What a small sample cannot tell you
A small swatch under showroom light is useful for shortlisting, but does not approve a whole elevation:
- How it reads across an elevation. Colour and texture aggregate. Review variation across the proposed area rather than extrapolating from one piece.
- The joint pattern. Invisible on a single piece, and often the most visually dominant thing about the finished surface.
- Module repetition. Whether a pattern reads as texture or as obvious repetition only shows at scale.
- How it behaves in the real light. Orientation, time of day and light colour temperature all change the reading.
- Batch variation. One piece does not establish the spread.
Takeaway: Match the sample level to the decision being approved.
Three levels, for three kinds of decision
Hand sample. Confirms material identity, rough colour, surface feel. Useful for shortlisting and potentially for a small local interior detail, but not for approving the appearance of a whole elevation.
Large sample or sample board. Big enough to show a full module and at least one joint. Viewed outdoors for external materials, at the real distance, ideally on the actual orientation. Use it to review the repeated surface; a small isolated detail may need only a hand sample.
Mockup or benchmark panel. A built section of the real assembly — substrate, support system, fixings, cladding, jointing, sealant — erected on or adjacent to the building. Tests appearance, buildability, tolerance and workmanship together, and becomes the agreed standard.
Viewing and approval record
Keep this record with the retained sample. These are editorial review fields, not a certification procedure.
| Field | Approval record |
|---|---|
| Product / batch / sample ID | ___ |
| Accepted range limits | Lightest / darkest; texture or pattern limits: ___ |
| Approved scope | Local detail / repeated field / named junction: ___ |
| Light / orientation / distance | ___ |
| Dry and wet state | Observed / not applicable / not checked: ___ |
| Photos | IDs / dates / matching viewing conditions: ___ |
| Decision and signature | Accept / reject / hold; person / role / date: ___ |
| Retained version | Sample location / drawing revision / later changes: ___ |
| Not established | Untested performance / unresolved details: ___ |
Takeaway: Record both what was accepted and what the sample did not establish.
Natural materials: approve a range
Stone, timber and brick vary within a source and between batches. A single approved sample is not a usable standard, because the delivered material will not match it exactly and there is then no agreed basis for accepting or rejecting.
Instead, approve a range: samples showing the acceptable extremes of colour, veining, texture and figure. Write the range into the specification. That gives both sides something to measure against.
Takeaway: Keep the accepted range samples, rather than relying on a single piece.
See natural vs engineered building materials for why this problem is structural to the material family rather than a supplier failing.
When a mockup is worth building
A mockup costs real money, so it earns its place when:
- The area is large enough that a mistake is expensive to remedy
- The assembly is unusual, or combines materials in a way the trades have not built before
- Appearance is contentious or hard to predict from samples
- Tolerances are tight and need to be proven achievable
- Multiple trades meet at a junction that has to be sequenced correctly
Decide from junction complexity, uncertainty and the cost of rework, not project size alone. A small residential project can still have a junction worth mocking up.
Takeaway: Decide from junction uncertainty and rework risk, not floor area alone.
What a mockup actually tests
- Appearance at scale, in real light, with real joints
- Buildability — whether the detail can be built as drawn
- Tolerance — whether the achievable accuracy matches what the design assumed
- Sequencing — which trade goes first, and what that means for access
- Workmanship standard — the reference all subsequent work is judged against
That last one is the reason mockups survive value engineering. It converts “good quality” from an adjective into a retained reference both parties have agreed on.
Takeaway: Record which appearance and workmanship decisions the mockup establishes.
Where this sits
Sample and mockup work belongs in design development and early technical design, alongside the material decision itself — see how architects select building materials and how to compare material durability and maintenance. The building materials and finishes guide covers the wider process.
Frequently asked questions
How big should a sample be?
Big enough to show at least one full module and a joint, viewed at the distance the finished surface will be seen from. For a facade that usually means panels rather than swatches; for a small interior detail a hand sample may be enough.
What is a benchmark or mockup panel?
A built section of the real assembly — substrate, fixings, cladding, jointing and sealant — erected on or near site. It sets the agreed standard for appearance and workmanship, and it is what later work is judged against.
Why do samples look different on site?
Light. Interior lighting has a different colour temperature and direction from daylight, and daylight itself changes through the day and with orientation. A material approved under showroom light regularly reads differently on a north elevation.
How do I approve a natural material that varies?
Approve a range rather than a piece — a set of samples showing the acceptable extremes of colour, veining and texture. Written into the specification, that range becomes the standard the delivered material has to sit within.
Sources
- Building Envelope Design Guide — Whole Building Design Guide, National Institute of Building Sciences Accessed August 27, 2026.
