How to use data to make a hit TV show | Sebastian Wernicke

133,338 views ใƒป 2016-01-27

TED


ืื ื ืœื—ืฅ ืคืขืžื™ื™ื ืขืœ ื”ื›ืชื•ื‘ื™ื•ืช ื‘ืื ื’ืœื™ืช ืœืžื˜ื” ื›ื“ื™ ืœื”ืคืขื™ืœ ืืช ื”ืกืจื˜ื•ืŸ.

ืžืชืจื’ื: Ido Dekkers ืžื‘ืงืจ: Zeeva Livshitz
00:12
Roy Price is a man that most of you have probably never heard about,
0
12820
4276
ืจื•ื™ ืคืจื™ื™ืก ื”ื•ื ืื“ื ืฉืกื‘ื™ืจ ืœื”ื ื™ื— ืฉืจื•ื‘ื›ื ื›ื ืจืื” ืœื ืฉืžืขืชื ืขืœื™ื• ืžืขื•ืœื,
00:17
even though he may have been responsible
1
17120
2496
ืœืžืจื•ืช ืฉื™ื™ืชื›ืŸ ื•ื”ื•ื ื”ื™ื” ืื—ืจืื™
00:19
for 22 somewhat mediocre minutes of your life on April 19, 2013.
2
19640
6896
ืœื‘ืขืจืš 22 ื“ืงื•ืช ื‘ื™ื ื•ื ื™ื•ืช ืฉืœ ื—ื™ื™ื›ื ื‘19 ืœืืคืจื™ืœ 2013.
00:26
He may have also been responsible for 22 very entertaining minutes,
3
26560
3176
ื™ื›ื•ืœ ืœื”ื™ื•ืช ืฉื”ื•ื ื”ื™ื” ื’ื ืื—ืจืื™ ืœ22 ื“ืงื•ืช ืžืื•ื“ ืžืฉืขืฉืขื•ืช,
00:29
but not very many of you.
4
29760
2256
ืื‘ืœ ืœื ืจื‘ื™ื ืžื›ื.
00:32
And all of that goes back to a decision
5
32040
1896
ื•ื›ืœ ื–ื” ื—ื•ื–ืจ ืœื”ื—ืœื˜ื”
00:33
that Roy had to make about three years ago.
6
33960
2000
ืฉืจื•ื™ ื”ื™ื ืฆืจื™ืš ืœืงื‘ืœ ืœืคื ื™ 3 ืฉื ื™ื.
00:35
So you see, Roy Price is a senior executive with Amazon Studios.
7
35984
4832
ืื– ืืชื ืžื‘ื™ื ื™ื, ืจื•ื™ ืคืจื™ื™ืก ื”ื•ื ืžื ื”ืœ ื‘ื›ื™ืจ ื‘ืื•ืœืคื ื™ ืืžื–ื•ืŸ.
00:40
That's the TV production company of Amazon.
8
40840
3016
ื–ืืช ื—ื‘ืจืช ื”ืคืงื•ืช ื”ื˜ืœื•ื•ื™ื–ื™ื” ืฉืœ ืืžื–ื•ืŸ.
00:43
He's 47 years old, slim, spiky hair,
9
43880
3256
ื”ื•ื ื‘ืŸ 47, ืจื–ื”, ืฉื™ืขืจ ืขื ืงื•ืฆื™ื,
00:47
describes himself on Twitter as "movies, TV, technology, tacos."
10
47160
4816
ื”ื•ื ืžืชืืจ ืขืฆืžื• ื‘ื˜ื•ื•ื™ื˜ืจ ื› "ืกืจื˜ื™ื, ื˜ืœื•ื•ื™ื–ื™ื”, ื˜ื›ื ื•ืœื•ื’ื™ื”, ื˜ืืงื•."
00:52
And Roy Price has a very responsible job, because it's his responsibility
11
52000
5176
ื•ืœืจื•ื™ ืคืจื™ื™ืก ื™ืฉ ื™ืฉ ืื—ืจื™ื•ืช ืจื‘ื” ื‘ืขื‘ื•ื“ื” ื›ื™ ื”ืื—ืจื™ื•ืช ืฉืœื• ื”ื™ื
00:57
to pick the shows, the original content that Amazon is going to make.
12
57200
4056
ืœื‘ื—ื•ืจ ืืช ื”ืชื•ื›ื ื™ื•ืช, ืืช ื”ืชื•ื›ืŸ ื”ืžืงื•ืจื™ ืฉืืžื–ื•ืŸ ืชื™ืฆื•ืจ.
01:01
And of course that's a highly competitive space.
