How-to · sample evaluation · pre-production approval · colour and placement checks · B2B corporate merch in Armenia

How to Evaluate a Merch Sample Before You Approve Production

Buyers usually treat a physical sample as proof, and then approve production against a photograph of it. A sample is not proof of anything; it is one unit, produced on one machine, by one operator, on one day, and it is a genuinely useful artefact within those limits. It can tell you whether the artwork prints on that garment colourway, whether the placement is where the specification says it should be, how the surface feels, whether the stitching holds together, and whether the labels are right. It cannot tell you what the run looks like at unit four hundred, how the print behaves after thirty washes, whether the size grading works across the run, or whether the cartons will be packed correctly. This guide is the procedure for getting the most out of the sample you are given and the cost you pay for it. It sets out what to prepare before the parcel opens, how to check colour under light that does not flatter it, how to measure placement and scale instead of judging them by eye, which variations are ordinary and which are defects, how to write feedback that produces a corrected sample rather than a second argument, and how to record the approval so that the production run has something to be measured against.

Request a sample against your spec Other how-to guides

What a sample proves and what it cannot

The single most common sampling error is treating the sample as a preview of the finished delivery. It is not. It is a unit of evidence with a narrow scope, and the scope is knowable in advance. Before anything is unpacked, it is worth being explicit about which questions this artefact can settle and which questions it will leave open, because that determines whether the sample is worth approving, what has to be checked in addition, and whether the programme has a risk that nobody has yet owned.

It can prove the colour on that garment colourway: Whether the design reads on the actual fabric colour rather than on white paper, whether a dark garment needed an underbase and got one, and whether the chosen colours are distinguishable at conversation distance
It can prove approximate placement and scale: Whether the mark sits where a wearer expects it, whether the artwork proportion survives the fabric, and whether the size is right for the intended position. Exact millimetre placement across a run is not provable from one unit
It can prove surface and hand: Whether the print or embroidery feels acceptable against the fabric, whether the ink sits into the weave or films on top, and whether the decoration survives being pulled at the seam
It can prove construction details: Label stitching, seam quality, thread colour, backing, closure hardware, zip function, lining, and whether components specified on the bill of materials are actually present
It cannot prove run-to-run consistency: One unit cannot reveal drift across a large batch. Ask instead what the supplier’s own acceptance check is at the end of a run, and on what basis a run is passed
It cannot prove durability: Wash and rub durability cannot be inferred from a fresh sample. What it can do is tell you whether the wrong process was specified for the fabric, which is the most common durability failure
It cannot prove size grading: One garment says nothing about whether sizes S through XXL were cut from the same block. Request a graded set when the programme depends on a fit across a large team
It cannot prove packing, labelling or delivery: Sample presentation is usually assembled by hand for the buyer. Carton planning, labelling, insert accuracy and transit risk have to be assessed from documents and from goods-in, not from the parcel the sample arrived in

Once the boundaries are clear, the sample becomes what it should have been: a fast, cheap check that catches the expensive mistakes before production starts. A sample that settles colour, placement and construction in one afternoon removes an entire category of risk from a decision that would otherwise be taken on trust.

What to sample for each service line

The correct sample depends on what the programme is buying, and a sample that is right for one service line is often irrelevant for another. The four lines below are the ones most corporate programmes run, and for each the useful sample is a different object with a different inspection routine.

Booth staff apparel

The decisive sample is a graded set, not a single garment. Ask for one item from the smallest, one from the middle and one from the largest size in the run, and check that the chest placement rises and falls with the body rather than sitting at a fixed distance from the collar seam. Check the garment itself for collar rib, shoulder seams and side seams, then confirm the decoration sits across a flat panel rather than a seam. Staff apparel is worn by people who did not choose to wear it, so comfort, sizing labels and colour consistency across the set matter as much as the mark. Wash one of the three samples before the programme if you have time; the result is more informative than any photograph.

High-volume giveaways

Sample the packaging, not just the item. A giveaway that arrives in bulk is handled in transit, in a store room and possibly through several pairs of hands, so ask for one unit packed exactly as the volume will be, and inspect it for movement inside the pack, for surface protection, and for whether the mark survives being dropped onto a desk. Then open a second sample from a different position in the packing sequence if one exists, because the unit that sat in the middle of a carton is not the same test as the one that sat on top. For a volume line, ask what they will do with the surplus held for reorders; that answer reveals whether the run was planned or bought.

