Most buyers spend their quality attention on the ribbon itself — the colour, the print, the width, the hand feel, the certification. They are right to, because those are the properties they specify and the properties they can see in a swatch. But on a finished programme, the majority of the defects that reach a customer and the majority of the complaints that reach a retailer are not weaving defects, dyeing defects or printing defects. They are converting defects. They are created in the last step, where a flat ribbon is turned into a shaped part: a bow wing, a bookmark end, a label strip, a card wrap, a fan-fold, a petal, a loop closure. The reason is straightforward. A flat ribbon on a roll is forgiving. It is inspected under tension, on a wide web, with generous tolerance. A cut part is not forgiving at all. It has a shape, it has edges, it has corners, and it has been through a machine that applied a great deal of force to it. The cut edge is where the ribbon's construction finally gets tested. A woven ribbon with a firm selvedge cuts cleanly. A wide, loosely woven construction with a fuzzy edge cuts into a ragged line that varies along its length. A satin with a slippery finish moves under the blade. A velvet or a brushed ribbon compresses under the tool and springs back. A flocked ribbon crushes. A wired ribbon, if the wire is close to the edge, gets exposed and cuts into the operator's hand. Every one of these is a material behaviour, and none of them is visible in a ribbon swatch. Corners and small features are where conversion is won or lost. A bow with a notched tail, a label with a rounded end, a slot in a trim strip, a tiny petal shape — these are where tool geometry, material behaviour and registration meet. A part that is perfect everywhere except in the corner is a part that gets rejected. The cut part changes the handling characteristics of the whole assembly. A cut ribbon has exposed edges, and exposed edges fray. Fraying is a function of the weave, and cutting accelerates it. In a gift set that is handled, carried, shelved and shipped, an edge that frays on day one looks like a defect on day thirty. And the economics live here too. Tooling is a real, often significant, one-off cost, and the way it is amortised across an order decides whether the unit price the buyer approves is the unit price that actually applies at the volume they eventually run. A quote that hides tooling inside the unit price makes the first order look expensive and the re-order look cheap. A quote that shows tooling separately lets the buyer make a decision about volume that is actually available to them. At Smith Ribbon, in Xiamen since 2004, we run die-cutting, laser cutting and knife cutting in-house alongside weaving, dyeing, printing and finishing, because the cut part is where our work is finished, not where it stops. This guide covers how to choose a conversion route, how to specify a cut part so that it can actually be inspected, how to price tooling honestly, and how to avoid the conversion step being treated as an afterthought.
There are three commercially significant routes for cutting ribbon into parts, and they are not interchangeable. Most programmes use one as the primary and the others as contingency or for low volumes. 1. Mechanical die-cutting. A steel rule is set into a wooden or metal board, and a hydraulic or pneumatic press drives a cutting blade through the ribbon against the board's cutting surface. This is the traditional method and it remains the standard answer for volume work on structured materials. Its strengths are real. A die costs money — typically in the low hundreds of dollars for a simple part, rising with complexity, size and the number of cutting edges — but once paid, the running cost per part is very low. The cut is repeatable: the same die produces the same shape every time, which is exactly what a brand needs when a trim has to match across production runs a year apart. Throughput is high. And because the die is a physical object, the cutting edge can be specified, honed, and replaced. Its limits are equally real. The die has to be made for a specific ribbon specification — if the ribbon changes width, thickness, construction or stiffness, the die's clearance has to be correct for the new material and the part may need a re-tool. Sharp corners need a specific, deliberately radiused approach, and very small internal radii are limited by what can be ground into a rule. A die that is dulled, whether by a hard ribbon construction, an abrasive print, a metallic pigment or a stray piece of board, produces a fuzzy edge and has to be reground. And because a die is a physical commitment, die-cutting is economically wrong for short runs, for one-off prototypes, and for a programme whose shape is still moving. 