How Do You Measure a Bag's Dimensions Correctly During Inspection?
A dimension is only correct against a stated datum, in a stated condition, with a stated tool. Bag faults blamed on size usually turn out to be two parties measuring different things. Agree the method in the tech pack and the inspection reading becomes evidence instead of an argument.

Write the method before the first unit is made: where the tape goes, what condition the bag is in, which units are used, who reads the scale, and what tolerance applies. Everything after that is arithmetic.
Dimension disputes are the most tedious category of quality argument, and almost always avoidable. Two people measure the same bag, get readings 8 mm apart, and each believes the other is wrong. Neither is. One measured across the widest point and the other measured seam to seam; one pulled the tape tight and the other let it lie; one read the bag empty and flat, the other read it as it came out of the carton. The disagreement is not about the bag at all.
This matters commercially for a reason that has nothing to do with appearance. Dimensions drive whether the product fits what the customer bought it to carry, how many units fit a carton, how many cartons fit a pallet, and what the freight quotation was based on. A dimension read carelessly at the factory becomes a claim six weeks later in another country.
What Counts as a Dimension on a Bag Specification?
More than three numbers. Height, width and depth describe the shell, but a bag specification usually carries at least seven sets of measurable figures, and each one has its own natural datum and its own failure mode.
| Measurement | Where the tool is placed | Error seen most often | What to record |
|---|---|---|---|
| Width | Square across the front panel at the stated datum — the widest point where none is named, seam to seam where a datum seam is named | Pulling the tape diagonally, or reading across a bulging pocket | The reading to one decimal place, plus which datum was used |
| Height | Vertically from the centre of the base to the stated top edge | Measuring to the top of handle piping or to a raised binding instead of the stated edge | The reading, and whether the bag was standing or laid flat |
| Depth | Square to the front panel at the stated point on the base | Reading across whatever happens to bulge, which varies by how the unit was packed | The reading and the point on the base where it was taken |
| Handle drop | From the stated top edge at the handle centre to the underside of the handle at its highest point | Measuring along the handle material rather than vertically | The reading, and whether the handle was lifted or lying |
| Strap length | End to end including the hardware, in the stated adjustment position | Recording a fully extended strap when the specification meant the middle position | The reading and the adjustment position |
| Opening or mouth | Along the stated edge, mouth relaxed, not pulled | Pulling the mouth open to read it | The reading and whether the closure was open or closed |
| Base footprint | Length and width of the base panel flat on the table | Reading it while the unit is loaded or has set unevenly | Both readings, taken on a flat surface |
Internal dimensions deserve separate treatment. A padded compartment either accepts the device the customer owns or it does not, so it is normally written as a minimum rather than as a plus-or-minus band. Writing "±1.0 cm" on an internal measurement invites exactly the reading nobody wanted: the small side of the tolerance.
Why Do Two People Measure the Same Bag Differently?
Six causes account for nearly every case. The datum is unstated. The condition — empty, flat, relaxed, settled — is unstated. The person applies tension to a soft material. The reading is taken with a cloth tape that has stretched. Units are mixed or rounded differently. And the fabric itself has not finished moving, which is a real effect with foam laminates and coated shells.

The last cause surprises buyers. A bag that has been folded, packed and shipped compressed will read differently on arrival from how it read the day it was boxed. Give it time to recover and re-read before concluding there is a problem. Conversely, do not accept "it will recover" as an answer to a reading taken at the factory on day one.
What Should Be Written Into the Specification?
A dimension clause does not need to be long. It needs to be complete in six respects: what is measured, from which datum, in what condition, in which units rounded how, with what tolerance, and by whom. Anything less, and the inspection reading is an opinion.
Six items every dimension clause must carry
- What is being measured, named explicitly — not "approximate size".
- The datum for each reading: the point or seam the measurement starts and ends at.
- The condition the unit is in: empty, laid flat, relaxed on a table, closure open or closed.
- The unit of measure and the rounding rule, applied consistently throughout.
- The tolerance per measurement, with internal usability figures written as minimums.
- Who reads the scale, and which tool — a shop-floor measurement is only comparable when both sides use the same one.
Put the drawing and the approved sampler on the same page, and say which governs. When a drawing and a physical sample disagree — and they do, more often than anyone admits — every later argument turns on which document the programme actually agreed to.
What Tolerance Belongs on a Soft Bag?
