1
00:00:06,130 --> 00:00:09,436
Hi my name is Leslie John, and I am a 
professor of marketing at Harvard 

2
00:00:09,436 --> 00:00:13,376
Business School. 
In my job I study why people make 

3
00:00:13,376 --> 00:00:17,654
irrational decisions. 
And I try to use these moments of 

4
00:00:17,654 --> 00:00:21,610
irrationality to help people help 
themselves. 

5
00:00:21,610 --> 00:00:25,630
In other words, I try to take all the 
silly things that we do and I try to use 

6
00:00:25,630 --> 00:00:30,914
them to people's own benefit. 
So have you ever wondered why we often 

7
00:00:30,914 --> 00:00:34,970
fail to do things that we know are good 
for ourselves? 

8
00:00:34,970 --> 00:00:40,342
For example, we buy gym club memberships 
only to see the membership cards collect 

9
00:00:40,342 --> 00:00:44,155
dust. 
We regularly pledge to exercise more, but 

10
00:00:44,155 --> 00:00:49,144
then day after day we come home to veg 
out in front of the computer. 

11
00:00:49,144 --> 00:00:54,160
People try to take their prescription 
medications on time, but they regularly 

12
00:00:54,160 --> 00:00:58,904
fail to do so. 
Or put differently, have you ever 

13
00:00:58,904 --> 00:01:05,62
wondered why people knowingly engage in 
harmful behavior like risky sex? 

14
00:01:05,62 --> 00:01:11,490
We'll, a lot of these behaviors can be 
explained by lapses in self-control. 

15
00:01:11,490 --> 00:01:14,870
And today, I'm going to talk to you about 
a way that we can try to overcome some of 

16
00:01:14,870 --> 00:01:20,112
these problems. 
So it should come really as no surprise 

17
00:01:20,112 --> 00:01:25,263
that if you pay people to do things, 
they'll often do things for you. 

18
00:01:25,263 --> 00:01:29,103
But it turns out that this is not the 
most efficient way to get people to do 

19
00:01:29,103 --> 00:01:32,950
things. 
My colleagues and I, Kevin Volpe and 

20
00:01:32,950 --> 00:01:37,834
George Lowenstein, we've been working on 
ways to make these expenses work even 

21
00:01:37,834 --> 00:01:44,130
better as in change more behavior. 
So lots of employers, they will pay for 

22
00:01:44,130 --> 00:01:47,586
their employees to use the gym, hoping 
that their employees will then use the 

23
00:01:47,586 --> 00:01:52,896
gym and be healthier. 
But we've been finding that these simple 

24
00:01:52,896 --> 00:01:58,356
direct subsidization, they're not really 
that efficient in getting people to go to 

25
00:01:58,356 --> 00:02:02,109
gym. 
It turns out that there's a way that 

26
00:02:02,109 --> 00:02:07,334
gives you much more bang for your buck. 
So what we do is we, the trick is instead 

27
00:02:07,334 --> 00:02:13,610
of just paying people to do things, we 
capitalize on people's irrationality. 

28
00:02:13,610 --> 00:02:17,563
So, in other words, we try to take these 
biases that people have in their 

29
00:02:17,563 --> 00:02:21,610
decisions and use them to people's 
advantage. 

30
00:02:21,610 --> 00:02:26,362
So today, I'll tell you about some work 
we've done where we've used two biases to 

31
00:02:26,362 --> 00:02:30,754
try to help people to lose weight, that 
is one, the over weighting of small 

32
00:02:30,754 --> 00:02:37,290
probabilities and two, loss aversion. 
Let me explain, so over weighting of 

33
00:02:37,290 --> 00:02:41,382
small probabilities. 
Have you ever wondered why people play a 

34
00:02:41,382 --> 00:02:47,460
lottery even though there's almost zero 
chance that they'll win? 

35
00:02:47,460 --> 00:02:52,603
Or why people bet on underdogs? 
Well, it turns out that we human beings, 

36
00:02:52,603 --> 00:02:56,153
we're actually really bad at 
understanding and estimating 

37
00:02:56,153 --> 00:03:01,450
probabilities. 
So, we tend to overestimate, extremely 

38
00:03:01,450 --> 00:03:05,548
unlikely events. 
So, extremely unlikely events feel to us 

39
00:03:05,548 --> 00:03:10,600
like they're much more likely to happen 
than they actually are. 

40
00:03:10,600 --> 00:03:15,712
So in other words, these low probability 
events exert a disproportionate influence 

41
00:03:15,712 --> 00:03:21,500
on our judgments and decisions. 
And this can explain a lot of crazy 

42
00:03:21,500 --> 00:03:26,808
beliefs that people have. 
For example, an ungodly number of 

43
00:03:26,808 --> 00:03:31,596
Americans think that they will win the 
lottery and in fact some people truly 

44
00:03:31,596 --> 00:03:37,772
believe that Winning the lottery. 
For them, winning the lottery is their 

45
00:03:37,772 --> 00:03:41,150
retirement plan. 
So how can we use this normally 

46
00:03:41,150 --> 00:03:46,620
pernicious bias to help people? 
Well, what let me explain how we did this 

47
00:03:46,620 --> 00:03:50,713
with weight loss. 
So instead of just paying people to lose 

48
00:03:50,713 --> 00:03:55,649
weight we took advantage of lotteries. 
So instead of paying someone a dollar to 

49
00:03:55,649 --> 00:04:00,200
lose weight. 
We gave people, say, a 1% chance of 

50
00:04:00,200 --> 00:04:04,380
winning $100. 
So importantly, the cost is the same, but 

51
00:04:04,380 --> 00:04:09,700
the difference in the impact on human 
behavior is actually quite extraordinary. 