2. Laser cutting. A focused beam vaporises the ribbon along the cut path, with no mechanical contact. This is the modern route and it has transformed what is possible at low volume and at fine detail. Its strengths are real. There is no tooling to buy, so it is the right answer for prototyping, for short runs, and for a programme that has not settled its shape yet. It holds very fine detail and very small features — small holes, fine internal radii, narrow webs between features — that a die rule cannot hold. It produces a consistent geometry from the first piece to the ten-thousandth, with no tool wear, which matters when the shape is complex. And because nothing touches the material, there is no die clearance to design and no blade deflection to manage. Its limits are equally real, and they are physical chemistry rather than opinions. The cut edge of a laser-cut part is different in character from a die-cut edge: it is locally heat-affected, and depending on the material it can be slightly discoloured, slightly browned, or slightly melted and re-solidified. For polyester the effect is usually small and can be acceptable; for some constructions and some colours it is visible under a bright light or against a light background. The thermal effect is greater on some fibres than others, and it interacts with the ribbon's own heat setting — a ribbon that has been heat-set can react differently from one that has not. Laser cutting also has a heat-affected zone, and on a narrow ribbon or on a delicate weave, the zone can matter. And the economics shift: laser cutting has no tooling cost but it does have a running cost per cut length, so at very high volume with a simple shape, a die will usually beat it on unit price. 3. Rotary knife and band-knife cutting. A circular or banded blade on a moving table cuts the ribbon against a moving board, and the part is either cut freehand to shape or fed through a die-cut stripper. This is the route for simple, straight or gently curved cuts on continuous ribbon — trimming to width, cutting to length, guillotine-type cuts — and for high-throughput slitting and shearing work. Its strengths are speed, low tooling cost and an almost unlimited length capability. Its limits are geometric: it is very good at straight lines and poor at tight curves and internal detail, and a ribbon with a soft or elastic handle can distort during the cut and produce a part that is dimensionally wrong even though the cut itself was clean. How to choose. Simple shape, high volume, stable geometry: die. Complex shape, low or uncertain volume, fine detail: laser. Long straight runs on continuous ribbon: knife. Most real programmes use a combination — laser for the prototype phase, then a die once the shape has frozen and the volume is real. That staged approach is usually the cheapest outcome and it is worth planning deliberately rather than discovering.
Buyers write extremely detailed specifications for the ribbon and almost none for the cut edge. Then they receive parts that look correct from a distance and are rejected on the table. Here is what a cut-edge specification actually needs to contain. Edge class. The first thing to define is what an acceptable edge looks like, because "clean" means nothing measurable. A workable way to define it is a small set of conditions the edge must satisfy. For most decorative trim programmes the useful classes are: Class A — Cut clean. The edge is continuous and straight along its length, with no visible fibre pull, no fraying at the edge, no raggedness or tears in the weave, and no loose threads. This is achievable on structured woven ribbon with a firm selvedge, on most satins, on grosgrain, and on many printed constructions, and it is what a gift bow or a retail trim normally needs. Class B — Cut clean with minor fibre pull acceptable. The edge is continuous but individual fibres may lift slightly at the edge without pulling threads out of the weave. Acceptable on many decorative and secondary trims where the edge is not handled hard. Class C — Cut with fibre pull permitted, no threads removed, no tears. The minimum. The edge may show some fibre disturbance but the weave must be intact and no loose thread ends may be present. This is realistic for very wide, loosely woven, brushed, flocked or fringed constructions. What you must not accept is an undefined edge. A specification that says "cut to shape, edges clean" without a class and without the conditions is a preference, and it will be judged against the supplier's own opinion of what clean means. Selvedge behaviour. A critical and frequently missed point: the edge quality of a cut ribbon depends on whether the cut runs along the selvedge or through the body of the weave. A ribbon cut through the body of the weave has its own yarn ends exposed at the cut line, and those ends will fray. A ribbon cut along the selvedge has the mill's own finished edge, which is a different and generally better situation — but the selvedge itself has to be good, and it does vary between constructions and between widths. Whenever the design allows, specify the cut to run on the selvedge, and where it does not, state the acceptable behaviour of the cut yarn ends explicitly. Corner and radius specification. Corners are where parts get rejected. State the minimum internal radius, the minimum external radius, and the minimum web width between features. These are not arbitrary: they set the limits of what the tooling can hold, and the die or the laser parameters that produce them. A part with a 0.5mm internal radius will be specified to a toolmaker and quoted, and it will come out at 0.5mm with the radius filled in and the corner stress-concentrated. A realistic minimum internal radius for mechanical die-cutting on ribbon is generally in the region of 1mm; laser can hold smaller, but the thermal effect grows. Tolerance. Cut-part tolerance is a different proposition from ribbon-width tolerance, because the