You can download a layout that packs your parts beautifully, take it to the shop, make two cuts, and discover there is no third cut you can make. The plan was never cuttable. This happens more than it should, and the reason is a constraint that optimizers sometimes quietly ignore.
What your saw can actually do
Put a sheet on a table saw, a track saw, or an industrial panel saw. Every one of those tools does the same thing: it makes a straight cut, all the way across, from one edge to the other. The blade enters at an edge and leaves at the opposite edge. It cannot start in the middle. It cannot stop halfway and turn.
So a cut always splits one piece into exactly two pieces. Then you take one of those and split it into two. And so on. This is called a guillotine cut, and it is the defining constraint of sheet-goods cutting.
The layouts that break it
Now picture four rectangles arranged in a pinwheel: one across the top-left, one down the right side, one across the bottom-right, one up the left. They interlock like a windmill. It is a genuinely efficient packing — very little wasted space.
Try to cut it. Any line you draw all the way across the sheet passes through at least one part. There is no first cut. The layout is perfectly valid as a packing and completely impossible as a cutting plan.
A layout is guillotine-cuttable if you can repeatedly split every piece edge to edge until only parts remain. If at any stage no full-width or full-height line misses every part, the plan cannot be made on a saw.
Why bad plans get produced
Rectangle packing is a well-studied problem, and most off-the-shelf algorithms solve the unconstrained version — get the most rectangles into the area. That is the right problem for a CNC router, a laser, a waterjet or a plasma table, where the head can plunge anywhere and follow any path.
It is the wrong problem for a saw. An optimizer that reports a higher yield by using non-guillotine layouts is not better; it is answering a question you did not ask. The yield figure is real, and the plan is useless.
How to spot one
Look at the drawing and try to find the first cut. Trace a line all the way across the sheet, top to bottom or left to right, that does not touch any part. If you cannot find one, put the plan down.
Then repeat the test on each piece the first cut creates. A cuttable plan survives this all the way to individual parts. It is a two-minute check and it will save you a sheet.
What you give up, and what you get
Being honest: the guillotine constraint does cost yield. A pinwheel layout genuinely fits more parts than any guillotine layout of the same set. Typically you give up a few percentage points.
In exchange you get a plan you can execute, in a defined order, on the tools you own. A layout you cannot cut has a yield of zero.
Where the exception applies
If you have a CNC router, a laser or a waterjet, none of this binds you. The head goes where it likes, parts can nest inside each other's curves, and you should use software built for true-shape nesting — you will get much better material use than any saw-based tool can offer.
For everyone working with a blade, the constraint is real and permanent. Kesim only ever produces guillotine-cuttable layouts, and every plan comes with the cut sequence in order — rip the sheet into strips, then crosscut each strip. Every cut runs edge to edge, because that is the only kind your saw can make.