A retail chain had approved a new fixture design and needed every branch to look identical. The constraint was not manufacturing — it was holding colour and fit across seven production batches shipped to different regions.
The design was already approved when we were brought in. Six fixture types, fourteen pieces per store, in a tinted acrylic that matched the brand colour. The rollout covered 120 locations in four shipping waves over eight weeks.
The failure mode everyone was worried about was colour. Tinted acrylic varies between sheet lots, and a customer walking from one branch to another in the same city will notice a difference in brand colour faster than they will notice a badly made joint.
The second risk was assembly. Store fitters are not joiners. A fixture that ships as fourteen parts with a drawing and no parts list gets assembled wrong, and the error is only discovered when a customer photographs it.
Two decisions were made before any material was ordered.
A retained master swatch. We produced physical swatches from three candidate sheet lots, the client approved one, and that swatch stayed in our inspection area for the life of the project. Every subsequent batch was measured against it under D65, targeting ΔE ≤ 2. Crucially, batches were not compared to each other — comparing to the previous shipment lets drift accumulate, whereas comparing to a fixed master resets the error every time.
Flat-pack with numbered hardware. Each store received one carton per fixture type, with hardware in a numbered bag and a printed parts list showing which bag goes with which step. Multi-carton fixtures were numbered in assembly order. This cost marginally more in packing time and eliminated the assembly-error category entirely.
The eight-week window ruled out anything that needed a development round, so the programme ran on one construction repeated across four fixture types: a countertop display stand, a tiered shelf unit, a lockable showcase and a floor POP display cabinet. Every one of them shares a 5 mm cast sheet wall, a 0.5 mm squareness callout and the same jig family, which is what allowed a single first article to release the whole programme.
The first physical object made was not a fixture, it was the bonding jig. Building the jig first means the first article is checked against the geometry that will actually produce 120 sets, rather than against a hand-built one-off that later proves impossible to repeat. The jig holds squareness to 0.5 mm across a metre, which at 1.8 m tall translates to under 1 mm of rack — invisible on a shop floor and enough to make doors hang true.
| Stage | What was fixed | Why it mattered across 120 stores |
|---|---|---|
| Sheet | One batch, held for the whole programme | Removes colour drift between store 1 and store 120 |
| Jig | Bonding jig built before the first article | The article is approved against repeatable geometry |
| Edge | Flame polish on visible edges, sawn elsewhere | Consistent look at a third of diamond-polish labour |
| Hardware | One cam lock, one hinge, one part number | A store team never has to identify a variant |
| Packing | Per-store carton, numbered hardware bag | A short shipment is visible on receipt, not on install day |
The one thing that could not be jigged was the store itself. Floors are not level and walls are not square, so every floor cabinet shipped with a 15 mm scribe allowance on the two faces that meet a wall, and a pack of shims. That allowance costs almost nothing at the cutting stage and it is the difference between an installer finishing in an hour and an installer cutting a panel on site.
Two things did not go to plan, and both are worth recording.
Batch four came in at ΔE 2.4. The sheet lot was within the supplier's own tolerance but drifted from our master. We rejected the sheets before cutting, which cost four days, and the client was notified the same day. The alternative — shipping slightly-off fixtures to 30 stores — would have been invisible on delivery and obvious in a year.
The first wave arrived with two cracked corners. Flat-pack cartons had insufficient corner protection for the courier used in that region. We added moulded corner inserts from wave two onward and back-shipped replacements for the eight damaged pieces. The fix was in packaging, not in the part.
120 stores fitted across four waves, seven production batches, no fit rework and no colour complaints. The retained swatch and the approved sample stayed with us, which means a repeat order in eighteen months can be matched to the original standard rather than to whatever the last batch happened to be.
The cost of the two interventions — retaining a master swatch and numbering hardware bags — was small relative to the project. Both are now standard on any rollout work we quote.
Twelve weeks from drawing to last carton is a realistic number for a 120-store rollout of four fixture types, and it breaks down in a way that is worth planning against.
| Phase | Duration | What can go wrong |
|---|---|---|
| Sampling and approval | 3–7 business days per round | Two extra rounds cost two weeks and cannot be recovered later |
| First article | 5–8 business days | A colour approved on a screen rather than a swatch |
| Production | 7–15 business days per batch | Sheet arriving late; hardware on back order |
| Packing and labelling | 3–5 business days | Store list changing during packing |
| Freight and install | Depends on geography | Site access; staged delivery by floor |
On budget, the programme is dominated by part count rather than by material. Four fixture types at 120 stores each is 480 units, and the difference between a four-part and an eleven-part design across that volume is measured in tens of thousands, not hundreds. The other large line is packing: per-store cartons with numbered hardware add 12–20 percent over a bulk pack, and on a rollout that is money well spent.
Where a chain is planning its first rollout, the single most useful thing to fix early is the colour standard. Everything else can be re-cut; a colour that was never signed off cannot.
More projects
Send the fixture drawings and the store count. The first thing we will want to settle is the colour swatch, because everything downstream depends on it.