cut introduces its own variation from tool wear, blade deflection, material stretch under clamping, and feed registration. A practical approach is to define tolerance on the two dimensions that matter most — overall length and the position of the cut feature — and to specify the rest as "to drawing" with a general tolerance that applies. State the general tolerance class explicitly rather than leaving it implicit, and state whether the tolerance is on the cut part as delivered or on the part after the receiving operation, because a cut part that is correct but relaxes when folded is not correct. Conditioning and measurement. As with any dimensional specification, the measurement conditions have to be stated. Cut parts are measured flat, under defined light tension, after a defined conditioning period, and the number of parts measured per lot should be stated. This matters more for cut parts than for flat ribbon because the geometry itself carries more variables. Burr direction and deburring. For die-cutting, specify the burr side and whether deburring is required or permitted. For laser, specify whether a heat-affected zone at the edge is acceptable and to what visual standard. These are the two lines that most often cause an argument on delivery, and both are trivially resolved at specification stage. Fraying and edge finish. If the end use has an exposed edge that will be handled, say so, and consider whether the edge needs heat-sealing, binding, over-locking or a covering trim. On ribbon constructions that fray — particularly loosely woven ones, chenille, and anything with a cut yarn core — an edge finish is often the right answer, and it should be a specified operation rather than an expectation that the ribbon will hold.
Tooling is where cut-part programmes most often get mispriced, and the mispricing surfaces later, at the moment the volume changes. The tooling cost is real and it is one-off. A simple single-part die typically runs from a few hundred US dollars upward, and the number climbs with part size, part complexity, the number of cutting edges, the number of parts in the set, the material being cut, and the tolerance required. If a programme has four sizes of the same shape, that is four tools, not one — though a multi-tool board can combine them, at a higher tooling cost and a shared wear life. State clearly in the quote what the tooling covers and what it does not, because "tooling free" in one quote and "tooling per size" in another is a common and expensive ambiguity. Tooling life is finite and is a property of the material. A die that cuts soft woven ribbon will last a great deal longer than a die that cuts a heavily woven, abrasive, or metallised construction. The edge quality degrades gradually before it fails, and the practical consequence is that a programme which runs for three years on one die will eventually see edge quality drift, and the buyer will experience it as a gradual quality decline rather than an event. The right practice is to record the die in the programme documentation with its material specification and a note that edge condition should be re-checked at a stated volume, and to have the supplier re-hone the rule rather than replace the whole board when that point arrives. Amortisation should be visible and should be a decision, not a trick. If a quote shows tooling separately, the buyer can make a real decision: run the die for a short programme and accept the unit cost, or commit to a volume that amortises it, or move to a laser route. If the quote hides tooling in the unit price, the first order looks expensive and every subsequent order looks like the supplier has cut their price, which produces exactly the wrong negotiating dynamic. A transparent tooling line, amortised across the committed volume, is good commercial practice and it protects both sides. Unit price is a function of volume, part complexity and material — and all three should be visible. Ask for the price at three volumes: the pilot, the realistic first-year volume, and the volume at which the programme would be considered successful. Most mills will quote these and it is a quick way to see whether a supplier's economics work for your programme or whether you need a different shape of agreement. Laser cutting changes the arithmetic, not the principle. No tooling, but a running cost per cut length and a cost per part that depends on complexity and on machine time. At low volume laser is usually cheaper in total cost of ownership; at very high volume on a simple shape, a die will usually win. Where you are genuinely unsure of the volume, laser is the rational starting point and the die becomes a rational later step. Count the whole tool, not just the cutter. Any jigs, fixtures, holding plates, feeding and stacking attachments, or a pattern board needed to hold a particular part on the ribbon web all count as tooling and should be quoted. The most common surprise in this area is not the die cost — it is the small fixture that nobody mentioned, which turns out to be a bespoke piece of engineering. Separate tooling from the part price and from the material price. Three lines on a quote, not one. Material, conversion, and tooling. When the three are blended, a buyer cannot tell whether a price rise came from the ribbon, the process or the amortisation, and cannot negotiate the part that actually moved.
