Hardware store inventory software has to do four things that a general retail system was never asked to do. It has to load tens of thousands of items from your distributor's catalog and keep their costs, UPCs, and discontinued flags current without anyone retyping them. It has to sell one item in several units, so a box of deck screws can be bought by the case, sold by the box, and sold loose by the pound while the system keeps a single count. It has to build purchase orders from reorder points across thousands of slow movers, because nobody walks every aisle before the weekly truck. And it has to let you count the store while it stays open, so the drift between the shelf and the screen gets caught before it shows up as a stockout or a write-off. If a system handles three of the four, the fourth will find you on a busy Saturday.
Why a hardware catalog breaks generic inventory tools
A clothing boutique might carry a few thousand items, and a coffee shop fewer than a hundred. A full-line hardware store carries tens of thousands, and most of them sell a handful of times a year: the specific brass hinge, the 3/8-inch coupling nut, the replacement cartridge for a faucet that went out of production in the last decade. The store makes its living on that depth, because the customer who finds the odd part at your counter is the customer who skips the drive to the big box next time.
That depth is what breaks generic tools. Typing 20,000 items by hand is not a project anyone finishes, so the catalog has to come from your distributor. Fasteners, chain, rope, wire, and pipe are bought in one unit and sold in another, so the item file has to understand units of measure. Reorder points set one at a time do not scale past the first few aisles, so the system has to suggest the order. And with that many items on the floor, counting everything once a year guarantees that the numbers are wrong for eleven months. We wrote up the four moments a hardware store outgrows Square or Clover, and three of the four are inventory.
Load the catalog from your distributor, then keep it current
The right way to build a hardware item file is to pull it from the distributor you buy from, whether that is Orgill, Emery Jensen, or House-Hasson, and then add items to your store as you decide to stock them. A good catalog load brings in the UPC, the description, your cost, and the suggested retail, so adding a product is a search instead of a data-entry session, and when a customer asks whether you can get something, the answer is in the catalog instead of on hold with the warehouse.
Loading is the easy half. The half that decides whether your inventory stays accurate is upkeep, because distributors change costs, swap UPCs, and discontinue items every week. Ask any vendor three questions about it. Do cost changes flow in automatically, and can you review them before they hit your shelf prices? Do discontinued flags arrive, so you stop reordering something that is already on its way out? And when you place an order, does it go to the distributor electronically and come back as a shipment you can receive against, so the purchase order and the truck agree line for line? If you are an Orgill dealer, our guide to ordering from Orgill walks that whole loop, and the EDI guide explains the documents underneath it.
Broken packs: one box of deck screws, worked
Broken packs are where hardware inventory goes to get lost. You buy a case, sell some boxes whole, open one box into the bulk bin, and sell the rest by the pound to the customer who needs a handful for a railing repair. If the system has a separate SKU for the case, the box, and the loose screws, every one of those moves needs someone to adjust three counts by hand, and nobody does, so by spring the case SKU shows twelve on hand and the shelf shows none.
The fix is one product with one unit you count it in, and every other unit as a conversion. Here is one item, a 3-inch exterior deck screw, run end to end with illustrative numbers. The counting unit is the pound, a box is 5 pounds, and a case from the distributor is 4 boxes, or 20 pounds, at a cost of $90 a case, which works out to $4.50 a pound. The sealed box sells for $30 and loose screws from the bin sell for $7.50 a pound.
| Step | What happens | Quantity | On hand (lb) |
|---|---|---|---|
| Receive | Two cases come off the truck | +2 cases = +40 lb | 40 |
| Sell | A deck builder buys three sealed boxes | -3 boxes = -15 lb | 25 |
| Break a box | One box gets opened into the bulk bin | no change | 25 |
| Sell | A homeowner buys a scoop for a railing | -1.5 lb | 23.5 |
| Sell | Another customer buys a bigger scoop | -2.25 lb | 21.25 |
| Count | Three sealed boxes (15 lb) plus the bin weighed at 6 lb | 21 lb counted | 21 |
Three things are worth noticing. Breaking the box moved nothing, because the system already counted the screws by the pound, so opening the box is a shelf change rather than an inventory transaction. The count came up a quarter pound short, which is a few dozen screws that spilled, got miscounted on a scoop, or went home in a pocket, and that is the size of variance a fastener bin produces in a healthy store. And the box price and the loose price differ on purpose, because the bin costs you labor and spillage that the sealed box does not, so the box needs its own price rather than five times the per-pound price.
That last point is where setups go wrong. When a sealed box carries its own price, double-check the cost side of the conversion, because a box set up as the wrong number of pounds will make the margin read negative on an item that is making money, or rosy on one that is not. Our units of measure guide covers the same setup for wire off a reel, rope by the foot, and chain, so we will leave the rest of the fastener drawer there.
Reorder points across thousands of slow movers
The reorder point formula is average demand over the time it takes to get more, plus a cushion of safety stock. For a hardware store on a weekly truck, "the time it takes to get more" is the lead time plus the week you wait for the next order day. Take the deck screws: they sell about 15 pounds a week in the season, the truck takes a week, and you order once a week, so you need two weeks of cover, or 30 pounds, plus a cushion of 10 pounds, for a reorder point of 40 pounds. When on hand plus what is already on order drops to 40 pounds, two cases go on this week's order. A setting that rounds purchase orders up to full cases keeps the system from ordering 17 pounds of screws your distributor ships only by the case. The full math, including how to size the cushion, is in our reorder point and safety stock guide.