Tolerance is a commercial value judgement expressed in millimetres. Too tight and the line rejects usable units and the dates slip; too loose and the customer notices. The bands below are a reasonable starting template for cut-and-sew construction. Agree them in writing rather than inheriting them.
| Dimension | Usual construction tolerance | When it becomes a major | Reason for the band |
|---|---|---|---|
| External dimension up to 30 cm | ±1.0 cm | Beyond the stated tolerance in a direction the customer can see | Cutting variation plus one seam allowance either side |
| External dimension above 30 cm | ±1.5 cm | Beyond the stated tolerance | Longer seams accumulate positional variation |
| Internal or usability dimension | Stated minimum, not a tolerance | Anything below the stated minimum | The customer's article either fits or it does not |
| Handle drop | ±1.0 cm | Beyond the stated tolerance | A handle sitting high changes where the bag carries |
| Strap length in the stated position | ±2.0 cm | Beyond the stated tolerance | Webbing stretch and slider position combine |
| Base footprint | ±1.0 cm | Beyond the stated tolerance | A load either sits flat or it does not |
| Carton quantity | No tolerance — verify the count | Any departure from the stated count | A count is a fact rather than a measurement |
Note the direction rule. Many programmes accept tolerance in both directions; others restrict certain dimensions — a minimum internal width, a maximum external height to fit a fixture or a letterbox aperture. Write directional requirements as inequalities, not as bands.
How Many Units Should Be Measured?
Every unit in the sample. A dimension is a per-unit attribute, exactly like a seam: there is no basis for deciding which units to measure and which to trust. If the sample is 50 units at Level II on a 500-piece lot, the inspector reads all 50, externally at least.
That is less burdensome than it sounds because measurement is quick, and it produces something valuable: a distribution. A mean and a range tell you far more than a pass or fail. A lot whose readings cluster just above nominal is healthy. A lot whose readings sit exactly at the tolerance edge on every unit tells you the pattern or the machine has drifted, even though nothing has failed yet — which is precisely what an inline reading would have caught earlier.
Internal dimensions get the same treatment. Where a device fit is claimed, read it with the reference article if one was provided, rather than with a rule alone. A reading of "fits a 15-inch laptop" is worthless unless somebody owns a reference unit and puts it in the bag.
What Tools Belong on the Inspection Table?
A steel tape or a rigid rule rather than a cloth tape, because cloth stretches and then quietly disagrees with everybody. Callipers for hardware bar stock and for measuring webbing thickness. A flat table large enough for the unit. A reference box or a set of reference blocks where volume is being verified. And ideally the reference article that the customer claims the bag must accommodate.
Conditioning is the discipline that gets skipped. A coated fabric read straight off a warm machine behaves differently from the same fabric the next morning, and a laminate that has spent a night under compression recovers at a rate nobody has written down. Read the first units and the last units of a lot under the same conditions, at a similar time of day, or you will attribute to the product what actually belongs to the clock.
None of this needs a calibration certificate from an accredited laboratory for ordinary product measurement — but the tools should be checked against a reference rule at a stated frequency, and any formal test result used in a contract should come from a laboratory working to ISO/IEC 17025. The distinction matters when a report is later used as evidence.
How Does a Dimension Connect to the Carton and the Freight?
Directly enough to be worth reading twice. Carton dimensions are derived from unit dimensions and how many units are packed per carton. Cartons per pallet follow from carton dimensions. Freight is quoted on volume or weight, and frequently on whichever is greater. A millimetre on the unit becomes a centimetre across a stacked pallet.
The chained consequence is the risk: a dimension drifting upward during a run can push the packed carton past the size the freight quotation assumed, reduce the number of cartons per pallet, and change the cost per unit landed. Nobody notices at the inspection table, because everyone is looking at whether ±1.0 cm is met. Build unit dimensions into the landed cost model rather than keeping them in a separate file.
What Does a Measurement Finding Look Like in a Report?
A good report gives you a reading per unit, the datum used, the condition, the tolerance, and the class assigned. "2 of 50 units read 1.4 cm over the stated 30.0 cm width at the widest point, laid flat, closure closed, steel rule" is usable. "Slight size deviation" is not, and it is astonishing how often that is all a buyer receives.

Ask for a photograph of the tool in place for any unit that fails. If a claim ever follows, the photograph is the difference between a discussion about credibility and a discussion about facts.
What If the Drawing and the Approved Sample Disagree?
Settle it before the first lot, not after. Most programmes say the approved sample governs appearance and the drawing governs measurable figures, with named exceptions. Whatever rule is chosen, write it down, because both Readings will otherwise be produced at exactly the moment nobody can afford to wait for a third one.
The cleanest practice is a single signed document that carries the drawing, the tolerance table, and the sample identification together. Sampling takes 6-10 working days from a signed-off tech pack, and that document is what "signed-off" means — every subsequent discussion, including the inspection checklist, points back to it.
Where Do Dimensions Fit in the Defect Classes?
