Small Parts, Fasteners and Assortment Cases

A drawer of mixed hardware is not a storage failure, it is a classification failure. Sockets sort on one axis and wrenches on one axis. Fasteners sort on five, all of which matter, and any scheme that does not decide up front which axis is dominant collapses into a pile within a year.

The taxonomy problem, quantified

Consider a modest automotive hardware stock, metric only:

• Diameter: M5, M6, M8, M10, M12 (5)
• Pitch: coarse and fine for most diameters (2)
• Length: 12, 16, 20, 25, 30, 40, 50, 60 mm (8)
• Head and drive: hex, flange hex, socket cap, button, flat countersunk (5)
• Property class: 8.8, 10.9, 12.9 per ISO 898-1 (3)
• Finish: plain, zinc per ASTM F1941, black oxide, stainless (4)

The full cross product is 4,800 distinct part numbers. Nobody stocks that, but the point stands: even after pruning to the realistic combinations, a working shop stock runs to several hundred distinct items, and every one of them looks broadly like the others in a bin.

Inch hardware adds a second thread standard on top. A 1/4-20 UNC and a 1/4-28 UNF are visually near-identical and mechanically incompatible, and ASME B1.1 defines both. SAE J429 grade markings (three radial lines for grade 5, six for grade 8) and ISO property class stamps (8.8, 10.9) are the only reliable way to tell a load-bearing fastener from a hardware-store one after it leaves its box, and they are unreadable across a bench.

The two sane sorting schemes

Scheme A: geometric, for shop stock. A strict hierarchy, applied in order.

1. Thread standard (metric / inch), as a separate cabinet, shelf or bin block
2. Grade or class, as a separate row or color band within that block
3. Diameter, as a column
4. Pitch (coarse before fine), adjacent within the diameter group
5. Length, ascending within the pitch group
6. Head style, as the last split, often the one you decide not to split at all

The strength of this scheme is that finding a part requires no memory, only the fastener's specification. The weakness is that assembling the hardware for a job means visiting six bins.

Scheme B: job kits, for what travels. A closed case per recurring task, holding every fastener and small part that task needs, in the quantities that task needs, with no sorting logic beyond "this job."

The strength is that a job kit turns a hardware search into a single grab. The weakness is duplication: the same M8 x 25 bolt exists in four kits and in the shop stock, so the total inventory is larger and reordering is more complex.

The practical answer for most shops is both, with a hard boundary. Shop stock is Scheme A and lives on a wall or a shelf. Travel kits are Scheme B, they are replenished from shop stock, and they are never used as overflow storage for shop stock. The moment a job kit becomes a place to put leftover hardware, it stops being a kit.

The bin systems

The two categories worth extra comment.

Small parts drawer cabinets achieve the highest density per square foot of any option, because they eliminate the open face that bins require. They also hide their contents completely, so they demand disciplined labeling and they punish a wrong put-away severely. They are excellent for a stable, mature inventory and poor for one still being discovered.

Modular system organizer inserts are the ones that connect to the rest of a kit. If your tools already ride in a stacking system, the organizer that clips to that stack removes a whole class of loading problem. The compatibility landscape is covered in cross-brand modular box compatibility and the ecosystem differences in Packout vs ToughSystem vs Systainer vs the rest.

Bin sizing arithmetic

Bins are sold by outside dimensions and used by internal volume. Two corrections apply: bins taper for nesting and molding release, and you cannot fill to the lip without spillage. Usable volume runs about 70 percent of the outside box.

Then estimate piece volume. A fastener's swept volume is the shank cylinder plus the head, and randomly poured hardware packs at roughly 45 percent of the container volume. So:

pieces = (outside volume x 0.70) / (piece volume / 0.45)

Those piece counts are derived from the arithmetic above, not from weighing bins, and they are deliberately given as approximations. Their purpose is to stop the two common sizing errors.

Error one: bins too big. A bin holding 1,700 M6 bolts when you use forty a year is not storage, it is a permanent obstacle at the front of the shelf. It also hides the reorder point, because a large bin looks full for years and then is suddenly empty.

Error two: bins too small for consumables. The opposite failure is a bin that needs refilling weekly. Set the bin size so a normal reorder cycle drops it to roughly one third full, which makes the visual level itself the reorder signal.

Portable versus shop-fixed

The split is not about which parts, it is about which failure you can tolerate.

