/ Library/ What Should We Remember/ Duration Neglect
Recall Bias № 067 · Last updated 6 June 2026

Duration Neglect.

"How long it took matters less than how it felt at the worst and last moments."

01Overview

Duration neglect is the finding that people poorly integrate the total length of an experience into their overall evaluation. A long wait with a smooth ending can feel better than a short wait that ends badly — and vice versa. Duration is underweighted relative to intensity and closure.

Designers obsess over total time on task, average handle time, and load speed metrics. Users remember whether the wait felt anxious, whether progress was visible, and how it ended. Optimising average duration while ignoring peak frustration and final impression is duration neglect in product form.

02Detailed explanation

Kahneman's cold-hand studies and service queue research show:

  • Adding a slightly longer ending that reduces peak pain improves remembered experience — even when total time increases.
  • Flat, unexplained waits feel longer in memory than variable waits with explained milestones — regardless of identical duration.
  • Users report "quick" experiences that took longer when peaks were mild and endings clear.

Duration neglect is the companion to peak–end rule: together they argue for designing the emotional timeline, not only the clock timeline.

03Why it exists

Memory compresses duration unless it is salient. Intensity and narrative closure are easier to encode than elapsed seconds.

Analytics default to averages. Peaks and endings require intentional instrumentation and qualitative follow-up.

The short version

Shaving five seconds off average load time may matter less than removing the terrifying spinner at second twelve.

04Effects on users

Users tolerate longer flows that communicate progress and finish confidently. They resent shorter flows that stall without explanation or end ambiguously.

Their retrospective complaints target "how awful it felt" more often than exact minutes — support logs reflect intensity, not duration logs.

05Effects on designers & teams

Teams optimise the wrong variable:

  • Speed-only KPIs. Faster backend, same anxiety UI — remembered as slow.
  • Ignored tail experiences. Checkout optimised; confirmation and email receipt neglected.
  • Queue design blind spots. Hold music and dead air treated as acceptable if average wait drops slightly.

06Practical takeaways

  • Map peaks and endings explicitly. Journey maps should tag emotional highs/lows, not only step count.
  • Invest in progress communication. Skeleton screens, staged disclosure, and honest ETAs reduce peak pain.
  • Design endings deliberately. Success states, receipts, and next-step clarity are memory-forming.
  • Consider benevolent duration. Sometimes adding a calming final step improves memory without harming outcomes.
  • Measure remembered experience. Post-task surveys about how long it felt — compared to actual timers.

07Design examples

Loading

Fast but frightening

API responses improve 30%. Spinner remains a blank screen with no progress. Satisfaction flatlines because peak uncertainty unchanged.

Support queue

Shorter hold, harsher end

Average wait drops two minutes. Calls now end with abrupt automated survey. NPS falls — ending got worse while duration improved.

Onboarding

Quick but cliff-ended

Users complete setup in three minutes — record time. No confirmation of what happens next. Day-two recall: "I guess it worked?"

Healthcare portal

Longer walk-through wins

Guided upload with milestones takes longer than instant upload. Fewer abandoned uploads and higher remembered ease — duration neglected in favour of managed peaks.

08Ethical risks

Exploiting duration neglect — making endings feel good while hiding total cost, fees, or commitment time — is a dark pattern cousin of misdirection.

Users making time-sensitive decisions (health, finance, legal) deserve honest duration signals, not only emotionally polished endings.

Self-test: Where have you optimised average time while leaving the worst moment untouched?

10Suggested reading