The Peak and the End

People do not remember experiences as averages. They remember the most intense moment and the last one, and they will happily choose more total suffering if it finishes better.

Here is an experiment that should not work the way it does. Participants are asked to hold a hand in painfully cold water for sixty seconds. Later they do a second trial: sixty seconds at the same painful temperature, followed by a further thirty seconds during which the water is warmed very slightly, still unpleasant but noticeably less so. Then they are told they must repeat one of the two trials and asked to choose.

A clear majority choose the longer one. They elect to endure ninety seconds of cold rather than sixty, which is objectively more total discomfort, because the extra thirty seconds improved the ending. Daniel Kahneman and colleagues ran versions of this repeatedly, and the result held. It is one of the cleanest demonstrations in psychology that remembered experience and lived experience are different quantities, and that we make our decisions using the wrong one.

The rule extracted from this body of work is called peak and end. When people evaluate an episode afterwards, their judgement is dominated by two data points: the moment of greatest intensity, and the final moment. Almost everything in between is discarded, and so, remarkably, is duration. Researchers named that last part duration neglect, and it means an experience twice as long is not remembered as twice as much of anything.

The finding moved out of the laboratory in a study that people still find slightly shocking. Donald Redelmeier and Kahneman worked with patients undergoing colonoscopies at a time when the procedure was performed without routine sedation, having them report discomfort in real time. Some patients had the scope left in place, stationary, for a short additional period at the end, adding time and mild discomfort but avoiding the sharp finish. Those patients rated the whole experience as less unpleasant afterwards, and follow up suggested they were more likely to return for the procedure again. Adding pain improved the memory, and the memory is what determines whether a person comes back.

Kahneman built a larger idea on this, dividing us into an experiencing self and a remembering self. The experiencing self lives in the present and has, on his estimate, an attention span measured in seconds. The remembering self writes the account afterwards, keeps the records and, crucially, makes all future decisions. The two disagree constantly. We do not choose between experiences. We choose between memories of experiences, and the memory was assembled by an editor with a strong preference for climaxes and endings.

Once you see this you cannot stop seeing it in badly designed encounters. A superb three hour meal ending with fifteen minutes of trying to get the bill is remembered as an evening with a problem in it. A hotel stay of impeccable quality ending at a checkout desk staffed by someone unhappy to be there loses more than a fair accounting would allow. A flight is judged largely by the boarding, the worst turbulence and the wait at the carousel, and almost not at all by the seven uneventful hours in between.

The commercial version of the insight is well known and frequently abused. Service recovery works partly because a problem resolved warmly late in an interaction can become the peak, and a well made ending is cheap relative to the difference it makes. That is legitimate design. It becomes manipulation when a business deliberately builds an unpleasant middle in the knowledge that a generous finish will be all anyone remembers, which is the logic of every airline that treats you badly for six hours and hands you a chocolate on descent.

For fragrance, the peak and end rule is not an analogy. It is a literal description of the product. A composition has an opening, a heart at its most intense, and a drydown, and it is unfolding in time exactly like the experiences in these studies. Which means the average pleasantness of a fragrance across eight hours is very nearly irrelevant to how it will be remembered. What will be remembered is its most intense moment and how it left.

This is why the drydown deserves more attention than the industry gives it, and why the opening gets far more than it deserves. Openings sell bottles, because openings are what a person smells on a paper strip in a shop, and the entire retail apparatus is built around that thirty second window. But the opening is not the peak and it is certainly not the end. The person who wears the thing will spend six hours with the base, and that is the part their remembering self will file. A fragrance that opens brilliantly and finishes thin is a dinner with a fight about the bill at the end of it.

It also suggests that the common complaint about a scent having no lasting power is often mis-stated. What is usually meant is not that the fragrance vanished, but that it ended without a conclusion. Something that fades to a clean skin note is not a failure of longevity. Something that fades into a vague sweetness with no relationship to what came before is a failure of ending, and the two get reported with the same words.

Extend the rule past the liquid and it applies to a house entire. The customer's peak is usually the first wearing, the moment the thing works on their own skin and they understand what they bought. Their ending is not the purchase. It is the last interaction they have with you: the note in the parcel, the reply when something went wrong, the way an order was handled six months later. Almost every business spends its budget on the first impression and its attention on the middle, and then allows the ending to be handled by whoever happens to be available.

The instruction that falls out of this is unglamorous and slightly uncomfortable, because it means deliberately engineering how things conclude, which can feel like a trick. It is only a trick if the middle is a lie. Build a good experience, put the intensity where it belongs, and then finish it properly, because the finish is the part that will be carried out of the room and remembered as the whole.