Loose hardware in a tool chest drawer is the worst of both. Small parts migrate under liners, into slide raceways, and eventually into every drawer below. If hardware must live in a chest, it lives in a closed case inside the drawer, never poured into the drawer itself. This is one of the drawers where a liner is a liability rather than an asset, for reasons set out in drawer liners compared.

The classic sealed-organizer failure

Gasket-sealed organizers are sold on their ingress rating, sometimes an IP claim under IEC 60529. That rating describes the seal between the lid and the case body. It says nothing about the internal partitions, and the internal partitions are where the actual failure happens.

The mechanism: divider walls in a molded or adjustable insert typically stop 1 to 2 mm short of the closed lid. That clearance exists so the lid closes without binding on a slightly warped divider. In a shop, on a bench, it is invisible. In a truck, every vertical acceleration bounces the contents, and small flat parts (washers, circlips, small screws) hop the wall. Two hundred miles later the M6 compartment contains M5s and the washers are in five places.

Fixes, in order of effectiveness:

1. Close the gap. A strip of closed-cell polyethylene or EVA foam, 3 to 5 mm thick, bonded to the underside of the lid, compresses onto every divider wall and seals each compartment individually. This is the real fix and it costs almost nothing. Cut it from foam offcuts, per the kaizen foam method.
2. Use fixed-divider cases for what travels. Molded-in walls are usually taller and stiffer than removable ones, and they cannot shift.
3. Fill compartments. A compartment filled to 80 percent has little room for parts to jump. A quarter-full compartment is a shaker.
4. Carry the organizer flat and level. Standing an organizer on edge in a van converts every divider into a shelf that hardware slides off. Mount organizers horizontally, per mounting modular boxes in a van.
5. Split by size class across cases, so that if migration does happen, the mixed parts are at least distinguishable at a glance. Mixing M6 and M8 is recoverable. Mixing M6 x 1.0 and M6 x 0.75 is not.

Labeling small bins

Small bins are the hardest labeling case: the label area is tiny, the text must encode a full specification, and the label is read at bin distance in poor light.

• Use a fixed field order and never vary it. M8-1.25 x 25 / 8.8 / ZP reads instantly once your eye knows where each field sits. Free-form descriptions do not.
• Put the discriminating field first. If a shelf block is all M8, the leading field should be pitch or length, not the M8 everyone can see from the block position.
• Encode grade with color, spell it with text. A yellow band for class 10.9 across a whole row, plus the text on each label. Color alone fails for roughly one in twelve men.
• Label the bin position on the shelf too. Then a bin returned to the wrong slot is visible, which is the same location-versus-object principle developed in shadow boards, labels and tool inventory control.
• Include the part number you reorder by. The single most useful field on a shop bin label is the supplier's SKU, because it turns a reorder into a copy rather than a lookup.
• Use durable stock. Bins live near solvents and hands. Laser-printed polyester or resin thermal transfer, not inkjet on paper.

Assortment economics

Assortment kits get criticized for containing sizes nobody uses. That criticism misunderstands what you are buying.

• A 1,000-piece general assortment in the 25 to 60 USD band as of late 2025 works out to a few cents per piece, but the distribution is set by the packer. You will exhaust three sizes and still own eight hundred of the rest.
• A bulk box of one size, say 100 pieces of M8 x 25 class 8.8 zinc, typically lands in the 10 to 25 USD band, roughly 10 to 25 cents per piece. Ten times the unit price, and every piece is one you wanted.
• The number that dominates both is the cost of not having the part. A shop hour is commonly billed in the 100 to 200 USD band, and a parts run is rarely under 30 minutes door to door. One avoided trip pays for a lot of hardware.

That arithmetic gives a clear policy:

1. Buy assortments to learn. A new shop or a new type of work does not know its own consumption. An assortment is a cheap survey.
2. Track what empties. The three or four compartments you keep refilling are your real inventory.
3. Convert those to bulk. Move the high-usage sizes into properly sized bins with a reorder line, and stop replacing them from assortments.
4. Keep the assortment as the long tail. It now serves its correct function, holding one of many things rather than many of one thing.
5. Never let the assortment be the bulk store. Refilling an assortment case from bulk boxes converts a survey tool into a badly labeled, poorly sized bin rack.

One more cost that gets ignored: weight. A full 1,000-piece steel assortment case runs on the order of 15 to 25 lb (7 to 11 kg). Two of them in a top drawer is a meaningful change to a cabinet's mass distribution and a real factor in the tipping analysis in drawer weight distribution and tip-over. Hardware is the densest thing in most shops. It belongs low.