Build, Test, Evaluate

Now build it — and find out

Steps 5 and 6 are where the design meets the world. You implement the solution by building it, then evaluate it by testing it against the criteria from your brief. Most of what you learn about a design, you learn here, which is why good designers build early and test often instead of polishing an idea on paper for weeks.

The five questions

As the build takes shape, designers keep asking five questions. Each one is a different way a good idea can still be a bad product.

Ergonomics is matching a design to the human using it: a handle that fits a hand, a button where a thumb lands, a desk at the right height. A bottle stand that holds the bottle perfectly but needs two hands to get the bottle out has failed ergonomics.

A fair test

A test only teaches you something if it is fair: you change one thing, keep everything else the same, and measure the result. Testing a paper bridge? Keep the paper, width, and span the same and change only the number of folds; measure how many pennies it holds. A variable is a thing that can change. Change two at once and you cannot tell which one caused the result.

Record every trial in a table — the variable, the result, and what you noticed. Data you wrote down can be compared and graphed. Data you remember cannot.

Failure is information

A test that fails is not a failed project. It is the loop doing its job: the result points at the next change. The tower that fell over tells you the base is too narrow. The stand that tipped tells you where the weight sits. Write down how it failed, make one change, test again. That is redesign, and the third version is usually the one that works.

Evaluation does not stop at the test

Real products keep being evaluated after they are sold. Companies pull items off the production line to test, survey the people who bought them, and read complaints and warranty claims. Software is tried by early users before it is released. Every one of those is step 6, still running.

Where Craft Lab fits

Craft Lab is a place to run cheap trials: lay out the sticks, snap the connectors, move things, and see what fits before you cut real material. Build it digitally, build it for real, test it, and bring what you learned back to the screen. Materials are limited. Tries on a screen are not.

Done is not "I built it." Done is "it meets every criterion, inside every constraint, and I have the test data to show it."

A team wants to know whether more folds make a paper bridge stronger. Design their test, then handle the result.

1. 1

   The team wants to test the effect of folds. Which ONE variable should change between trials?

   💡 Only the thing you are asking about changes. Everything else stays fixed.

   Choose one, then Check.

   ○Number of folds○Number of folds and the type of paper○Width of the strip○How fast the pennies are added

   Check
2. 2

   Sort each part of the test: keep it the same, change it, or measure it.
3. 3

   Tap the test procedure in order.
4. 4

   Every strip held fewer pennies than the brief requires. Choose every move that fits the design process.

✅ Check: Build, Test, Evaluate