13
61280
2336
ื•ื›ืžื•ื‘ืŸ ืฉื–ื” ืชื—ื•ื ืชื—ืจื•ืชื™ ืžืื•ื“.
01:03
I mean, there are so many TV shows already out there,
14
63640
2736
ืื ื™ ืžืชื›ื•ื•ืŸ, ื™ืฉ ื›ืœ ื›ืš ื”ืจื‘ื” ืชื•ื›ื ื™ื•ืช ื˜ืœื•ื•ื™ื–ื™ื” ืฉื›ื‘ืจ ืงื™ื™ืžื•ืช,
01:06
that Roy can't just choose any show.
15
66400
2176
ื•ืจื•ื™ ืœื ืคืฉื•ื˜ ื™ื›ื•ืœ ืœื‘ื—ื•ืจ ื›ืœ ืื—ืช.
01:08
He has to find shows that are really, really great.
16
68600
4096
ื”ื•ื ื—ื™ื™ื‘ ืœืžืฆื•ื ืชื•ื›ื ื™ื•ืช ืžืžืฉ ืžืžืฉ ื˜ื•ื‘ื•ืช.
01:12
So in other words, he has to find shows
17
72720
2816
ืื– ื‘ืžื™ืœื™ื ืื—ืจื•ืช, ื”ื•ื ื—ื™ื™ื‘ ืœืžืฆื•ื ืชื•ื›ื ื™ื•ืช
01:15
that are on the very right end of this curve here.
18
75560
2376
ืฉื ืžืฆืื•ืช ืžืžืฉ ืขืœ ื”ืงืฆื” ื”ื™ืžื ื™ ื‘ืขืงื•ืžื” ื”ื–ืืช.
01:17
So this curve here is the rating distribution
19
77960
2656
ื”ืขืงื•ืžื” ื”ื–ืืช ืžืฆื™ื’ื” ืืช ื—ืœื•ืงืช ื”ืจื™ื™ื˜ื™ื ื’
01:20
of about 2,500 TV shows on the website IMDB,
20
80640
4376
ืฉืœ ื‘ืขืจืš 2500 ืชื•ื›ื ื™ื•ืช ื˜ืœื•ื•ื™ื–ื™ื” ื‘ืืชืจ IMDB,
01:25
and the rating goes from one to 10,
21
85040
2896
ื•ื”ืจื™ื™ื˜ื™ื ื’ ื ืข ื‘ื™ืŸ 1 ืœ10,
01:27
and the height here shows you how many shows get that rating.
22
87960
2976
ื•ื”ื’ื•ื‘ื” ื›ืืŸ, ืžืจืื” ืœื›ื ื›ืžื” ืชื•ื›ื ื™ื•ืช ืงื‘ืœื• ืืช ื”ืจื™ื™ื˜ื™ื ื’ ื”ื–ื”.
01:30
So if your show gets a rating of nine points or higher, that's a winner.
23
90960
4696
ืื– ืื ื”ืชื•ื›ื ื™ืช ืฉืœื›ื ืžืงื‘ืœืช ืจื™ื™ื˜ื™ื ื’ ืฉืœ 9 ื ืงื•ื“ื•ืช ื•ืžืขืœื”, ื–ืืช ื”ืฆืœื—ื”.
01:35
Then you have a top two percent show.
24
95680
1816
ื•ืื– ื™ืฉ ืœื›ื ืืช ื”ืชื•ื›ื ื™ื•ืช ืฉื ืžืฆืื•ืช ื‘2 ืื—ื•ื– ื”ืขืœื™ื•ื ื™ื.
01:37
That's shows like "Breaking Bad," "Game of Thrones," "The Wire,"
25
97520
3896
ืชื•ื›ื ื™ื•ืช ื›ืžื• "ืฉื•ื‘ืจ ืฉื•ืจื•ืช", "ืžืฉื—ืงื™ ื”ื›ืก", "ื”ืกืžื•ื™ื”",
01:41
so all of these shows that are addictive,
26
101440
2296
ืื– ื›ืœ ื”ืชื•ื›ื ื™ื•ืช ื”ืืœื” ืฉืžืžื›ืจื•ืช,
01:43
whereafter you've watched a season, your brain is basically like,
27
103760
3056
ื›ืฉืื—ืจื™ ืฉืืชื ืจื•ืื™ื ืขื•ื ื”, ื”ืžื•ื— ืฉืœื›ื ื‘ืขืฆื ื›ืžื•,
01:46
"Where can I get more of these episodes?"
28
106840
2176
"ืื™ืคื” ืื ื™ ื™ื›ื•ืœ ืœืžืฆื•ื ืขื•ื“ ืžื”ืคืจืงื™ื ื”ืืœื”?"
01:49
That kind of show.