Mostly minor, and that surprises buyers. A unit reading 3 mm outside tolerance but usable, invisible to a customer sitting next to it on a shelf, is a minor in almost every classification. A unit reading outside a stated minimum — the sleeve the device will not enter, the carton that will not take the stated count — is a major, because function is materially reduced.
Zero tolerance does not apply to dimensions as a rule, with one exception worth naming: any dimension the regulation or the destination market controls. An aperture requirement, a small-parts gauging condition, or a cord length limit quoted in a standard is not a tolerance question. It sits in the critical list.
How Should Capacity Be Stated and Checked?
Two methods exist and they do not agree with each other. Dimensional arithmetic multiplies three external readings, which overstates what a gusseted soft bag actually carries. Volumetric measurement fills the bag with a known medium and measures what went in, which understates nothing and overstates nothing — but depends on how firmly it is filled. Say which method your figure uses. Most claims whose method was never stated turn out to have used whichever number looked better.
Where a capacity figure appears in retail-facing material, check it against the destination market's rules on substantiating a claim. An unsupported number on a hangtag is a compliance exposure long before it is a quality one.
Put the datum and the tolerance into the tech pack
Send the drawing with the measurement points marked — every point, unit, tolerance and condition comes back written into the inspection checklist before sampling starts.
Frequently Asked Questions
What unit should bag dimensions be written in?
Pick one and use it everywhere. Mixing centimetres in one part of the drawing and inches in another is a reliable source of avoidable arguments, and rounding between them adds its own error.
Should the bag be measured empty or filled?
Empty and relaxed unless the specification says otherwise. Record the condition in the report, so that a later reading taken under different circumstances can be compared honestly.
How do you measure the depth of a bag that will not stand up?
Lay it on its side on a flat table with nothing inside, read square to the front panel at the stated point, and record that you did it this way. Do not let the panels fall open or get compressed by the tape.
What tolerance applies when none was written down?
None that you can enforce. The factory will reasonably argue that any reading met an unstated expectation, and the inspector has nothing to judge against. Fix it before sampling.
Is a 5 mm difference a defect?
It is a reading. Whether it is a defect depends on the tolerance agreed and on whether it is visible or affects use. Classify by effect, not by the number alone.
How many units get measured in a 500-piece lot?
Every unit in the sample — 50 units at General Inspection Level II — because a dimension is a per-unit attribute rather than a lot-level one.
Do dimension findings count as major or minor?
Usually minor when the unit stays usable and the departure is not visible. Major when a stated minimum is breached or function is reduced. Critical only where a regulation controls the figure.
Can a small dimension change affect freight cost?
Yes. Unit size determines carton size, which determines the pallet count and the volume the quotation was based on. Build unit dimensions into the landed cost model.
Should handle drop be included in the bag height?
Conventionally no — handle drop is read separately from the top edge. Say so on the drawing, because plenty of drawings are read the other way.
What if the approved sample measures outside the drawing?
Whichever document the tech pack says governs. Settle the precedence before the first lot; arguing about it afterwards costs more than writing one clause would have.
Does the inspector need calibrated tools?
For ordinary measurement, tools checked against a reference rule at a stated frequency are appropriate. Results used as formal evidence in a contract should come from an ISO/IEC 17025 laboratory.
Should folded or rolled goods be measured flat?
Yes, and given time to recover first. A unit read while still compressed from packing reads differently from the same unit a day later.
Should internal dimensions have a tolerance band?
Prefer stating a minimum. A band implies both directions are acceptable, which is rarely what the customer meant by a compartment size.
People Also Ask
How do you measure bag dimensions correctly?
Against a stated datum, in a stated condition, with a stated tool, and with the tolerance written down. Height, width and depth are read square on a flat surface with the unit empty and relaxed.
What tolerance should be allowed on bag dimensions?
Around ±1.0 cm on external readings up to 30 cm and ±1.5 cm above that, agreed in writing. Internal usability figures are better stated as a minimum.
Why do bag measurements differ from the specification?
Usually because the two parties used different datums, different tape tension, different units, or read the bag in different conditions rather than because anything went wrong in making it.
How is bag capacity measured?
Either by filling it with a known medium or by multiplying external readings. The two give different answers, so state which method produced the figure.
What is handle drop on a bag?
The vertical distance from the stated top edge of the bag, at the handle centre, to the underside of the handle at its highest point — read vertically, not along the handle material.
Sources and further reading
Every figure on this page is either our own programme data or attributed below. Open the originals before you write them into a specification.
- ISO 2859-1:1999 Sampling procedures for inspection by attributesInternational Organization for Standardization
- ASTM International — standards indexASTM International
- ISO/IEC 17025 — Competence of testing and calibration laboratoriesInternational Organization for Standardization