with the ledger, and you are recording the correction: a recount found 12 where the ledger said 15, a box got water-damaged, something walked away. Adjust is the honest "here is what is actually true now," with a reason attached.
That is almost the whole vocabulary; stock you carry back to the supplier gets its own entry too. No purchase orders, no bin locations, no barcode ceremony. A crew of three can learn it in one tailgate conversation:
"New rule, ten seconds, non-negotiable: grab stock for a job, log a Use against that job before you close the van door. Pick stock up at the supplier, log a Receive before you pull out of the lot. That's it. If the count's ever wrong anyway, say so and we adjust it. Nobody's in trouble for a wrong count, only for a silent one."
The tone of that last line matters more than it looks. The moment a wrong count becomes an accusation, people stop logging honestly, and the system dies of politeness.
On-hand at a glance, and the reorder moment
The daily payoff is the unglamorous one. Before you leave for a job, the on-hand count answers "do we have enough of X?" from your phone, in the driveway, instead of from a walk to the van and a rummage through bin three. The 10:40 a.m. supplier run gets prevented at 7:10 a.m., which is when prevention is cheap.
The weekly payoff is the reorder. Instead of restocking by vibe, wandering the supplier aisles buying what looks low, you pull the inventory list and see every tracked item with its current on-hand. The items sitting at 2 when you like to carry 20 are your shopping list. Ten minutes, once a week, and the truck starts each Monday stocked against the list of things that actually stop jobs.
Two honest notes on this. First, the list tells you what you have; deciding what "low" means for each item is still your judgment. You know that fittings burn fast in furnace season and slow in July. Second, the count is only as true as the logging. If Use entries get skipped for a week, the report confidently shows stock you do not have, which is worse than ignorance. That is why the habit is framed as non-negotiable and why Adjust exists: a Friday recount of the top ten items takes five minutes and pulls the ledger back to truth.
Why not just edit a single number?
This is a design detail that sounds technical and is actually about trust.
The obvious way to track stock is a single editable number: fittings, 15. When someone takes six, they change it to 9. Simple, and it fails in a specific, predictable way. When the count is wrong, and it eventually will be, an editable number can tell you nothing about why. Was it a miscount at intake? A job that never got logged? Two people "fixing" the number in opposite directions? The number just sits there, wrong, with no history, and after the second bad experience nobody trusts it, and after nobody trusts it, nobody maintains it. That is the lifecycle of most inventory spreadsheets.
Zeus stores movements instead, as a running ledger. Every Receive, Use and Adjust is a permanent entry: who, what, how many, when, against which job. The on-hand count is just the running total of that history. Only the newest movement can be corrected or removed in place; anything older is fixed by a new entry that says so, the way an accountant reverses a bad line rather than erasing it.
What that buys a small crew, concretely:
Discrepancies become answerable. The ledger says 15, the bin says 9. Scroll the movements: Thursday shows a Receive but no Uses, and Thursday was the day of the big install. The missing entry is found, logged, done. A two-minute reconciliation instead of a shrug.
Job costs get their materials. Because Use entries name a job, each job accumulates its real materials consumption. When you look at what a job actually made you, the fittings are in the math instead of vanishing into general overhead.
The count stays trustworthy under many hands. With three people touching stock, an editable number is a rumor. A ledger where every change has a name and a timestamp is a record. Nobody has to remember; the history remembers.
Shrinkage becomes visible. If Adjust entries keep writing stock down with no story, you have learned something real about a bin, a site, or a person. Uncomfortable, but far better known than unknown.
What this is not
Keep expectations straight, because oversold inventory systems are how crews end up back at the spreadsheet graveyard.
This will not count anything for you. There is no magic. The ledger knows what you tell it, ten seconds at a time, and a crew that will not log Uses will not be saved by any tool. Start with ten items, not fifty, and let the habit prove itself on the things that hurt most.
It is deliberately light. If you are running serialized equipment, multiple warehouses, or hundreds of fast-moving SKUs with purchasing workflows, you have a real warehouse problem and should buy warehouse software. Zeus's inventory is scaled for the truck-and-a-small-shop reality: the twenty stoppers, a handful of movement types, one list you can read in a minute.
And it will not stop the occasional supplier run. Jobs surprise you; that is the trade. The goal is not zero runs. It is turning the routine stockout, the one caused purely by nobody knowing, into a rare event. That one is fully preventable, and it was always the expensive kind, because it was never priced into the day.
Frequently asked questions
How many items should I start with?
Ten to twenty. Write down every item that has caused a mid-job supplier run in recent memory. That list, not a catalog, is your starting inventory. Add items only when they earn it by stopping a job or by being expensive to double-buy. A short list maintained beats a long list abandoned, and you can grow it any week.
My helper will never remember to log Uses. Now what?
Attach the log to a physical moment, not to memory: stock leaves the van, entry goes in, door closes. Same reflex as locking the truck. Keep the rule blame-free for wrong counts and firm about silent ones. And recount your top ten on Fridays for the first month; the gap between ledger and bin tells you honestly whether the habit is landing, while the movement history shows you exactly which day it slipped.
What about material bought for one specific job? Should that pass through inventory?
Usually no. Special-order material that arrives and goes straight into a job the same week can be booked to the job directly as a cost; running it through Receive and Use adds ceremony without information. Inventory earns its keep on shared stock: the items many jobs draw from, where the question "how much is left?" has no owner. That distinction keeps the system small.
Does this work when the crew has no signal?
Yes. Movements logged in a basement or a rural site record on the device and sync when coverage returns, so the ledger catches up on its own. The one habit that matters there is logging at the moment of use anyway, rather than "when I'm back in service," because the entry you postpone is the entry you forget. The fuller offline story is covered in its own article on this blog.
ZEUS