29
109040
1200
ืกื•ื’ ื›ื–ื” ืฉืœ ืชื•ื›ื ื™ื•ืช.
01:50
On the left side, just for clarity, here on that end,
30
110920
2496
ื‘ืฆื“ ื”ืฉืžืืœื™, ืจืง ืœืฉื ื”ื‘ื”ืจื”, ื›ืืŸ ื‘ืงืฆื” ื”ื–ื”,
01:53
you have a show called "Toddlers and Tiaras" --
31
113440
3176
ื ืžืฆืืช ืชื•ื›ื ื™ื•ืช ืฉื ืงืจืืช "ืคืขื•ื˜ื•ืช ื•ื›ืชืจื™ื"--
01:56
(Laughter)
32
116640
2656
(ืฆื—ื•ืง)
01:59
-- which should tell you enough
33
119320
1536
--ื•ื–ื” ืืžื•ืจ ืœื”ื’ื™ื“ ืœื›ื ืžืกืคื™ืง
02:00
about what's going on on that end of the curve.
34
120880
2191
ื‘ื ื•ื’ืข ืœืžื” ืฉืงื•ืจื” ื‘ืงืฆื” ื”ื–ื” ืฉืœ ื”ืขืงื•ืžื”.
02:03
Now, Roy Price is not worried about getting on the left end of the curve,
35
123095
4161
ืขื›ืฉื™ื•, ืจื•ื™ ืคืจื™ื™ืก ืœื ืžื•ื“ืื’ ืžืœื”ื’ื™ืข ืœืงืฆื” ื”ืฉืžืืœื™ ืฉืœ ื”ืขืงื•ืžื”,
02:07
because I think you would have to have some serious brainpower
36
127280
2936
ื‘ื’ืœืœ ืฉืื ื™ ื—ื•ืฉื‘ ืฉืฆืจื™ืš ื™ื›ื•ืœืช ื—ืฉื™ื‘ื” ืจืฆื™ื ื™ืช
02:10
to undercut "Toddlers and Tiaras."
37
130240
1696
ื‘ืฉื‘ื™ืœ ืœื’ื‘ื•ืจ ืขืœ "ืคืขื•ื˜ื•ืช ื•ื›ืชืจื™ื".
02:11
So what he's worried about is this middle bulge here,
38
131960
3936
ืื– ืžื” ืฉื”ื•ื ืžื•ื“ืื’ ืžืžื ื• ื”ื•ื ื”ื‘ืœื™ื˜ื” ื”ืืžืฆืขื™ืช ืฉื ืžืฆืืช ื›ืืŸ,
02:15
the bulge of average TV,
39
135920
1816
ื”ื‘ืœื™ื˜ื” ืฉืœ ื˜ืœื•ื•ื™ื–ื™ื” ืžืžื•ืฆืขืช,
02:17
you know, those shows that aren't really good or really bad,
40
137760
2856
ืืชื ื™ื•ื“ืขื™ื, ื”ืชื•ื›ื ื™ื•ืช ื”ืืœื” ืฉืœื ื›ืœ ื›ืš ื˜ื•ื‘ื•ืช ื•ืœื ื›ืœ ื›ืš ืจืขื•ืช,
02:20
they don't really get you excited.
41
140639
1656
ืœื ื‘ืืžืช ื’ื•ืจืžื•ืช ืœืš ืœื”ืชืจื’ืฉ.
02:22
So he needs to make sure that he's really on the right end of this.
42
142320
4856
ืื– ื”ื•ื ืฆืจื™ืš ืœื•ื•ื“ื ืฉื”ื•ื ื ืžืฆื ืžืžืฉ ื‘ืงืฆื” ื”ื™ืžื ื™ ืฉืœ ื–ื”.
02:27
So the pressure is on,
43
147200
1576
ืื– ื™ืฉ ืœื—ืฅ,
02:28
and of course it's also the first time
44
148800
2176
ื•ื›ืžื•ื‘ืŸ, ื–ืืช ื’ื ื”ืคืขื ื”ืจืืฉื•ื ื”
02:31
that Amazon is even doing something like this,
45
151000
2176
ืฉืืžื–ื•ืŸ ืขื•ืฉื” ืžืฉื”ื• ื›ื–ื”,
02:33
so Roy Price does not want to take any chances.
46
153200
3336
ืื– ืจื•ื™ ืคืจื™ื™ืก ืœื ืจื•ืฆื” ืœืงื—ืช ืกื™ื›ื•ื ื™ื.
02:36
He wants to engineer success.
47
156560
2456
ื”ื•ื ืจื•ืฆื” ืœื”ื ื“ืก ืืช ื”ื”ืฆืœื—ื”.
