Low-Fidelity Prototyping: Mock-ups and Models
Cardboard, foam, clay and paper, sizing for people, and what a mock-up can and cannot tell you.
What a mock-up is for
A low-fidelity prototype is made quickly from cheap, easy materials: cardboard, foam, clay, wood, paper and tape. It answers questions of size, shape, layout, feel and flow, long before anything needs to be accurate or strong. Because it costs minutes, you can make several, break them and learn from each.
Worked example: a cardboard case against a printed one
- A handheld case 150 × 80 × 30 mm has a surface area of 2 × (12 000 + 4 500 + 2 400) = 37 800 mm² = 378 cm².
- In corrugated card at about $0.0002 per cm² it costs $0.08 and 20 minutes to cut and tape. A 3D print of the same shape uses about 55 g of filament ($1.38) and takes about 5 hours of printing: 15 times slower.
- So test fit, grip and layout in card first; print only when the shape is settled (see Rapid Fabrication: Choosing a Method).
Materials and their uses
| Material | Good for | Tools | Notes |
|---|---|---|---|
| Corrugated cardboard | Boxes, enclosures, layouts, brackets at full size | Knife, ruler, hot glue, tape | Cheap and fast; not strong or accurate |
| Foam board (foam core) | Cleaner enclosures, models with edges | Sharp knife, glue gun, pins | Cuts cleanly; dents easily |
| Styrofoam and EVA foam | Organic shapes, handles, grips | Hot wire, rasp, sandpaper | Carve and sand by hand; EVA is soft |
| Modeling clay (oil-based) | Hand grips and ergonomic forms | Fingers, tools | Easy to change; scan or cast it afterwards |
| Wood and MDF scraps | Rigid frames and jigs | Hand tools, saw, drill | Heavier; more work |
| Paper and card (origami, folds) | Folding mechanisms, covers, sheet-metal shapes | Scissors, craft knife | Test folds and bends before cutting metal |
| Building bricks, construction kits | Mechanisms, linkages, gear trains | None | Quick proof of motion |
| Hot glue, tape, 3D pen | Joining and sketching in space | Glue gun, 3D pen | Quick fixes |
| Thin plywood (laser-cut) | Flat parts with accuracy | Laser cutter (see Laser Cutting) | Fast once the file is ready |
Testing with people: size, grip and reach
People vary, and a design must fit a range of them. Body dimensions are roughly normally distributed, so you can describe a dimension by its mean and standard deviation, and design to cover the 5th to the 95th percentile (mean ± 1.645 σ), which includes 90 % of people.
- Clearances (door openings, finger holes, gaps for a hand) are set by the largest person: design for the 95th percentile.
- Reach (distance to a control, a high shelf) is set by the smallest person: design for the 5th percentile.
- Adjustment (a seat, a strap, a handle) covers the whole range; fixed sizes cover only some.
- Use published data (anthropometric tables for your population and age) for real work; the values here are illustrative.
Worked example: how big should a button be?
- Suppose the width of the pad of an adult finger has a mean of 16 mm and a standard deviation of 1.6 mm (illustrative). The 95th percentile is 16 + 1.645 × 1.6 = 18.6 mm.
- A touch target the size of the 95th-percentile finger can be pressed accurately by nearly everyone: about 18 to 19 mm (the often-quoted minimum of 9 to 10 mm is for a fingertip pressing one key; larger for gloves and for stress).
- A hand grip: if grip diameter has a mean of 34 mm and σ 3.5 mm, a handle for 90 % of users should work from 28 mm to 40 mm; a single fixed diameter near the middle (about 34 mm) fits most, and a soft or tapered handle covers the spread.
- Test it: clay and foam handles of three diameters, handed to ten people, tell you more than the numbers.
| Button width | Narrower than this | Wider than this |
|---|---|---|
| 12 mm | 1 % of users have fingers narrower than this | 99 % have wider fingers |
| 14 mm | 11 % of users have fingers narrower than this | 89 % have wider fingers |
| 16 mm | 50 % of users have fingers narrower than this | 50 % have wider fingers |
| 18 mm | 89 % of users have fingers narrower than this | 11 % have wider fingers |
| 20 mm | 99 % of users have fingers narrower than this | 1 % have wider fingers |
Weight and balance
A hollow mock-up weighs a fraction of the real thing, and weight changes how a handheld device feels and where it is held. Add ballast (washers, coins, sand, lead shot) to match the mass, placed where the real parts will be.
Worked example: ballast for a handheld tool
- The real product will weigh 250 g (battery, board, motor, case) and the cardboard mock-up weighs 40 g: the ballast is 210 g, i.e. about 38 coins of 5.67 g, or six 35 g batteries’ worth.
