There is a category of design that never appears on a screen except as a preview: the things that get printed, cut, folded and handed to somebody. A gift box, a paper crane, a bunting banner, a pop-up card. They look simple, and they fail in ways that flat design never does, because they have a second geometry underneath the visible one.
Get that geometry wrong and the artwork can be beautiful and the object will still not close.
The dieline is the real drawing
Anything assembled from flat material starts as a dieline: the outline of the shape before folding, with every cut, crease and glue tab marked. It is a technical drawing, and it is what the object actually is. The artwork is decoration applied afterwards.
By convention the lines are color coded: solid for cut, dashed for fold, and a further distinction between a mountain fold, which folds away from you, and a valley fold, which folds towards you. Anyone who has assembled flat pack furniture from a diagram with the two confused knows exactly how much that convention matters.
Three things belong on any dieline and are frequently missing:
- Glue tabs, with enough surface to hold and a bevelled edge so they slide into place instead of catching.
- Material thickness allowance. Card has depth. A box designed to exact internal dimensions will not close, because each fold consumes a little material and the walls take up space. Thicker board needs more.
- Grain direction. Paper and card have a fibre direction. Fold along the grain and you get a clean crease; fold across it and the outer surface cracks, which is instantly visible on any dark or heavily printed area.
The artwork has to know where it will end up
This is where most homemade paper craft goes wrong. A pattern that looks lovely on the flat sheet ends up with the interesting part on the underside of a flap, and the plain area facing out.
Before decorating anything, you need to know which region of the flat sheet becomes which part of the finished object. On a box that is reasonably intuitive. On anything folded repeatedly it is not intuitive at all: in a paper crane, the part that becomes the head starts as a corner of the square and travels a long way, turning over more than once on the journey.
This is the single strongest argument for a preview that actually folds. A printed diagram shows you a sequence of flat drawings; it cannot show you which piece of your photograph ends up on the wing. Watching the fold happen, and being able to stop halfway and turn the half-folded shape around, answers in seconds a question that is otherwise answered by printing, folding and starting again.
Two sides, and the mirror problem
Folded paper shows both of its faces. That is what gives a crane its two-tone wings and a box a contrasting lining, and it is why designing only the front is designing half the object.
Printing both sides introduces a problem that catches everyone once. The back of the sheet must be mirrored, because it is printed on the reverse: what is on the left of the front is on the right of the back. Print the back the same way round as the front and every element lands in the wrong place, and the error is invisible until the sheet is folded.
Two further practicalities. Duplex printing has registration tolerance, usually a millimetre or two, so any design that depends on front and back aligning exactly will disappoint; give yourself margin. And thin paper shows through, which turns a strong dark design on the back into a gray ghost on the front. For two-sided work, heavier stock is not a luxury.
Cut lines, and why they are never exactly where you drew them
Whether the cutting is done by a machine, a blade or a pair of scissors, it lands close to the line rather than on it. The response is the same as in commercial print: extend color past the edge, and keep anything important away from it.
For home cutting, bleed of 2 to 3 millimetres past the cut line is sensible, and 3 to 5 millimetres of safe margin inside it. Cutting machines that read registration marks are more accurate, and still not perfect, particularly on thicker material where the blade deflects slightly.
One tip that saves a great deal of frustration: score before you fold, always. A crease made by folding alone is uneven, and on card it cracks. Scoring with a blunt point along a ruler produces a fold that is straight, sharp and repeatable, and takes about ten seconds.
Instructions are part of the design
If somebody other than you will assemble the thing, the instructions matter as much as the artwork, and they are usually an afterthought.
What works: one drawing per step, showing the state before the action, with an arrow indicating the movement. Fold lines drawn in the same convention throughout. A number, not a paragraph. Where a step is genuinely difficult, an additional view from a different angle.
What does not work: a photograph of the finished object with a caption saying “fold as shown”.
Choosing the material before the design
Weight is measured in grams per square metre and it changes what is possible more than any other single choice.
- 80 to 100 is ordinary printer paper. Fine for origami and anything folded many times, useless for anything that must stand up.
- 120 to 160 is the useful middle: holds a crease, takes ink well, still folds without cracking.
- 200 to 300 is card. Needs scoring before every fold, and will crack across the grain if you do not.
- Above 300 starts to need proper creasing tools and generous allowances for thickness.
Origami is the exception that proves the rule: a model with twenty folds needs thin paper, because twenty layers of anything heavier will not close. Traditional origami paper is around 60 grams for exactly this reason.
Prototype in the worst material you have
The habit that saves the most time: before printing anything on good stock, make the whole thing once in plain paper with no artwork on it at all.
You are not testing the design, you are testing the geometry. Does it close, do the tabs reach, is the thickness allowance right, does the fold sequence work in the order you wrote it. Five minutes with a scrap sheet catches nearly every structural error, and structural errors are the ones that cost a whole print run.
Then print one good copy, assemble it, and only then make forty.