02:39
He needs a guaranteed success,
48
159040
1776
ื”ื•ื ืฆืจื™ืš ื”ืฆืœื—ื” ืžื•ื‘ื˜ื—ืช,
02:40
and so what he does is, he holds a competition.
49
160840
2576
ืื– ืžื” ืฉื”ื•ื ืขื•ืฉื”, ื”ื•ื ืขื•ืจืš ืชื—ืจื•ืช.
02:43
So he takes a bunch of ideas for TV shows,
50
163440
3136
ืื– ื”ื•ื ืœื•ืงื— ืžืกืคืจ ืจืขื™ื•ื ื•ืช ืœืชื•ื›ื ื™ื•ืช ื˜ืœื•ื•ื™ื–ื™ื”,
02:46
and from those ideas, through an evaluation,
51
166600
2296
ื•ืžื”ืจืขื™ื•ื ื•ืช ื”ืืœื”, ื‘ืืžืฆืขื•ืช ื”ืขืจื›ื”,
02:48
they select eight candidates for TV shows,
52
168920
4096
ื”ื ื‘ื•ื—ืจื™ื 8 ืžื•ืขืžื“ื™ื ืœืชื•ื›ื ื™ื•ืช ื˜ืœื•ื•ื™ื–ื™ื”,
02:53
and then he just makes the first episode of each one of these shows
53
173040
3216
ื•ืื– ื”ื•ื ืคืฉื•ื˜ ื™ื•ืฆืจ ืืช ื”ืคืจืง ื”ืจืืฉื•ืŸ ื‘ื›ืœ ืื—ืช ืžื”ืชื•ื›ื ื™ื•ืช ื”ืืœื”
02:56
and puts them online for free for everyone to watch.
54
176280
3136
ื•ืžืขืœื” ืื•ืชื ืœืจืฉืช ื‘ื—ื™ื ื ืฉื›ืœ ืื—ื“ ื™ื›ื•ืœ ืœืฆืคื•ืช.
02:59
And so when Amazon is giving out free stuff,
55
179440
2256
ืื– ื›ืฉืืžื–ื•ืŸ ืžื—ืœืงืช ื“ื‘ืจื™ื ื‘ื—ื™ื ื,
03:01
you're going to take it, right?
56
181720
1536
ืืชื ืชืงื—ื• ืื•ืชื, ื ื›ื•ืŸ?
03:03
So millions of viewers are watching those episodes.
57
183280
5136
ืื– ืžื™ืœื™ื•ื ื™ ืฆื•ืคื™ื ืจื•ืื™ื ืืช ื”ืคืจืงื™ื ื”ืืœื”.
03:08
What they don't realize is that, while they're watching their shows,
58
188440
3216
ืžื” ืฉื”ื ืœื ืžื‘ื™ื ื™ื, ื‘ื–ืžืŸ ืฉื”ื ืฆื•ืคื™ื ื‘ืชื•ื›ื ื™ื•ืช ืฉืœื”ื,
03:11
actually, they are being watched.
59
191680
2296
ื‘ืขืฆื, ืขื•ืงื‘ื™ื ืื—ืจื™ื”ื.
03:14
They are being watched by Roy Price and his team,
60
194000
2336
ื”ื ื ืฆืคื™ื ืขืœ ื™ื“ื™ ืจื•ื™ ืคืจื™ื™ืก ื•ื”ืฆื•ื•ืช ืฉืœื•,
03:16
who record everything.
61
196360
1376
ืฉืžืชืขื“ื™ื ื”ื›ืœ.
03:17
They record when somebody presses play, when somebody presses pause,
62
197760
3376
ื”ื ืžืชืขื“ื™ื ื›ืฉืžื™ืฉื”ื• ืœื•ื—ืฅ ื ื’ืŸ, ื•ื›ืฉืžื™ืฉื”ื• ืœื•ื—ืฅ ื”ืฉื”ื”,
03:21
what parts they skip, what parts they watch again.
63
201160
2536
ืขืœ ืื™ื–ื” ื—ืœืงื™ื ื”ืฆื•ืคื™ื ืžื“ืœื’ื™ื, ื•ืื™ื–ื” ื—ืœืงื™ื ื”ื ืจื•ืื™ื ืฉื•ื‘.
03:23
So they collect millions of data points,
64
203720
2256
ืื– ื”ื ืื•ืกืคื™ื ืžื™ืœื™ื•ื ื™ ื ืงื•ื“ื•ืช ืžื™ื“ืข,
03:26
because they want to have those data points
65
206000
2096
ื‘ื’ืœืœ ืฉื”ื ืจื•ืฆื™ื ืœื”ืฉืชืžืฉ ื‘ื ืชื•ื ื™ื ื”ืืœื”
03:28
to then decide which show they should make.
