What it is
Recency bias is the tendency to weight recent information more heavily than older information, even when the longer-run record is more statistically reliable. It is reinforced by the emotional salience of recent events and by the way most reporting surfaces the latest data point rather than the trend. In fast-moving environments, recency bias can masquerade as responsiveness - but responsiveness grounded in noise rather than signal is just instability with a better reputation.
Where it shows up
In product and technology strategy, recency bias appears when a recent quarter of weak sales, a single poor sprint, or a newly published competitor feature triggers a fundamental change of direction before the underlying trend has been established. Build-vs-buy decisions made in the wake of a specific event - a public outage from the current vendor, a high-profile competitor build - often reflect recency bias rather than a systematic assessment of long-run fit and cost.
What Rubicon Probity does
When Rubicon Probity detects a decision record at the Decide stage where the stated trigger is a recent event and the supporting evidence is concentrated in the last quarter or less, it raises a NOTE flag and requests a longer-horizon data view - at minimum 12 months of comparable data - before the decision is finalised. The record must show that the recent signal is representative rather than anomalous.
Detection questions
- How does the evidence from the last quarter compare to the 12- or 24-month trend - and does the trend support the same conclusion?
- If this triggering event had not occurred in the last 90 days, would you still be making this decision now?
- Have you confirmed that the recent data point is a signal rather than noise before using it to justify a strategic change?