Shisha buyer guide · Article 212
How Many Shisha Charcoal Cubes Should Buyers Measure?
Choose measurement counts around the decision, lot coverage, observed variation and method repeatability. Keep distinct cubes separate from axes and repeat readings.
01 · Start with the decision
There is no useful universal cube count.
The right shisha charcoal cube measurement sample size depends on what the buyer needs to decide, which stock the sample represents, how much variation is expected and how measurements will be used. A quick comparison of identified samples, an investigation of a dimension concern and a formal lot acceptance inspection require different plans. Write the question before choosing a number.
Define the characteristic: individual edge dimensions, cube-to-cube consistency or the count of cubes outside an agreed specification. Estimating an average dimension is different from assessing the frequency of nonconforming cubes. A count selected for one purpose should not silently become an acceptance rule for another.
This guide provides a buyer planning worksheet, not a fixed sample-size table. It does not prescribe an ISO or AQL procedure, legal dimensional tolerance, guaranteed confidence level or current Unimax test result. Confirm the applicable specification and inspection scope for the inquiry.
02 · Define the population and coverage
More cubes from one carton do not cover missing lots.
Record product reference, nominal cube size, specification revision, lot identity and the cartons available for selection. Define whether the decision concerns an evaluation sample, one production lot or several separately identified lots. Keep the boundary visible in both the plan and the final report.
Separate three counts: cartons selected, distinct cubes measured and readings taken on those cubes. Measuring three axes of one cube gives three readings on one cube; it does not create three independently selected cubes. Repeating a reading on that cube increases the reading count without adding product coverage.
Lot size is part of the planning information, but a simple percentage of the shipment is not a universal solution. Distribution across lots and carton positions, the selection design and the purpose of inference also matter. An apparently large cube count may still omit an entire lot.
Use the carton-and-lot sampling guide to map selection locations. Preserve carton and cube IDs so results can be reviewed within and across their source groups. If access limits coverage, document the limit rather than presenting the accessible cartons as the whole shipment.
03 · Match effort to the consequence
Use risk to define the evidence needed, not to invent certainty.
Identify what a wrong decision would affect: a sample comparison, a customer’s dimensional requirement, a release decision or an unresolved complaint. Record relevant changes such as a new product configuration, an unfamiliar lot or inconsistent earlier observations. These are reasons to review the plan with the responsible QC team, not automatic proof that a particular count is sufficient.
For an exploratory sample review, agree enough identified observations to answer the limited comparison question and report it as exploratory. For an order release, establish the applicable decision criteria and sampling procedure before measuring. Where a named acceptance standard is required, ask the responsible inspector to confirm its applicability and the actual plan parameters.
Define how unexpected findings trigger review, additional collection or a hold. Keep the original measurements when the plan expands. Do not measure until an attractive average appears, then discard the earlier observations; that changes the evidential basis of the decision.
Any count proposed by a supplier, buyer or inspector should come with a rationale: the population, selection basis, characteristic, decision rule and limitations. A round number without those details is a workload estimate, not a demonstrated confidence claim.
04 · Review observed variation
Keep the distribution visible alongside the average.
Retain each cube’s readings, units and identity. Review the spread and any differences between carton or lot groups rather than reporting only a pooled mean. Two groups can share a similar average while containing different individual dimensions; an average alone cannot establish that every cube meets a limit.
If observations differ materially across groups, investigate the source before pooling them. Check identity, method and measurement consistency as well as possible product differences. Decide whether separate reporting or a revised collection plan is needed, and record the reviewer’s reasoning.
For a formal estimate, required precision, confidence and assumptions about variability inform sample-size planning. Relevant prior data or an identified preliminary study may help inform that design. Do not borrow a variability figure from another product or an unexplained report and treat it as established for the current lot.
Confidence statements require the corresponding statistical design and assumptions. A larger convenience sample does not automatically justify a shipment-wide confidence interval, and zero observed nonconforming cubes does not prove that none exist in the unexamined stock.
05 · Check measurement repeatability
Separate measurement variation from product variation.
Agree the measuring instrument, measurement locations and orientation, units, recording precision and treatment of chipped or irregular edges. Identify the instrument and the relevant verification or calibration record. Its display resolution alone does not establish its accuracy or the suitability of the complete method.
Use a documented repeatability check on identified cubes: repeat the agreed measurement under the same defined conditions, retaining the readings. If the repeat readings vary enough to affect the intended decision, resolve the method or instrument issue before interpreting small cube-to-cube differences.
Keep repeated readings linked to the original cube ID. When different operators or instruments are involved, record that change and compare the measurement process as appropriate. Repeated measurements of one cube answer a measurement-consistency question; they are not extra independent product samples.
If a result is close to an agreed dimensional limit, apply the agreed decision procedure, including any treatment of measurement uncertainty. Do not invent a rounding allowance or automatically call the result a pass. Record an unresolved borderline result as requiring review.