66
208120
2696
ื›ื“ื™ ืœื”ื—ืœื™ื˜ ืื™ื–ื” ืชื•ื›ื ื™ืช ื”ื ื™ื™ืฆืจื•.
03:30
And sure enough, so they collect all the data,
67
210840
2176
ื•ื›ืžื•ื‘ืŸ, ื›ืš ื”ื ืื•ืกืคื™ื ืืช ื›ืœ ื”ืžื™ื“ืข,
03:33
they do all the data crunching, and an answer emerges,
68
213040
2576
ื”ื ืขื•ืฉื™ื ืืช ื›ืœ ืขื™ื‘ื•ื“ ื”ืžื™ื“ืข, ื•ืชืฉื•ื‘ื” ืขื•ืœื”,
03:35
and the answer is,
69
215640
1216
ื•ื”ืชืฉื•ื‘ื” ื”ื™ื,
03:36
"Amazon should do a sitcom about four Republican US Senators."
70
216880
5536
"ืืžื–ื•ืŸ ืฆืจื™ื›ื” ืœื™ืฆื•ืจ ืกื™ื˜ืงื•ื ืขืœ ืืจื‘ืขื” ืกื ื˜ื•ืจื™ื ืจืคื•ื‘ืœื™ืงืื™ื."
03:42
They did that show.
71
222440
1216
ื”ื ืขืฉื• ืืช ื”ืชื•ื›ื ื™ืช ื”ื”ื™ื.
03:43
So does anyone know the name of the show?
72
223680
2160
ืื– ื”ืื ื›ื•ืœื ืžื›ื™ืจื™ื ืืช ื”ืฉื ืฉืœ ื”ืชื•ื›ื ื™ืช?
03:46
(Audience: "Alpha House.")
73
226720
1296
(ืงื”ืœ: "ื‘ื™ืช ืืœืคื.")
03:48
Yes, "Alpha House,"
74
228040
1456
ื›ืŸ, "ื‘ื™ืช ืืœืคื."
03:49
but it seems like not too many of you here remember that show, actually,
75
229520
4096
ืื‘ืœ ื ืจืื” ื›ืื™ืœื• ืœื ื”ืจื‘ื” ืžื›ื ืคื” ื–ื•ื›ืจื™ื ืืช ื”ืกื“ืจื” ื”ื”ื™ื, ืœืžืขืฉื”,
03:53
because it didn't turn out that great.
76
233640
1856
ืžืคื ื™ ืฉื”ื™ื ืœื ื™ืฆืื” ื›ื–ื” ื˜ื•ื‘.
03:55
It's actually just an average show,
77
235520
1856
ื”ื™ื ืœืžืขืฉื” ืจืง ืชื•ื›ื ื™ืช ืžืžื•ืฆืขืช,
03:57
actually -- literally, in fact, because the average of this curve here is at 7.4,
78
237400
4576
ืœืžืขืฉื” -- ืžื™ืœื•ืœื™ืช, ืœืžืขืฉื”, ื‘ื’ืœืœ ืฉื”ืžืžื•ืฆืข ืฉืœ ื”ืขืงื•ืžื” ืคื” ื”ื•ื 7.4,
04:02
and "Alpha House" lands at 7.5,
79
242000
2416
ื•"ื‘ื™ืช ืืœืคื" ืงื™ื‘ืœื” 7.5,
04:04
so a slightly above average show,
80
244440
2016
ืื– ืžืขื˜ ืžืขืœ ื”ืชื•ื›ื ื™ืช ื”ืžืžื•ืฆืขืช,
04:06
but certainly not what Roy Price and his team were aiming for.
81
246480
2920
ืื‘ืœ ื‘ื”ื—ืœื˜ ืœื ืžื” ืฉืจื•ื™ ืคืจื™ื™ืก ื•ื”ืฆื•ื•ืช ืฉืœื ื• ื›ื™ื•ื•ื ื• ืืœื™ื•.
