Approximation is acceptable when the case needs direction rather than precision, but you should make the level of accuracy explicit. The estimate must still be good enough to support the decision.

A five-percent difference may not matter if one option is twice as large as another. It matters much more when the alternatives are close or the decision depends on a threshold.

Round numbers deliberately and tell the interviewer what you are doing. That makes it possible to judge whether the shortcut is reasonable.

Before approximating, ask what precision the decision requires. Then sanity-check the result so a convenient estimate does not create the wrong order of magnitude. Keep the answer anchored in the actual decision rather than trying to satisfy an imagined recruiter rule. For more detail, see the mental math guide.

Approximation is great when the answer is nowhere near the decision threshold.