Carry-on quality control should be aimed at the six component groups the customer actually operates: the main zipper, wheels and housings, the trolley handle, the shell hinge zone, face fabric and stitching, and hardware finish. Almost everything else in the bag is static — these six decide whether the product comes back.
Ask anybody who has run a luggage repair desk, and you will hear the same short list come back: zippers, wheels, trolley handle, cracked hinge zones, blown seams, flaked plating. Buyers rarely return a bag because the colour was two points off. They return it because something they touch every day stopped working in month three.
That list is not random, and it is not bad luck. It is the shadow of a spec sheet — usually somebody else's spec sheet, inherited unchanged, never written down. This article walks through the six component groups that decide whether a carry-on comes back, what "good" looks like for each one, and the test that catches the failure before your customer does.
Why the failure list is so short
A carry-on has a few hundred parts, but almost none of them cycle. The shell is static. The lining just sits there. The parts that fail are the ones the customer operates — the zipper they drag shut twice a day, the wheels that take every curb, the handle they extend with one hand while holding a coffee. These are wearing parts, and wearing parts live and die by tolerance and material grade, not by styling.
Which leads to the uncomfortable conclusion: most returns are designed in, not made in. A factory running at acceptable quality will still deliver thousands of failures if the zipper spec is one grade too light, because no amount of inspection upgrades a component that was never built to last.
1 · Zipper and slider: the number-one claim
Across every market we have worked, the main-opening zipper generates more warranty claims than any other single component. The failure modes are predictable: coil chains that separate under a packed bag, sliders whose autolock wears out and creeps open, and tape that frays where the stitching bites it at each end of the opening.
What to specify: chain type (coil versus metal, and why the answer differs between a hardside perimeter zip and a softside main opening), chain size for the main opening — this is one place where stepping up one size is cheap insurance — autolock sliders rather than non-locking ones, tape grade matched to the fabric, and bar-tack reinforcement at both ends of the opening where stress concentrates.
What to test: a reciprocatory (cycle) test on the finished zipper assembly, and a slider lock-strength check. Cycle counts in brand test matrices commonly run from several hundred to a few thousand cycles depending on product grade — the number matters less than having one, in writing, that both sides accept.
2 · Wheels and housings: where "durable" goes to die
Wheel complaints rarely mention the wheel. They say "it pulls to one side", "it sounds loud on pavement", or "the wheel fell off". Underneath those symptoms: a tread compound that flat-spots, a bearing that seizes after a wet terminal, or — most common of all — a housing that cracked away from the shell after a curb impact, because it was screwed into plastic instead of through a metal reinforcing plate.
What to specify: tread material and hardness, bearing type, and above all the fixing method — housings mounted through a metal plate into the shell structure, not into bare plastic. If a supplier offers a "replaceable" wheel system, test the replacement claim literally: order one spare set and fit them with the tools a customer would have.
What to test: tread abrasion or drum-run testing for wear, and a housing load test that simulates the case being yanked up a kerb by the trolley handle, fully packed.
3 · Trolley handle: the part customers touch most
The trolley system is the one component the customer operates with one hand while distracted, so small defects are loud. Bent tubes from gate-checking, a button lock that stops catching, and side-to-side wobble that makes a well-made case feel like a market stall — that last one is a tolerance issue, not a strength issue, and it is almost never written down unless somebody insists.
What to specify: tube wall gauge, number of lock stages, and an explicit maximum wobble or play figure at full extension. The play tolerance is the single most skipped line in carry-on spec sheets, and it is the one customers judge in the first ten seconds.
What to test: an extend-retract cycle test performed with a side load applied — cycling the handle vertically with no load tells you nothing about how it behaves in a terminal.
4 · Shell and hinge zone: the hardside failure map
Hardside shell material choices — polycarbonate, polypropylene, ABS and blends — trade impact behaviour, flex and weight against each other, and the right answer depends on your market. But whatever the material, the geography of failure is the same: cracks start at the hinge zone and at the corners, after impact, and they start more readily in cold weather for materials whose impact performance drops with temperature.
