Hedonic adaptation
Emotional reactions to a changed situation fade as it becomes the new normal.
After a change in circumstances, feeling tends to drift back toward its prior baseline. The drift is faster and more complete for some changes than others, and adaptation to sustained bad conditions is often partial. Attention is a driver: people stop noticing what stays constant.
The spike and the drift back
How it shows up in software
A speed improvement, a redesign or a new feature produces a spike in sentiment that decays into expectation. The team that shipped it still feels the win; users have rebaselined. This is why last quarter's headline feature stops appearing in interviews, and why a product can improve steadily while satisfaction stays flat.
Using it well
- Expect post-launch sentiment to decay and set the measurement window accordingly. Read satisfaction at four to eight weeks, not at day three.
- Vary the surface rather than repeating it. A weekly digest that changes its lead section holds attention longer than the same template.
- Spend on removing recurring irritations rather than adding another one-off flourish, since people adapt to a pleasant surprise faster than they adapt to a daily annoyance.
- Make improvement visible at the moment it is used, for example showing the saved step, instead of announcing it once in a changelog.
Where it turns manipulative
- Escalating stimulation to outrun adaptation, which is the mechanism behind streak pressure, variable reward: An unpredictable payoff, which sustains a behaviour longer than a predictable one. loops and notification inflation.
- Treating adaptation as licence to degrade quality slowly, on the theory that users will rebaseline to the worse version.
- Withholding a fix so it can be released later as a delight moment.
Where you have seen it
Spotify
Discover Weekly refreshes its contents on a fixed cadence, so the surface stays the same while the payload changes.
Duolingo
Rotates visual themes, character animations and lesson framings over time rather than shipping one fixed reward screen.
What the research says
- Brickman, Coates and Janoff-Bulman, 1978Contested
Compared lottery winners, people with paraplegia and controls on rated happiness. Winners were not reliably happier than controls, and the accident group was less far below controls than expected.
Very small samples, cross-sectional, and the famous summary overstates it. The accident group was measurably less happy than controls. Treat it as the origin of the idea, not proof of it.
- Lucas, Clark, Georgellis and Diener, 2003Well evidenced
Followed a large German panel over 15 years. People adapted substantially to marriage but did not fully return to baseline after unemployment.
This is the finding that corrects the simple treadmill story: adaptation is real, incomplete, and varies by event type.
- Frederick and Loewenstein, 1999Mixed evidence
A review across noise, imprisonment, income and disability, cataloguing where adaptation is strong and where it stalls.
Grades are a judgement about the evidence, not about how useful the idea is. Plenty of contested effects are still worth knowing, as long as you do not cite them as settled.