Lead-capture gifts

Sample the item in the context it will be used in, which for most programmes means in a hand, in daylight, next to a phone. Small surfaces punish small artwork errors, so this line is where print registration, edge definition and colour separation are easiest to catch. Check every surface the mark occupies, including anything that curves. Check whether the item is comfortable to hold and whether it survives being put in a bag with keys in it. If the item will carry a captured detail rather than a brand mark, the fit of the printed area to the object matters more than the print itself, and it is the first thing to inspect.

VIP and speaker kits

Request the full kit, assembled as it will ship, and check it against the approved component list line by line rather than against the photograph in the proposal. Count the pieces. Open the notebook and count the pages. Check that the pen writes and that the battery cover fits. For this line, component substitution is the dominant risk, not decoration, and a full-kit sample is the only thing that catches it. Also check the outer pack itself, since a kit that arrives flat and unprotected has already lost the presentation it was bought for. One kit in hand beats any number of component swatches.

Prepare before the parcel opens

Ten minutes of preparation turns a sample review from an impression into a measurement. Do this before the box is delivered, not after the box is open, because the most useful checks need the specification and the artwork in front of you at the same time as the garment.

Have the approved artwork file open on a screen, at the same revision number that was approved. Print it at the size it will appear on the garment, or at least measure it on screen against a ruler and mark the intended print width in centimetres on the paper. Have the specification sheet open, specifically the lines covering fabric composition, weight, garment colourway, decoration method, position and dimensions, thread or ink colours, and labels. These are the lines you are about to check the physical object against, and a deviation only means something when it deviates from a written number.

Then set up the light and the surface. Inspect in neutral daylight or under a high-colour-temperature lamp, never under a warm domestic bulb and never against a wooden table or a coloured wall, because both the light and the background will shift your judgement of colour. A sheet of plain white paper and a grey card give you a neutral reference in the frame. Use a steel ruler or a tape measure, and a magnifier for anything embroidered or printed on a textured surface. Lay everything flat, and have a second person available for the colour comparison, because colour judgement is strongly affected by the eye that has just been staring at a screen.

Finally, decide in advance who signs the approval and how long that approval is valid for. A sample approval without a named signatory, a date and a statement of what it covers is not an approval, and it will be treated by everyone downstream as a general expression of enthusiasm.

Colour: the check that decides more programmes than the rest

Colour is where samples most often expose a real problem and where buyers most often wave it through. It is also the check most damaged by the conditions it is performed in. Judging a print under a warm bulb against a wooden table will make an accurate match look wrong and an inaccurate one look tolerable, so the conditions above are not a refinement, they are the method.

Compare against the garment, not against the screen. Hold the garment in neutral daylight with white paper behind it and compare the print to the colourway shown on the specification. The screen is a light source with its own colour management and it will not agree with a printed surface, which is exactly the mismatch that makes buyers distrust accurate work and accept inaccurate work.

Check the colour the design depends on. Two-colour marks that pass on a light garment frequently fail on a dark one, because a light ink on a dark field has to sit on a printed underbase and that underbase has a thickness. Look at the edge of a light element against a dark background in raking light, and you will see underbase that has not quite covered. It reads as a halo, and it is the most common rejection reason on dark garments.

Check separation at conversation distance and at arm’s length. A design that reads correctly at arm’s length may show a visible gap or a dot structure at conversation distance, which is the distance a badge or a sleeve is actually seen at. Step back to the distance of the interaction and look again.

Check the reverse. Turn the garment over. Ink that has soaked through the fabric, or a patch of cured ink on the inside of a seam, is a sign that the process was pushed harder than the cloth could take, and it usually appears on dark garments and on heavy fabrics. It is not fatal, but it tells you something about the settings used and you should know it before you approve.

Judge against a stated tolerance, not against a feeling. Ask the supplier what colour tolerance they work to and have them measure the sample. A printed colour judged by eye can differ from a target and still be inside a tolerance that everyone accepts; the same colour can look perfect to you and be outside the one the supplier works to. Getting the tolerance on the record converts the argument from taste into measurement, and it is the single most useful question in this section.

Placement, scale and orientation: measure, do not eyeball

Placement is where most informal sample reviews fail, because a mark can look right and be two centimetres wrong. Eyeballing works for a single shirt hung on a chair and fails for a garment laid flat on a table, and the difference between those two observations is exactly the number you need to write on the sheet.

Measure from a fixed reference. For a chest print, measure the vertical distance from the collar seam down to the top edge of the print, and the horizontal offset from the sleeve seam. For a back print, measure from the collar seam to the top of the mark and the overall print width. For an embroidery placement, measure the centre of the mark from the shoulder seam and from the centre line. Whatever reference you choose, it must be a seam and it must be written into the specification, because the seam is what moves between sizes and what a factory actually aligns to.