04:10
Meanwhile, however, at about the same time,
82
250320
2856
ื‘ื™ื ืชื™ื™ื, ืขื ื–ืืช, ื‘ืขืจืš ื‘ืื•ืชื• ื”ื–ืžืŸ,
04:13
at another company,
83
253200
1576
ื‘ื—ื‘ืจื” ืื—ืจืช,
04:14
another executive did manage to land a top show using data analysis,
84
254800
4216
ืžื ื”ืœ ืื—ืจ ื”ืฆืœื™ื— ืœืคืชื— ืชื•ื›ื ื™ืช ืžื•ืฆืœื—ืช ื‘ืฉื™ืžื•ืฉ ื‘ื ื™ืชื•ื— ืžื™ื“ืข,
04:19
and his name is Ted,
85
259040
1576
ื•ื”ืฉื ืฉืœื• ื”ื•ื ื˜ื“,
04:20
Ted Sarandos, who is the Chief Content Officer of Netflix,
86
260640
3416
ื˜ื“ ืกืจื ื“ื•ืก, ืฉื”ื•ื ืžื ื”ืœ ื”ืชื•ื›ืŸ ื”ืจืืฉื™ ืฉืœ ื ื˜ืคืœื™ืงืก,
04:24
and just like Roy, he's on a constant mission
87
264080
2136
ื•ืžืžืฉ ื›ืžื• ืจื•ื™, ื”ื•ื ื‘ืžืฉื™ืžื” ืงื‘ื•ืขื”
04:26
to find that great TV show,
88
266240
1496
ืœืžืฆื•ื ืชื•ื›ื ื™ืช ื˜ืœื•ื•ื™ื–ื™ื” ืžืขื•ืœื”,
04:27
and he uses data as well to do that,
89
267760
2016
ื•ื”ื•ื ื’ื ื”ืฉืชืžืฉ ื‘ืžื™ื“ืข ื›ื“ื™ ืœืขืฉื•ืช ื–ืืช,
04:29
except he does it a little bit differently.
90
269800
2015
ืื‘ืœ ื”ื•ื ืขื•ืฉื” ื–ืืช ืžืขื˜ ืื—ืจืช.
04:31
So instead of holding a competition, what he did -- and his team of course --
91
271839
3737
ืื– ื‘ืžืงื•ื ืœืขืจื•ืš ืชื—ืจื•ืช, ืžื” ืฉื”ื•ื ืขืฉื” -- ื•ื”ืฆื•ื•ืช ืฉืœื• ื›ืžื•ื‘ืŸ --
04:35
was they looked at all the data they already had about Netflix viewers,
92
275600
3536
ื”ื™ื” ืฉื”ื ื”ื‘ื™ื˜ื• ื‘ื›ืœ ื”ืžื™ื“ืข ืฉื›ื‘ืจ ื”ื™ื” ืœื”ื ืขืœ ืฆื•ืคื™ ื ื˜ืคืœื™ืงืก,
04:39
you know, the ratings they give their shows,
93
279160
2096
ืืชื ื™ื•ื“ืขื™ื, ื”ืจื™ื™ื˜ื™ื ื’ ืฉื”ื ื ื•ืชื ื™ื ืœืชื•ื›ื ื™ื•ืช,
04:41
the viewing histories, what shows people like, and so on.
94
281280
2696
ื”ืกื˜ื•ืจื™ื•ืช ื”ืฆืคื™ื”, ืื™ื–ื” ืชื•ื›ื ื™ื•ืช ืื ืฉื™ื ืื•ื”ื‘ื™ื, ื•ื›ืš ื”ืœืื”.
04:44
And then they use that data to discover
95
284000
1896
ื•ืื– ื”ื ืžืฉืชืžืฉื™ื ื‘ืžื™ื“ืข ื”ื–ื” ื›ื“ื™ ืœื’ืœื•ืช
04:45
all of these little bits and pieces about the audience:
96
285920
2616
ืืช ื›ืœ ืคื™ืกื•ืช ื”ืžื™ื“ืข ื”ืงื˜ื ื•ืช ื”ืืœื• ื‘ื ื•ื’ืข ืœืงื”ืœ:
04:48
what kinds of shows they like,
97
288560
1456
ืื™ื–ื” ืกื•ื’ื™ื ืฉืœ ืชื•ื›ื ื™ื•ืช ื”ื ืื•ื”ื‘ื™ื,
04:50
what kind of producers, what kind of actors.
98
290040
2096
ืื™ื–ื” ืกื•ื’ ืฉืœ ืžืคื™ืงื™ื, ืื™ื–ื” ืกื•ื’ ืฉืœ ืฉื—ืงื ื™ื.
04:52
And once they had all of these pieces together,
99
292160
2576
ื•ื‘ืจื’ืข ืฉื”ื™ื” ืœื”ื ืืช ื›ืœ ืคื™ืกื•ืช ื”ืžื™ื“ืข ื‘ื™ื—ื“,
04:54
they took a leap of faith,
100
294760
1656
ื”ื ืขืฉื• ืงืคื™ืฆื” ืฉืœ ืืžื•ื ื”,
04:56
and they decided to license
101
296440
2096
ื•ื”ื ื”ื—ืœื™ื˜ื• ืœืืฉืจ
04:58
not a sitcom about four Senators
102
298560
2456
ืœื ืกื™ื˜ืงื•ื ืขืœ ืืจื‘ืขื” ืกื ื˜ื•ืจื™ื
05:01
but a drama series about a single Senator.