Approval discipline on a cut programme is where the most expensive mistakes live, because a programme can be approved beautifully at ribbon stage and rejected entirely at part stage. Approve a part, not a ribbon. A ribbon sample tells you about colour, hand and width. It tells you nothing about whether the part will cut cleanly, whether the corner will hold, or whether the edge will fray. The sample that needs to be signed is a cut part in the production construction, in the production colour, at the production width, produced by the production route. If the programme is going to be laser-cut, the approval sample must be laser-cut. If it is going to be die-cut, the approval sample should be off the actual die, because die-cut and laser-cut parts from the same ribbon are different parts in every practical sense. Approve the edge explicitly, and physically. "Edges clean" in a purchase order is a dispute waiting to happen. Approve a physical standard sample — a part you have signed as the edge reference — and keep it. Then inspect against the reference, not against a phrase. Where the edge class is lower than Class A, write down what specifically is permitted, and consider requiring that loose threads be trimmed rather than left. Watch the transition from prototype to production tooling. A laser prototype approved by the brand is not automatically a die-cut production part. The two routes differ at corners, at fine features, at the thermal zone, and often at dimension. When a programme moves from laser to die, treat it as a new sample requiring approval — it is a small cost against a very large one if the change is not made deliberately. Define the inspection method for parts. Parts are inspected as parts: a count against the order quantity, a dimensional sample of a stated number of units, an edge assessment against the reference standard, and a visual assessment for the print or decoration if the part carries any. Specify the sampling plan and the acceptance criteria, and specify what happens on failure. AQL sampling conventions are widely used in the industry for this kind of inspection and are a reasonable default where you have nothing more specific. Log the cutting programme alongside the ribbon programme. Part number, shape, material spec, colour spec, quantity per set, conversion route, die or programme ID, tolerance, edge class, approval date, and the approved physical reference. When the programme grows — a fourth colour, a second size, a new market with a different retailer packaging requirement — this log is what tells you which parts need new tooling and which are already covered. Plan the first production run with the die present, not after. The first run off a new tool should be checked against the reference before the balance of the order is produced and packed. A die that is running slightly off will produce an entire pallet of parts before anyone notices, and the standard practice is to run a short first section, check it, and then continue. At Smith Ribbon, since 2004, we have cut ribbon parts in our own Xiamen plant for brand owners in 50+ countries, and the programmes that run without drama are the ones where the cut part was treated as a specified component with its own approval and its own reference sample, rather than as a finishing step on a ribbon order. We will tell you when a shape is going to be difficult or expensive to cut, and we will say so at the sample stage rather than at the tooling invoice.