That arithmetic is worth doing for the few hundred items that drive your sales and your contractor accounts. It is not worth doing for the 15,000 items that sell a few times a year, and that is where most hardware stores either overbuy or give up. Run the long tail on min and max levels instead: a min of one and a max of two on the brass hinge, a min of two and a max of four on the faucet cartridge, so the system restocks what sold without anyone calculating standard deviations for a toilet flapper. Then have the system build the suggested purchase order from those levels and from recent sales history, and spend your time reviewing the order rather than writing it. A forecast that compares recent weeks with the same weeks last year earns its keep on seasonal goods, since a reorder point set for ice melt in January will bury you in it by March.
Counting the store while it stays open
No hardware store wants to close for a weekend to count 20,000 items, and a store that does it once a year is working from wrong numbers most of the year. Cycle counting splits the job into small slices on a schedule: the fast, high-dollar items monthly, the middle of the store quarterly, and the slow tail once or twice a year. Counting by aisle or by vendor section fits a hardware floor well, and counting a vendor's section the week before its order goes out catches the drift exactly when it would cost you. Our cycle counting guide lays out the schedule.
Two habits make the counts worth more than the time they take. Count blind, so the person counting enters what is on the shelf instead of confirming whatever number the screen shows. And count in the same unit you stock in, so the fastener bin gets weighed or counted by the pound if that is how the item is set up, rather than estimated in boxes. The variance those counts turn up is your measured shrink, and our shrink guide shows how to read it.
What to hand the AI, and what to keep
With tens of thousands of items, the most useful thing software can do is point you at the few hundred that need attention this week. That is a good job for AI, as long as it reads the same sales and inventory records the register writes rather than a copy from last month. The questions worth asking it are the ones you would ask a sharp inventory manager: where cash is tied up in dead stock, which items went negative this week, which costs changed on the last delivery, and what the suggested order looks like for each vendor.
Keep the decisions. AI can tell you a hinge has not sold in a year and what it ties up, but it does not know that the contractor who buys it twice a year also buys a pallet of decking every spring. The dead stock guide has the routine for working that list once you have it.
Eight tests to run in a hardware inventory demo
Bring your own items and your own distributor to the demo, and ask the rep to run these live rather than show slides:
- Add ten items from your distributor's catalog. Pick a mix you stock, from a fastener to a plumbing part, and time it. It should be a search and a click per item, with UPC and cost filled in.
- Show a cost change coming in from the distributor. Ask where you review it and how it reaches shelf prices and labels.
- Set up a box of screws sold by the box and by the pound. One product, one count, a separate box price, and a partial-pound sale that rings up cleanly.
- Receive a case and sell it three ways. Watch the on-hand count after each step and compare it with the table above.
- Build a suggested purchase order for one vendor. Check that it uses your min and max levels and sales history, rounds to full cases, and lets you edit before sending.
- Send the order electronically and receive the shipment against it. The truck should arrive as a list you confirm, not a packing slip you retype.
- Run a cycle count of one aisle from a phone while the store is open. Confirm it starts blind and posts the variance when it completes.
- Ask the system where your cash is tied up in dead stock. The answer should name items and dollars from your data, not a generic report you have to filter for an hour.
How Rundoo handles it
Rundoo is the AI-first POS for independent supply stores, and our hardware store POS page shows how it fits a deep aisle. On the catalog side, Orgill's item master files load into Rundoo with UPCs, item names, and descriptions ready to import, purchase orders go to Orgill over EDI with your store's pricing on the order, shipment notifications load automatically for receiving, and price changes, warehouse inventory with discontinuation flags, and item maintenance updates flow in, with price changes reviewed and applied from the price changes page when you are ready. The Emery Jensen catalog loads from EJD's ElecCom system with UPCs and item details, and the House-Hasson catalog loads with UPCs, descriptions, costs, and suggested retail, with Rundoo checking for updates every 15 minutes.
Units of measure are conversions on a single product record, set up in Admin > Product > Units of measure and on each product, with separate defaults for stocking, pricing, sale, order, and transfer. Decimal quantities are a per-product switch, a unit can carry its own price under Alternate pricing, and a Case quantity setting rounds purchase orders to full cases. One plain limit: when one distributor sends boxes and another sends pieces for the same product, unit alignment on EDI still has edge cases being worked through, so check those items after the first order.
The purchase order builder offers Replenishment, Min & Max, and Predictive methods per vendor, and Predictive compares the last few weeks this year with the same weeks last year to build the season into the suggestion. Cycle counts can be scoped by vendor, category, product line, or product tag, run from the Staff app on any iPhone or Android by scanning, and start blind. Another plain limit: Rundoo has no dedicated shrink transaction or shrink reason codes, so measured shrink comes from counts, and damaged units you keep go through a store-use account. And Dooey, the AI on every screen, answers questions like "Where is cash tied up in dead stock?" from the same records the register writes.