Measure the print, not the garment. Get the width and the height of the decorated area in centimetres and compare with the artwork specification. A design that has been scaled to fit a print area can also have been scaled to fit the wrong one, and the difference is visible only as a number. If the artwork was supplied as a drawing or a photograph rather than as print-ready artwork, check the aspect ratio against the original; a stretched mark on a knit is easy to miss because it looks slightly larger than expected and nobody knows why.

Check orientation on every position. Turn the garment over and check that the back is not mirrored, that a left chest mark is on the left chest as the wearer sees it, and that a cap mark reads the right way round on the front of the cap rather than the back. Orientation errors are cheap to fix on a sample and expensive on a run, and they usually originate in the artwork file rather than in production.

Check the mark against the garment’s own structure. Look at whether the print crosses a seam, sits over a pocket edge, sits above the wearer’s hand on a side position, or lands exactly where a jacket closure will cover it. These are placement questions that the specification should already answer and that a physical sample is uniquely able to answer, because they depend on how this garment sits on a body.

Stitching, construction and surfaces

Decoration and garment construction are inspected together, because a technically perfect print on a badly made garment still fails the wearer. Take the garment apart mentally: pull at the seams, look at the inside of the neck, check the labels, and check that the decoration behaves when the fabric is stretched.

For embroidery. Check thread colour against the specification under the same neutral light. Look at the reverse: bobbin threads should not be visible on the face, the backing should be present and should not be cut so close that the edge shows through, and loose threads mean the trimming was rushed. Check stitch density on a fine fabric, where a dense fill stiffens the panel and makes the garment uncomfortable. Check that a letterform with a long stroke has not narrowed or dropped a stitch, and check registration where two colours meet, because embroidery has no overlap and the second colour is laid into the first. Then stretch the garment. A print that has not been properly stabilised will pucker or distort along the stitch line when pulled.

For prints. Check edge definition under magnification, particularly on fine lines and small text, which is where a digital process and a screen process diverge most visibly. Check registration between colour layers, and on a dark garment check the underbase edge described earlier. Look for ink pooling in the middle of large solid areas, which indicates a cure or flash problem. Flex the fabric gently: ink that has not bonded into the fibre will show fine cracking, and a plastisol film will feel stiff and may crack along a fold line. Finally, check the back of the garment for show-through and for a cured-on ink mark on the inside of a seam.

For the garment itself. Check the seams by pulling them, check the collar for a comfortable rib and for a clean attachment, check labels against the specification including size and care content, and check every closure and fastener. Check zips and hardware by actually using them. On an unlined cap, check that the closure sits correctly and that the panels meet cleanly at the front. Whatever the specification says about fabric composition and weight, this is the moment to confirm the garment is that fabric, because a substituted fabric is the defect a sample is best placed to catch and worst placed to detect later.

Ordinary variation versus a defect

Buyers reject good samples and accept bad ones because neither side has written down what counts as a defect. Agreeing the table below before the sample arrives removes most of that friction, and it protects both sides: the buyer gets a consistent standard, and the supplier is not asked to remake work that met an unstated expectation.

Observation	Ordinary variation	A defect
Colour	A shade difference inside the stated tolerance, or a slight shift between units printed in different sessions	A colour outside the stated tolerance, a wrong ink or thread named on the specification, or a colour that reads differently on the garment than on the approved proof
Placement	The mark sitting inside the agreed placement tolerance, which should be stated in millimetres on the specification	The mark outside that tolerance, over a seam, mirrored, or on the opposite chest from the specification
Scale	A print filling its agreed area without touching a seam or a fold	A print scaled beyond its agreed area, distorted in proportion, or touching a seam or closure
Surface	A print that is firm rather than soft, and a hand that differs slightly from the approved garment	Cracking when flexed, visible ink on the reverse, an unprinted area inside the design, or a colour layer offset from another
Stitching	A trace of thread trimming on the reverse under the backing, and a slight difference in density between panels	Bobbin thread visible on the face, a missing backing, loose threads, a dropped stitch in a letterform, or a colour change mid-stitch
Garment	Shade variation within the dye lot, and labelling that is correct but printed slightly off-centre	The wrong colourway, the wrong size, a missing care or composition label, a failed zip, or a fabric that is not the specified composition or weight
Packaging	A presentation carton assembled slightly differently from the proposal	No individual protection on a transit-sensitive item, missing inserts, or a carton that does not match the agreed packing plan

The tolerance column is the important one. A placement tolerance that is not written down will be argued about; a placement tolerance of a stated number will be accepted or rejected. Ask the supplier what tolerance they work to and record it on the specification. If they work to a looser tolerance than your brand needs, that is information you need before production, not after.

