Picking a material is really about trade-offs: strength, stiffness, weight, cost, corrosion, temperature and manufacturability all pull in different directions. A structured pass beats reaching for "aluminium because it's light" every time.
Start with the function, not the material
Ask what the part actually has to do. Resist bending without flexing? You want stiffness (Young's modulus E), and geometry matters more than material. Carry load without breaking? You want strength (yield). Survive millions of cycles? Fatigue strength. Not corrode? That may override everything.
Strength-to-weight and stiffness-to-weight
For anything that moves or flies, what matters isn't absolute strength but strength per unit weight (specific strength) and stiffness per unit weight. This is why titanium and carbon-fibre win in aerospace despite the cost, and why a material-property chart plotted as strength vs density is the classic selection tool — you want the upper-left corner.
You need a beam that is lighter but just as stiff in bending, and you are free to change its depth. Which material index should you maximise?
The index depends on the shape, because the shape is what you are allowed to change. A tie can only get fatter in proportion, so it wants E/ρ. A beam can get deeper, and depth enters bending stiffness cubed, so a softer material can buy back its modulus with section — that gives E1/2/ρ. A plate, constrained in width, gives E1/3/ρ. It is why wood and composites beat steel for beams and lose to it for ties, and why "use a stiffer material" is the wrong instinct on its own.
Try it — the selection line
Then layer in the constraints
Cost (raw + processing) usually decides between candidates that both work. Corrosion: stainless, coatings, or design out galvanic couples (dissimilar metals). Temperature: modulus and strength fall with heat; plastics soften early. Manufacturability: can you actually cast/machine/weld it? Compare candidates on real numbers with the material property tool and weight calculator.
A workable method
List the must-haves (hard constraints), rank the nice-to-haves, shortlist 3–4 materials that pass the constraints, then compare them on the ranked criteria with real values. It's faster than agonising and it leaves a paper trail for the design review.