What does it mean that the recency bias?

Recent events disproportionately influence your perception of what is normal or likely.

Example: Your company had two bad quarters after twelve good ones. In the leadership meeting, the team debates whether to abandon the current strategy — as though fourteen quarters of evidence no longer matter because the last sixty days felt painful. Nobody references the full dataset. The most recent experience has silently overwritten their calibration of what is normal.

Try this: Pick a domain where you recently changed your mind or shifted your behavior — investment allocation, a judgment about a colleague, a habit you dropped. Write down the event that triggered the shift. Now write down the full history: the last 12 months, 3 years, or whatever the relevant window is. Ask yourself: does the recent event justify the weight I gave it, or did I let the last data point rewrite my entire model?