Almost every cut-part problem has a design answer, and the design answer is nearly always cheaper than the quality answer. These are the ones we see repeatedly. Edge fraying in the finished assembly. Cause: a loosely woven or cut-yarn construction cut through the body of the weave. Design answers: cut along the selvedge wherever the geometry allows; specify a selvedge-backed construction; use a covering trim over the raw edge; or specify an edge finish operation — heat sealing, binding, over-locking. Choose the answer before the tooling is cut, because changing the geometry later means a new die. Fuzzing and matting on brushed, flocked or velvet constructions. Cause: the blade crushes and shears the pile. Design answers: specify a fuzz-tolerant edge class; reduce cutting speed; choose a route that supports the pile; or design the part so the fuzzy edge falls inside a fold or under a trim band. Wired ribbon damage. Cause: the wire is closer to the cut line than the die clearance accounts for. Design answers: specify a minimum wire-to-edge distance; state the wire material and gauge in the ribbon spec so the tool can be built with correct clearance; or choose a wired construction with a covered edge. This one is also a safety issue on the shop floor, so it is worth specifying explicitly. Corner distortion and tip lift. Cause: a sharp internal radius and a narrow web combined with material springback. Design answers: state a realistic minimum internal radius; avoid isolated sharp points on soft constructions; and consider a slight corner radius as a deliberate design feature rather than a defect. Marked or crushed edges on satin and high-lustre constructions. Cause: the part is clamped and the blade or the die shoe presses the surface. Design answers: specify a non-marking support surface under the cutting zone; where possible specify a cut that avoids the calendered face; and inspect against the edge reference rather than against an idealised straight line. Registration drift on printed ribbon. Cause: the print is registered to the ribbon width, and the cut position moves relative to the print as the ribbon feeds. Design answers: state the cut-to-print registration tolerance; specify a tie point so the cut is anchored to the print rather than to the roll; and require the die or laser programme to carry the registration from the printed roll. Part-to-part dimensional scatter from feed slippage on soft constructions. Cause: elastic or slippery ribbon feeding unevenly under the blade. Design answers: specify a gripper or backing support; state the tension tolerance at the cutting station; and require a first-piece check on each new roll loaded, not only once per order. Counts that do not reconcile. Cause: parts nested or stacked in a way that makes counting hard, or offcuts that are not accounted for. Design answers: specify the counting convention — pieces per carton, pieces per set, and the offcut recovery expectation — and require the count on the carton label to be verified by a second check before the carton is sealed. The pattern across all of these is that the failure was designed into the part before anyone cut it, and every one of them is preventable at the specification stage at essentially no cost. The ones that cost money are the ones discovered after the tooling has been paid for.
Not every mill that weaves ribbon also cuts it well, and cutting is a separate discipline with a separate skill base. These are the questions that separate the two. Do they cut in-house? Cutting is often subcontracted. That is a legitimate commercial choice and not a defect in itself, but you should know, because a subcontracted cutting step is a step where your tolerances are not under your supplier's direct control and where the production sequence is longer than the quote implies. Can they show you a cut part, from your own programme, in your own material? A finished part in front of you answers questions that a photograph cannot: how the edge actually looks, how the corner actually holds, how the part feels in the hand. Will they specify the edge class themselves? A supplier who quotes a cut ribbon without stating what edge quality they are committing to is not managing the variable. If they push back and say "it depends on the material", that is a good answer — but it should come with a proposed class for your material, not with a shrug. Do they tell you when a shape is expensive or difficult to cut? Honest suppliers will decline work that is not viable and will say what the problem is — a radius too tight, a material too elastic, a part too large for the machine. The ones who quote everything are the ones you find out about later. How is the tooling stored, versioned and protected between orders? A programme that re-orders in two years needs the same die. Ask where it is, how it is stored against humidity, how it is versioned against material revisions, and whether it is maintained. This is not a trivial question at a mill that cuts for many customers. What is the re-order lead time once the tool exists? This should be materially shorter than the first-order lead time, and if it is not, ask why. Do they quote tooling, conversion and material as separate lines? A supplier who will show you the three components has a cost model you can reason about. A supplier who gives you one blended number has a cost model you cannot interrogate, which means you cannot renegotiate the part that has actually moved. Will they handle the small quantities and the pilot, or only the volume? A programme is usually designed at low volume and then produced at high volume. A supplier that can do both, and can tell you where the crossover sits, is more useful than one that only wants the volume. At Smith Ribbon, in Xiamen since 2004, we run weaving, dyeing, printing, finishing and cutting in one plant — 15,000 m², 200+ staff, 100,000 metres of daily output — so the cut part is something we control end to end rather than something we subcontract and hope for. We quote tooling as a separate line, we state the edge class in the quotation, and we will tell you when a shape is a bad idea before the tooling is cut. If you are comparing a ribbon mill against a ribbon converter, that difference is worth pricing.