What to specify: material by name, not by look; sheet gauge; and an explicit hinge-zone reinforcement detail. If you sell into markets with real winters, ask what the material does at low temperature in a loaded drop, and put the answer in the spec.
What to test: a loaded drop test from practical handling heights onto edges and corners — not just flat — and, for cold-climate markets, a low-temperature impact check.
5 · Fabric and stitching: face fabric, lining and the handle root
A seam is not a component — it is what two pieces of fabric become after they are sewn together. The component is the fabric itself: the face fabric on a softside bag, and the lining on a softside or a hardside one. Specify the fabric and the seams that join it follow; specify only the seams and you are testing the symptom.
On softside product the failure map shifts to the fabric and, above all, to the handle root: the point where the carry handle or shoulder strap anchoring meets the body. Every lift of the bag is a pull on the fabric at that point. When it tears out, the customer's review writes itself.
What to specify: the face fabric by construction and weight — yarn type and count, weave, and any coating or lamination that carries the water-resistance claim — plus a lining material with a stated tear strength. Then the way it is joined: stitch density (stitches per centimetre, per seam type), bonded nylon thread rather than staple spun thread — it survives abrasion and resists rot — and bar-tack reinforcement at every stress point: handle roots, strap anchors, zip ends, and where webbing crosses a seam.
What to test: seam slippage on the main fabric, tear strength on the lining, and a pull test on the loaded carry handle that cycles the load rather than just holding it.
6 · Hardware finish: how a bag reads "cheap"
None of this makes the bag stop working, and that is exactly why it gets skipped. But plating that clouds or flakes after a humid summer, a zipper pull that loses its coating, or a rivet that loosens under load — each one tells your customer that the bag was costed down to the cent they could not see. Returns at renewal time start here.
What to specify: plating class with a salt-spray exposure requirement for metal hardware, and rivet-versus-screw decisions on anything load-bearing. None of these lines is expensive. All of them are visible within the first season of use.
What to test: salt-spray exposure on plated hardware from the actual production run — not the showroom sample.
From failure list to spec sheet
Every item above becomes a row in a table your supplier signs. The table has four columns and no adjectives:
| Component | What goes in the spec line | Test to demand |
|---|---|---|
| Main zipper | Chain type & size, autolock slider, tape grade, bar-tacked ends | Cycle test; slider lock strength |
| Wheels | Tread material & hardness, bearing type, fixing through metal plate | Tread abrasion / drum run; housing load |
| Trolley handle | Tube gauge, lock stages, maximum play at full extension | Extend-retract cycles with side load |
| Shell & hinge | Material by name, sheet gauge, hinge-zone reinforcement detail | Loaded drop (edges & corners); cold-impact if relevant |
| Fabric & stitching | Face fabric construction & weight, lining tear strength, stitch density, bonded nylon thread, bar tacks at stress points | Seam slippage; lining tear; handle-root pull |
| Hardware finish | Plating class + salt-spray requirement, fixing method for load-bearing parts | Salt spray on production parts |
Two disciplines make the table work. First, every acceptance criterion is a number or a named method — "high quality" is not a spec. Second, the sheet is revision-controlled: when a component changes between pilot-run and bulk, the sheet changes with it, and the old version is retired. Most "the sample was fine but the bulk is not" stories are revision-control stories.
The shortcut: ask for the factory's own library
Before your pilot run, ask the factory to show you their internal spec sheets for the components above — the ones they hold against their own production, not the one they typed for your enquiry. A factory that keeps a component spec library, and updates it when something fails in the field, is a factory you can build a brand on. A factory that can only produce a quotation has told you something too.
This is the working reference behind our Luggage Spec Library stage: specify the six groups, test the failure each spec prevents, and keep the sheet alive. The checklists and templates in Free resources give you the starting table.
Frequently asked: carry-on quality control
Building a spec sheet for a new carry-on programme?
Send us your current tech pack — or just the product you are planning. We will tell you which of the six groups is thinnest in it, what we would add before pilot-run, and why. Whether or not it turns into a project, you keep the answer.
Start a Conversation →