How old can a candidate's example be before a behavioral interviewer starts discounting it?

## Why age matters at all A behavioral interviewer is not collecting biography. They are trying to predict how you will behave in the role starting shortly after the loop ends, and the closer an example sits to the present, the less inference they have to do. An example from last quarter is close to a direct observation. An example from many years ago is a claim about a person who has since changed jobs, levels, tools and habits, and the listener must silently discount it by an amount they will not tell you. ## The practical floor Aim for roughly the last two to three years for your main examples. That window is long enough to contain a real project cycle — a migration, an on-call rotation you changed, a team you led — and short enough that the person you are describing is recognisably the person in the room. The floor is not a rule the interviewer recites; it is the point at which they start asking follow-ups like 'is that still how you'd approach it?' or 'what have you done more recently along those lines?' Those questions are the discount becoming audible. If you hear one, the example was too old for the axis it was meant to evidence. ## When reaching back is correct Three situations justify an older example. **A rare signal with no recent instance.** Some axes only get exercised when circumstances cooperate. If the last time you led a group through a genuine crisis was several years ago, a recent, milder substitute may not carry the signal at all. A real older example beats a thin recent one. **A scope you no longer operate at.** Roles move sideways as well as up. Someone who ran a large infrastructure programme and then joined a smaller team may find that the recent work simply does not reach the stakes the prompt is asking about. **A genuinely formative correction.** Failure prompts tolerate more age than most, because the value is in the correction loop, and a correction that has held for years is stronger evidence than one made last month whose durability is unknown. ## How to reach back without paying for it Name the age, do not hide it, and close the gap yourself. 1. **Flag it in one clause.** 'The clearest example is a few years old — I'll say why it's still the right one.' Volunteering the age reads as calibration; being caught by a follow-up reads as padding. 2. **Land the answer in the present.** End with what you would do differently now, or where the same judgment has shown up since in smaller form. This is the move that converts age into a growth signal rather than a staleness signal. 3. **Strip the perishable detail.** The parts of an old story that date fastest are the surrounding conditions — how deploys worked then, what tooling existed, what the team structure was. Keep the decision and the reasoning, compress the machinery to a sentence. 4. **Do not stack them.** One older example inside a round is a considered choice. Three is a pattern, and the pattern the listener perceives is that nothing notable has happened to you lately. ## The failure this interacts with Age and drama fail together more often than either fails alone. Old stories are usually the ones that got polished by retelling, so they are smoother and more exciting than anything recent — which is exactly why the reflex reaches for them. A candidate who tells a vivid outage from long ago instead of a quiet, on-signal decision from last quarter loses twice: the discount for age, and the missing evidence for the axis. Before reaching back, check that the older example wins on signal and not merely on polish. ## A short calibration For each of the five themes that recur, ask whether you have an example inside the two-to-three-year window that would satisfy the axis. Wherever the answer is no, you have found either a gap worth filling with current work, or a place where an older example must be flagged and landed in the present. Doing that check before the loop is far cheaper than discovering the gap while a behavioral interviewer waits.