Die-cutting wins on simple shapes at high volume, because the tooling is a one-off and the unit cost at volume is lower, with excellent repeatability once the tool is made. Laser cutting wins on complex shapes, fine detail, short runs, prototyping and any programme whose geometry is still moving, because there is no tooling and geometry is consistent from the first piece. Most programmes use laser for development and a die once the shape has frozen and the volume justifies it.
It depends on the construction and the end use, which is why the specification needs a named edge class rather than the word clean. A structured woven ribbon with a firm selvedge can usually be cut clean with no fraying; a loosely woven or wide construction will show some fibre pull, and a cut through the body of the weave exposes yarn ends that will fray. Define the class, approve a physical reference part, and inspect against the reference rather than against a phrase.
Cutting exposes yarn ends at the cut line, and those ends are not finished the way the mill selvedge is. On a loosely woven construction the cut yarns pull out under handling. The design answers are to cut along the selvedge wherever the geometry allows, specify a selvedge-backed construction, cover the raw edge with a trim band, or specify an edge finish such as heat sealing or binding. All of these are much cheaper than a field failure.
A simple single-part die typically starts in the low hundreds of US dollars and rises with part size, complexity, the number of cutting edges, the material and the tolerance. Multi-part sets and multi-size programmes cost more because each geometry needs its own tool. The tooling is a one-off and should be quoted as a separate line so that amortisation across your committed volume is visible rather than buried in the unit price.
It can. A laser cut edge is locally heat-affected and may show slight discolouration, browning or a melted-and-re-solidified character, and the size of that effect depends on the fibre and on the ribbon's own heat-setting history. On polyester it is usually minor and acceptable, but on some constructions and colours it is visible under bright light. Specify an acceptable visual standard for the heat-affected zone, and approve a physical sample cut by the production route rather than by hand.
You should not, if the programme involves cutting. A ribbon sample shows colour, width and hand, but tells you nothing about edge quality, corner behaviour or whether the shape will hold the required tolerance. Approve a cut part produced in the production construction, in the production colour, by the production route, and keep it as the physical reference for inspection.
Use a defined sampling plan rather than an ad hoc check. A common approach is AQL sampling on the agreed characteristics, combined with a stated dimensional sample size and an edge assessment against your approved reference part. State the sampling plan and the acceptance criteria in the specification, along with what happens when a lot fails, so that an inspection result becomes a defined business event rather than a negotiation.
A ribbon programme is finished when the ribbon is cut into parts, not when the ribbon is woven. The cut is where the material's real behaviour finally shows itself, where the economics of the whole order are actually decided, and where a programme that was approved on a swatch can still fail on a table. The good news is that almost every cut-part problem has a design answer, and nearly all of them are available at the specification stage for no cost at all: name the edge class, set a realistic radius, cut on the selvedge where the geometry allows, approve a physical part rather than a ribbon, and price the tooling honestly. That is a discipline we have applied since we began weaving and converting ribbon in Xiamen in 2004, and it is why buyers who start a cut-part programme with us tend to add shapes rather than replace suppliers.
If you are planning a private-label programme where the ribbon is cut into parts — bows, bookmarks, label strips, card wraps, fan-folds, petals or closure loops — send us the shape or a drawing, the width, the material, the estimated volume and the edge tolerance you need, and we will come back with the conversion route we would recommend, the tooling cost as a separate line, and the edge class we would commit to. Email xmmsd@126.com, or reach us on WhatsApp and phone at +86 137 7995 1780 — we reply within 24 hours.