Each chef makes their menu available to the public in their restaurant. All other things being equal, which menu is the most successful over the next six months? The answer is Chef B's menu.
When Chef A and B meet up again in 6 months time, Chef A is dumbfounded that his award wining menu has been outsold by the runner up and it is Chef B now drinking champagne. Chef B says his success is down to word of mouth recommendations. This isn't a trick in lateral thinking or a gimmick, so what gives?

The story here (taken from Kahneman's book) starts out with colonoscopies, painful colonoscopies. The question under investigation was how we understand our experience of pain or pleasure. Patients undergoing the medical procedure were asked to rate at regular intervals how much pain they were in. At the end they were asked to rate how painful the overall procedure was. The results were astonishing. Patients that endured a 10 minute procedure with a peak pain (rated 8/10) that quickly came to an end rated the operation more painful than those who endured a 25 minute procedure, also with peak pain rated 8/10 but where the pain subsided towards the end of the procedure. This seemed very odd, Patient B endured more pain than Patient A (same intensity but longer), but rated the procedure less painful.
Kahneman conducted further experiments in the area. Volunteers were told they were to undergo three tests. In the first test, they put their hand in a bath of painfully cold water for 60 seconds. In the second test, they were asked to put their hand in the same bath for 60 seconds after which time hot air was pumped into the bath to raise the water temperature modestly where their hand remained for a further 30 seconds. Put another way, Test 2 is Test 1 plus 30 seconds more of a lower level of pain. For Test 3, the subjects were told they could choose which test to undergo: they could either repeat Test 1 or repeat Test 2. 80% chose to undergo Test 2 so enduring 30 seconds of additional and unnecessary pain. Again, what gives?

Confusing experience with the memory of it is a compelling cognitive illusion... the experiencing self does not have a voice. The remembering self is sometimes wrong, but it is the one that keeps score... and it is the one that makes decisions.
Using statistical analysis in these and numerous other experiments, he defined the peak-end rule for memory:
The retrospective rating [of pain/pleasure] was well predicted by the average of the level of pain (pleasure) reported in the worst moment of the experience and at its end.
This then is how we remember and this is how we decide. We care about peaks, and we care about how it ended, other stuff simply gets lost in the event of recall. I wondered how this applied to my memories which in turn led me to wonder if my recall of meals that I liked or disliked at restaurants were from my 'experiencing self' or my 'remembering self'. The more I thought about it, the more Kuhneman's discussion made sense. In a Hollywood blockbuster film, the biggest explosion should be at the end of the film. Why? Because when you leave the cinema, the ending is disproportionately important in how you remember the film.
What does this mean for restaurants? It means that if you want customers to remember that they had a good time, then the meal must include both a distinct high point and a strong ending. Three practical implications immediately come to mind:
- the best (remembered) meals need a 'peak' enjoyment point. In reviews and blogs, this is surely what is termed 'the wow factor'. How many times have people said of the high point of the meal 'it's worth going back for that alone'. This is a crucial verbal clue that this particular thing is strongly memory forming and will therefore play a disproportionate part in how the meal is remembered.
- the last food that the customer eats will be particularly important in their memory of the restaurant. In a full meal, that means that a weak or strong dessert will disproportionately affect customers memories of the restaurant when later asked 'what's the food like?'. Presumably, limp petits fours should also be avoided (always of course but especially) if customers consider these as part of the meal.
- 'what's the restaurant like' is a very different question to 'what's the food like' and it may be the customer's last experience of the restaurant that is memory forming. This will often be receiving/paying the bill, and we all know how irritating it is waiting and wanting to pay the bill when there's no staff around to help. A very costly error it would seem, more than most restaurants recognise.

The Fat Duck for example gives every diner before they leave the restaurant a bag of brilliant confectionery, his famous 'kids in a sweet shop' offering. Accordingly, the last experience you have of The Fat Duck is not receiving a bill (unpleasant), but of receiving a gift (very pleasant indeed) which according to the above discussion is an absolute game changer in how you will remember The Fat Duck. Heston's flagship is arguably lavished with more praise than any other restaurant in the UK.
Still don't think so? Would it be the same or different if the sweet bag, the restaurant's 'present' to you was given at the beginning of the meal just after you sat down, let's call it a welcome gift? Under these circumstances it loses all impact. Off the top of my head, I can think of half a dozen Mayfair restaurants that also give their customers some kind of treat to take home at the end of the meal. Somewhere along the line, they've understood this makes sense even if they can't always articulate why.
At el Bulli, the final flourish was a chocolate box that was brought to the table following dessert, I've never seen one better. We had 40 courses at el Bulli, I can now remember less than half without prompting, but I'll never forget the chocolate box; is that because it was better than the other courses, or because it was last? Possibly both if Adria was aware of the memory formation phenomena so ensuring your very last experience of el Bulli was sublime. Similarly with Paul Ainsworth at Number 6, a brilliant meal was followed by his award winning dessert 'a Trip to the Fairground' which is arguably the most memorable dessert we have ever had in a restaurant: genius?
We started off considering two menus by Chef A and Chef B scoring 30 points and 27 points respectively. When the peak-end rule is applied however, Chef A scores 7 points [(8+6)/2] while Chef B scores 8.5 [(9+8)/2] having a higher peak and a stronger finish. The judges mark the meal 'as experienced' so Chef A wins, but the people who go to the restaurant will tell their friends about it the next day on an 'as remembered' basis so will remember Chef B's menu as being the better of the two.
There's lots of important considerations for Kahneman's findings in many important areas, but for restaurants, there are some very crucial lessons to take on board if true. Given the choice for example of improving every course on the menu a little bit or improving one course by lots, go with the latter to increase peak enjoyment. Ask yourself if the meal has a wow factor rather than 'is it consistently good'. Dessert should not be neglected and always make sure a diner's last experience of the restaurant is a smile, a goodbye and a thank you. Diners themselves may disagree with this analysis but presumably they are the same ones who, if the hand in the cold water bath experiment is true, voted for more pain rather than less, for rational disagreement is the very essence of a cognitive illusion.
For sure, in all of this Daniel Kahneman may not be right... but he has a Nobel Prize that says he is.