- Put the heavy items where they will really be: a battery of 70 g at the back, 60 mm behind the grip, moves the center of gravity by 70 × 60 ÷ 250 = 16.8 mm. Measure the balance point with a knife edge, adjust the ballast, and check how it feels in use.
Scale models: what scales and what does not
A model at a scale k of the real size has lengths k times as large, areas k², volumes and masses k³. Because strength depends on area and weight on volume, a small model is relatively stronger and stiffer than the real thing, and does not behave like it under load.
Worked example: a 1:2 scale model
| Quantity | Scales as | Half-size model |
|---|---|---|
| Length | k | 0.50× |
| Area (and load capacity) | k² | 0.25× |
| Mass (and weight) | k³ | 0.13× |
| Stress under self-weight | k | 0.50×: half the real stress |
| Natural frequency of a beam | 1 ÷ k | 2.0×: twice as stiff-sounding |
A half-size model of a bridge carries its own weight with half the stress of the full size, so it can survive designs that the real one would not. Use scale models for shape, layout and appearance, not for strength, unless you scale the loads and materials as well (similitude).
Paper prototypes and wizard-of-Oz tests
For a device with a display or an app, draw the screens on paper at the real size and let a person “use” them while you play the computer, swapping the sheet as they tap. A wizard-of-Oz test hides a person behind the interface who gives the responses a program would. Both test the flow and the wording, long before any code exists.
Worked example: a screen drawn to size
- A display of 70 × 40 mm and 128 pixels across has a pixel pitch of 0.55 mm. A 9 mm touch target is about 16 pixels: draw the buttons that large and see how many fit (about 7 across).
- Print the frame at 1:1 on card, cut the window, and slide the paper screens behind it: the layout is tested in minutes.
A one-afternoon mock-up session
| Step | Minutes |
|---|---|
| Sketch and cut 3 variants | 45 |
| Assemble with tape and glue | 30 |
| Test with 3 people | 45 |
| Record and decide | 15 |
Total 135 minutes (about 2.3 hours) for three variants tested with three people: nine observations and a decision. Compare with a day for one printed model.
What a mock-up cannot tell you
- Strength, fatigue and wear: test with real materials (see Testing, Iteration and Feedback).
- Heat, noise, vibration, battery life: these need working parts (see Electronics in a Prototype).
- Tolerances and assembly: a mock-up cut by hand has errors of a millimeter or more; tolerance questions need a precise process (see Mechanisms, Fasteners and Snap-Fits).
- Manufacturability: what is easy in card can be impossible to mold (see From Prototype to Production).
- Real use: people behave differently with a toy than with a finished product: confirm with a later prototype.
Troubleshooting
| Problem | Likely cause | Try |
|---|---|---|
| Users are polite, not honest | Mock-up looks too finished, or the test is leading | Make it rougher; ask them to perform tasks; watch, do not ask |
| Mock-up falls apart in testing | Too weak for handling | Reinforce with tape and card ribs; handle it yourself |
| Does not feel like the real thing | Wrong weight or balance | Add ballast in the right places |
| Size right in the mock-up, wrong in the print | Inaccurate cutting or forgotten parts | Measure the mock-up; model the real parts at full size |
| Cannot decide between variants | No criteria | Score them (see Rapid Fabrication: Choosing a Method) or test with more people |
Related
- Prototyping Techniques: the overview of this guide.
- The Prototyping Process and Its Stages: the ladder of prototypes.
- CAD, Simulation and Digital Prototypes: the digital version.
- Laser Cutting and 3D Printing: the next step up.
FAQ
What is the best material for a quick mock-up?
Corrugated cardboard and hot glue for boxes and layouts; foam board for cleaner enclosures; clay or foam for grips and organic shapes.
How accurate does a mock-up need to be?
Only as accurate as the question: ±1 mm is fine for whether a hand fits a handle, not for whether two parts mate.
How do I test ergonomics?
Make clay or foam versions in several sizes and hand them to people across the range; use published percentile data for design limits.
Can I use a scale model to test strength?
Not directly: strength and weight scale differently. A scale model is stronger relative to its weight than the real thing.
How many people should test a mock-up?
Three to five people per round usually shows the main problems; run several rounds (see Testing, Iteration and Feedback).
References
- Pheasant, S. and Haslegrave, C. M., Bodyspace: Anthropometry, Ergonomics and the Design of Work, CRC Press.
- Snyder, C., Paper Prototyping, Morgan Kaufmann; Ulrich, K. T. and Eppinger, S. D., Product Design and Development.
- Published anthropometric databases (for example ANSUR II, DINED) for the population you design for.
Last reviewed: October 2026. Body dimensions and costs here are illustrative: use published data for design.