06 · Compare scenarios before committing the count
The same measurement workload can answer different questions.
Scenario A: a buyer receives an identified pre-order evaluation sample and wants to compare dimensions with a stated specification. The measurement plan can support that sample comparison. It cannot establish the variation of an unseen production lot, regardless of how many repeat readings are taken.
Scenario B: an inspection covers multiple documented lots. Collecting all cubes from one accessible carton leaves other lots unrepresented. Review the carton-and-lot coverage and keep separate source identities before deciding how many distinct cubes to measure from the selected material.
Scenario C: repeated readings on the same identified cube disagree near the decision boundary. Increasing the number of other cubes does not by itself resolve the measurement-method concern. Check repeatability, instrument suitability and the decision procedure, then document the justified next step.
These scenarios intentionally set no universal cube count. Complete the worksheet with the buyer and inspector, approve the rationale before collection and report the actual counts achieved. If the original objective cannot be supported by the available evidence, narrow the conclusion or revise the plan instead of overstating certainty.
Buyer tool · define the count and its rationale
Measurement-count decision worksheet
Inspection ID ___; company / market ___; plan revision ___; buyer / inspector ___; review date ___. Agree the purpose and method before entering the proposed counts. This worksheet does not calculate a certified sample size.
| Planning area | Required definition | Evidence / rationale | Control / next action |
|---|---|---|---|
| Decision / characteristic | Comparison / investigation / acceptance ___; characteristic ___ | Specification revision ___; decision rule ___ | QC owner ___; scope agreed ___ |
| Population | Product / nominal size ___; lot IDs ___; lot size / units ___ | Carton population ___; accessible stock ___ | Excluded or unknown stock ___ |
| Selection coverage | Cartons selected per lot / position ___; selection method ___ | Source list ___; cube-to-carton IDs ___ | Access gaps / replacement rule ___ |
| Distinct cube count | Proposed cubes per source group ___; total distinct cubes ___ | Rationale ___; inspector / method reference ___ | Do not count axes or repeat readings as extra cubes. |
| Risk / variation basis | Decision consequence ___; prior or preliminary data ___ | Observed group differences ___; limits of prior data ___ | Review or expansion trigger ___ |
| Measurement method | Instrument ID ___; method / orientation ___; units ___ | Verification record ___; edge-damage treatment ___ | Operator ___; recording precision ___ |
| Repeat readings | Cube IDs ___; repeated reading plan ___; conditions ___ | Observed repeatability ___; method issues ___ | Resolve measurement concerns before inference. |
| Confidence / limits | Descriptive review or formal inference ___; assumptions ___ | Precision / confidence rationale if applicable ___ | No unsupported batch pass rate or guarantee. |
| Approval / report | Planned / actual cartons, cubes and readings ___ | Deviations ___; findings ___; open conditions ___ | Approve / revise / hold ___; reviewer / date ___ |
Count reconciliation: cartons accessed ___; distinct cubes measured ___; axes or defined dimensions per cube ___; repeat readings ___; total readings ___. Keep these quantities in separate report fields.
Decision record: objective supported ___; evidence limitations ___; unresolved measurement issues ___; approved next step ___; reviewer / date ___.
Cube measurement-count questions
Is there one correct number of cubes for every shipment?
No. Define the decision, population, selection design, variation and required evidence. A named formal inspection method needs its own applicable parameters and decision rules.
Do three axis readings count as three sampled cubes?
No. They are readings on one cube. Report distinct cubes, measured dimensions and repeat readings separately.
Does measuring more cubes guarantee every cube meets the specification?
No. A sample supports only the conclusions justified by its selection design, method and decision rules. Do not turn zero observed issues into a guarantee for all unmeasured stock.
What if repeat readings on the same cube disagree?
Review the instrument, positioning, method and defined conditions before interpreting small differences between cubes. More product measurements do not automatically resolve a measurement-process problem.
Method references
Sample count and measurement consistency need separate reasoning.
NIST’s sample-size guidance relates planning to variability, desired precision and confidence under a specified statistical setup. Its measurement guidance distinguishes short-term repeatability from longer-term sources of variation. These general principles inform the questions here; they do not prescribe a charcoal-specific count or acceptance plan.
NIST: selecting sample sizes · NIST: measurement variability
Commercial discussion
Agree what the measurements need to establish.
Share your product requirements, destination market, dimensional specification requested and inspection objective. Request current specifications, identified samples or an agreed inspection scope. Confirm the product, sample availability and measurement arrangements for the inquiry.
Discuss an inspection scopeFactory and quality evidence · Shisha charcoal products · Contact Unimax
Related buyer guides
Evaluate a shisha charcoal sample · Prepare a pre-shipment inspection · Sample across cartons and production lots · Identify samples and batches