103
301040
2880
ืืœื ืกื“ืจืช ื“ืจืžื” ืขืœ ืกื ื˜ื•ืจ ื‘ื•ื“ื“.
05:04
You guys know the show?
104
304760
1656
ืืชื ืžื›ื™ืจื™ื ืืช ื”ืชื•ื›ื ื™ืช?
05:06
(Laughter)
105
306440
1296
(ืฆื—ื•ืง)
05:07
Yes, "House of Cards," and Netflix of course, nailed it with that show,
106
307760
3736
ื›ืŸ, "ื‘ื™ืช ื”ืงืœืคื™ื," ื•ื ื˜ืคืœื™ืงืก ื›ืžื•ื‘ืŸ, ื”ืฆืœื™ื—ื• ืขื ื”ืชื•ื›ื ื™ืช ื”ื–ื•,
05:11
at least for the first two seasons.
107
311520
2136
ืœืคื—ื•ืช ื‘ืฉืชื™ ื”ืขื•ื ื•ืช ื”ืจืืฉื•ื ื•ืช.
05:13
(Laughter) (Applause)
108
313680
3976
(ืฆื—ื•ืง) (ืžื—ื™ืื•ืช ื›ืคื™ื™ื)
05:17
"House of Cards" gets a 9.1 rating on this curve,
109
317680
3176
"ื‘ื™ืช ื”ืงืœืคื™ื" ืžืงื‘ืœ ื“ืจื•ื’ ืฉืœ 9.1 ืžื”ืขืงื•ืžื” ื”ื–ื•,
05:20
so it's exactly where they wanted it to be.
110
320880
3176
ืื– ื–ื” ื‘ื“ื™ื•ืง ื‘ืžืงื•ื ื‘ื• ื”ื ืจืฆื• ืฉื–ื” ื™ื”ื™ื”.
05:24
Now, the question of course is, what happened here?
111
324080
2416
ืขื›ืฉื™ื•, ื”ืฉืืœื” ื”ื™ื ื›ืžื•ื‘ืŸ, ืžื” ืงืจื” ืคื”?
05:26
So you have two very competitive, data-savvy companies.
112
326520
2656
ืื– ื™ืฉ ืœื›ื ืฉืชื™ ื—ื‘ืจื•ืช ืชื—ืจื•ืชื™ื•ืช ื•ืชืื‘ื•ืช ืžื™ื“ืข.
05:29
They connect all of these millions of data points,
113
329200
2856
ื”ืŸ ื—ื™ื‘ืจื• ืืช ื›ืœ ืžืœื™ื•ื ื™ ื ืงื•ื“ื•ืช ื”ืžื™ื“ืข ื”ืืœื•,
05:32
and then it works beautifully for one of them,
114
332080
2376
ื•ืื– ื–ื” ืขื‘ื“ ื‘ืื•ืคืŸ ื™ืคื™ืคื” ืœืื—ืช ืžื”ืŸ,
05:34
and it doesn't work for the other one.
115
334480
1856
ื•ื–ื” ืœื ืขื•ื‘ื“ ืœืื—ืจืช.
05:36
So why?
116
336360
1216
ืื– ืœืžื”?
05:37
Because logic kind of tells you that this should be working all the time.
117
337600
3456
ืžืคื ื™ ืฉื”ืœื•ื’ื™ืงื” ืกื•ื’ ืฉืœ ืžืกืคืจืช ืœื›ื ืฉื–ื” ืฆืจื™ืš ืœืขื‘ื•ื“ ื›ืœ ื”ื–ืžืŸ.
05:41
I mean, if you're collecting millions of data points
118
341080
2456
ืื ื™ ืžืชื›ื•ื•ืŸ, ืื ืืชื ืื•ืกืคื™ื ืžืœื™ื•ื ื™ ื ืงื•ื“ื•ืช ืžื™ื“ืข
05:43
on a decision you're going to make,
119
343560
1736
ืขืœ ื”ื—ืœื˜ื” ืฉืืชื ืขื•ืžื“ื™ื ืœืขืฉื•ืช,
05:45
then you should be able to make a pretty good decision.
120
345320
2616
ืื– ืืชื ืฆืจื™ื›ื™ื ืœื”ื™ื•ืช ืžืกื•ื’ืœื™ื ืœืขืฉื•ืช ื”ื—ืœื˜ื•ืช ื“ื™ ื˜ื•ื‘ื•ืช.
05:47
You have 200 years of statistics to rely on.