How to reject a sample so the correction actually happens

Rejecting a sample is normal and it costs days. Rejecting a sample badly costs weeks, because a vague rejection produces a new sample that solves the wrong problem. The feedback format below works because it is short, specific, and anchored to the specification rather than to a preference.

Reference the specification line. Not “the print is too high” but “chest print: top edge to be between X and Y millimetres below the collar seam per the approved specification; measured Z.” A number anchored to a line in a document cannot be argued with later, because both parties are looking at the same page.

Describe the deviation once, without opinion. State what was found and where. Where a photograph exists, attach it with the garment in the frame and a ruler visible, because a photograph without a scale reference cannot be argued about and is usually rejected on procedural grounds alone.

State the required correction and ask whether it is a production fix or an artwork fix. This question saves the most time of any single item on this page. If the deviation originated in the approved file, correcting it at the sample stage means updating the artwork and re-proofing. Discovering the same deviation at goods-in means a reprint of the entire run. Ask explicitly, and have the answer recorded in writing.

Ask what changed to prevent a repeat. If the deviation came from a misread specification, ask for the corrected internal instruction. If it came from a machine setting, ask what the setting was and what it is now. A correction without a cause produces a correction without a fix, and the same deviation returns on the run.

Then approve the corrected sample explicitly, with a date and a reference. The approval should state which revision of the artwork it applies to and what it covers, so that a later change to the file automatically invalidates it. Keep the approved sample, label it with the job reference and the date, and store it with the specification. When the run arrives, the approved sample is the reference against which the first cartons are checked, and it is the evidence behind any claim made within the acceptance window.

Questions buyers ask about samples

Does approving a sample guarantee that the production run will match it?

No, and no supplier can offer that, because a sample is one unit made at one moment and a run is a process that drifts. What a supplier can offer is an acceptance procedure: a stated tolerance for colour, placement and dimension, a check against that tolerance at the end of the run, a stated reject threshold, and a documented decision to pass or fail the run before it ships. Ask for that procedure in writing during evaluation, because it is a far better predictor of what will arrive than the sample itself. In practice, the sensible control is to have the approved sample retained and to check the first cartons from the actual run against it before the balance is opened. A sample approval is evidence of intent, not a guarantee of conformity.

Should I order and pay for samples, or accept the ones offered at no charge?

Order one at the rate that applies to your main order rather than accepting a promotional sample. The distinction matters because a sample produced under a different rate can be produced differently, and because a paid sample places the supplier under the same commercial expectation as the order. Whichever route you take, define in advance what the sample is expected to prove, and put that definition in writing. A sample supplied without a stated purpose becomes an object that either satisfies an undefined requirement or is rejected on grounds nobody wrote down. For a garment range, request the graded set rather than one representative size, since the size question cannot be answered from a single garment.

What should happen if the sample arrives damaged or does not match the specification at all?

Treat it as an early signal about the order, not as an inconvenience. Photograph the parcel as it arrived, including the outer packaging and any seals, before opening it further. Record the deviation in the same structured form you would use for any other rejection, reference the specification line, and send it before you have emotionally invested in the garment. A supplier who replaces a damaged sample quickly and without argument has demonstrated something useful about handling a defective shipment. A supplier who argues about a damaged sample is telling you how the goods-in dispute at the end of the order will be handled, and that is worth knowing several weeks earlier.

Is Corp-Merch a partner or official supplier to any of the organisations or bodies mentioned on this page?

No, Corp-Merch is an independent B2B merchandise supplier working with corporate buyers in Armenia and the wider EAEU. References on this page to standards bodies, certification systems, testing methods and trade guidance are informational only and describe how sampling and inspection work in general. They imply no partnership, sponsorship, endorsement or official-supplier relationship with any organisation or body named here, and no trademark or certification mark is claimed as owned by Corp-Merch. Every project is quoted individually against a written specification, and nothing published on this page is a binding commercial offer.

This page is written for procurement, office, HR and marketing teams at Armenian companies who order custom merchandise and need a defensible way to inspect a sample before committing to production. Corp-Merch is an independent B2B merchandise supplier and takes no position on any other supplier named in any description, comparison or example on this page. Tolerances, measurement references and inspection routines described here are working methods, not a certification, an audit result or a standard, and the acceptance thresholds that finally apply are those agreed in your own written specification. Timings, process descriptions and any reference to testing or certification practice are typical for the category and will vary with specification, quantity and destination. Nothing on this page is a quotation, a price or a commitment to supply.