52
00:04:09,700 --> 00:04:14,55
So what we did was we gave participants 
in our study a weight loss goal, so we 

53
00:04:14,55 --> 00:04:19,830
told them every day what their weight 
loss goal was. 

54
00:04:19,830 --> 00:04:24,252
And at the beginning of the study, each 
participant got to choose a lucky number, 

55
00:04:24,252 --> 00:04:28,975
a two digit lucky number. 
And now, every day of the study, we put 

56
00:04:28,975 --> 00:04:34,220
all the numbers into a hat and we drew 
out one lucky number every day. 

57
00:04:34,220 --> 00:04:40,520
So if your lucky number was drawn, you 
could win $100, pretty great. 

58
00:04:40,520 --> 00:04:45,545
But there was one catch and there was one 
aspect that made this lottery scheme way 

59
00:04:45,545 --> 00:04:49,715
more potent. 
That is, you would only get the $100 if 

60
00:04:49,715 --> 00:04:54,660
you had actually attained your weight 
loss goal that day. 

61
00:04:54,660 --> 00:04:58,412
So, what we did was, suppose your number 
was drawn, but you hadn't lost the weight 

62
00:04:58,412 --> 00:05:03,650
that you were supposed to lose that day. 
What we did, we sent you a really mean 

63
00:05:03,650 --> 00:05:07,594
text message that said well 
congratulations, you're number came up, 

64
00:05:07,594 --> 00:05:13,910
but unfortunately, you didn't meet your 
weight loss goal today. 

65
00:05:13,910 --> 00:05:17,760
You would have won a $100 if you had only 
lost the weight. 

66
00:05:17,760 --> 00:05:21,864
so this incentive scheme, not only does 
it leverage the power of over weighting 

67
00:05:21,864 --> 00:05:25,284
of small probabilities, it feels like 
you're more likely to win than you 

68
00:05:25,284 --> 00:05:30,490
actually are. 
But it also harnesses the strong 

69
00:05:30,490 --> 00:05:34,910
motivations that people have to avoid 
regret, avoid that sense of, oh, what if 

70
00:05:34,910 --> 00:05:39,650
my number came up and I hadn't lost the 
weight. 

71
00:05:39,650 --> 00:05:44,72
so the second bias that I'll to you about 
before I tell you about the results of 

72
00:05:44,72 --> 00:05:48,818
this study. 
We also in this study made use of loss 

73
00:05:48,818 --> 00:05:54,220
aversion, which is a classic bias, loss 
aversion. 

74
00:05:54,220 --> 00:06:00,0
So loss aversion, refers to how the pain 
of losing something is much more poignant 

75
00:06:00,0 --> 00:06:05,525
to us, then is the pleasure that we get 
with an objectively equivalently sized 

76
00:06:05,525 --> 00:06:11,333
gain. 
So in other words, people hate losing 

77
00:06:11,333 --> 00:06:15,190
money way more than they enjoy getting 
money. 

78
00:06:15,190 --> 00:06:20,440
And just like the over weighting of small 
probabilities, this bias can explain a 

79
00:06:20,440 --> 00:06:25,315
lot of suboptimal behavior, for example, 
it explains why investors hold onto 

80
00:06:25,315 --> 00:06:31,720
losing stocks for far too long, when 
really they should. 

81
00:06:31,720 --> 00:06:34,658
They should see their losers. 
They should get rid of their losing 

82
00:06:34,658 --> 00:06:37,153
stocks. 
And again just like with the over 

83
00:06:37,153 --> 00:06:41,47
weighting small probabilities we use this 
bias in this case to help people lose 

84
00:06:41,47 --> 00:06:44,500
weight. 
How do we do this? 

85
00:06:44,500 --> 00:06:48,973
So what we did was we gave dieters the 
opportunity to put their own money down 

86
00:06:48,973 --> 00:06:53,542
towards losing weight. 
So they could wager their own money 

87
00:06:53,542 --> 00:06:57,141
towards losing weight. 
So they put their money down, and if they 

88
00:06:57,141 --> 00:07:01,17
didn't lose their weight, if they didn't 
lose, if they didn't attain their weight 

89
00:07:01,17 --> 00:07:05,256
loss goal, then they forfeited or lost 
their money. 

90
00:07:05,256 --> 00:07:09,569
If they did attain their goal then they 
would get their money back, plus a match. 

91
00:07:09,569 --> 00:07:12,260
So in other words, they were make, they 
were making a bet. 

92
00:07:12,260 --> 00:07:15,476
They were begging on themselves that they 
would lose weight. 

93
00:07:15,476 --> 00:07:20,588
Now, how do these two different incentive 
systems that are designed to capitalize 

94
00:07:20,588 --> 00:07:25,274
on irrationality, how do these incentive 
systems fare when we compare them to a 

95
00:07:25,274 --> 00:07:31,472
control incentive condition? 
Well, the results after four months, 

96
00:07:31,472 --> 00:07:36,229
participants that were randomly assigned 
to use these incentive schemes, they lost 

97
00:07:36,229 --> 00:07:42,31
way more weight than the people in the 
control condition, good news. 

98
00:07:42,31 --> 00:07:46,855
however, the tricky thing, even trickier 
than losing weight in the first place, is 

99
00:07:46,855 --> 00:07:51,176
keeping weight off once we remove the 
incentives. 

100
00:07:51,176 --> 00:07:54,419
And I think I'll save that for another 
lecture or two altogether because that's 

101
00:07:54,419 --> 00:07:57,906
another big problem that we're currently 
trying to solve. 

102
00:07:57,906 --> 00:08:03,106
so summing up, today I've talked about 
how we can leverage or capitalize on 

103
00:08:03,106 --> 00:08:08,306
people's irrationality to help people 
make better decisions and live better 

104
00:08:08,306 --> 00:08:13,93
lives. 
Thanks.