121
347960
2216
ื™ืฉ ืœื›ื 200 ืฉื ื” ืฉืœ ืกื˜ื˜ื™ืกื˜ื™ืงื•ืช ืœื”ืกืชืžืš ืขืœื™ื”ืŸ.
05:50
You're amplifying it with very powerful computers.
122
350200
3016
ืืชื ืžื’ื‘ื™ืจื™ื ืื•ืชืŸ ืขื ืžื—ืฉื‘ื™ื ืžืžืฉ ื—ื–ืงื™ื.
05:53
The least you could expect is good TV, right?
123
353240
3280
ื”ื›ื™ ืžืขื˜ ืฉืืชื ื™ื›ื•ืœื™ื ืœืฆืคื•ืช ืœื• ื–ื” ื˜ืœื•ื™ื–ื™ื” ื˜ื•ื‘ื”, ื ื›ื•ืŸ?
05:57
And if data analysis does not work that way,
124
357880
2720
ื•ืื ื ื™ืชื•ื— ืžื™ื“ืข ืœื ืขื•ื‘ื“ ื›ืš,
06:01
then it actually gets a little scary,
125
361520
2056
ืื– ื–ื” ืœืžืขืฉื” ื ืขืฉื” ืžืขื˜ ืžืคื—ื™ื“,
06:03
because we live in a time where we're turning to data more and more
126
363600
3816
ืžืคื ื™ ืฉืื ื—ื ื• ื—ื™ื™ื ื‘ื–ืžืŸ ื‘ื• ืื ื—ื ื• ืคื•ื ื™ื ืœืžื™ื“ืข ื™ื•ืชืจ ื•ื™ื•ืชืจ
06:07
to make very serious decisions that go far beyond TV.
127
367440
4480
ื›ื“ื™ ืœืขืฉื•ืช ื”ื—ืœื˜ื•ืช ืžืžืฉ ื—ืฉื•ื‘ื•ืช ืฉืขื•ื‘ืจื•ืช ืžืขื‘ืจ ืœื˜ืœื•ื™ื–ื™ื”.
06:12
Does anyone here know the company Multi-Health Systems?
128
372760
3240
ื”ืื ืžื™ืฉื”ื• ืคื” ืžื›ื™ืจ ืืช ื”ื—ื‘ืจื” "ืžืขืจื›ื•ืช ืžื•ืœื˜ื™-ื”ืœื˜'"?
06:17
No one. OK, that's good actually.
129
377080
1656
ืืฃ ืื—ื“. ืื•ืงื™ื™, ื–ื” ื˜ื•ื‘ ืœืžืขืฉื”.
06:18
OK, so Multi-Health Systems is a software company,
130
378760
3216
ืื•ืงื™ื™, ืื– "ืžืขืจื›ื•ืช ืžื•ืœื˜ื™-ื”ืœื˜'" ื”ื™ื ื—ื‘ืจืช ืชื•ื›ื ื”,
06:22
and I hope that nobody here in this room
131
382000
2816
ื•ืื ื™ ืžืงื•ื•ื” ืฉืืฃ ืื—ื“ ืคื” ื‘ื—ื“ืจ
06:24
ever comes into contact with that software,
132
384840
3176
ืœืขื•ืœื ืœื ื™ื’ื™ืข ืœืžื’ืข ืขื ื”ืชื•ื›ื ื” ืฉืœื”ื,
06:28
because if you do, it means you're in prison.
133
388040
2096
ืžืคื ื™ ืฉืื ื›ืŸ, ื–ื” ืื•ืžืจ ืฉืืชื ื‘ื›ืœื.
06:30
(Laughter)
134
390160
1176
(ืฆื—ื•ืง)
06:31
If someone here in the US is in prison, and they apply for parole,
135
391360
3536
ืื ืžื™ืฉื”ื• ืคื” ื‘ืืจืฆื•ืช ื”ื‘ืจื™ืช ื‘ื›ืœื, ื•ืื– ื”ื ืžื’ื™ืฉื™ื ื‘ืงืฉื” ืœื—ื ื™ื ื”,
06:34
then it's very likely that data analysis software from that company
136
394920
4296
ืื– ื–ื” ืžืื•ื“ ืกื‘ื™ืจ ืฉืชื•ื›ื ืช ื ื™ืชื•ื— ื”ืžื™ื“ืข ืžื”ื—ื‘ืจื” ื”ื”ื™ื
06:39
will be used in determining whether to grant that parole.
137
399240
3616
ืชื”ื™ื” ื‘ืฉื™ืžื•ืฉ ื‘ืœืงื‘ื•ืข ืื ืœื”ืขื ื™ืง ืืช ื”ื—ื ื™ื ื”.
06:42
So it's the same principle as Amazon and Netflix,
138
402880
2576
ืื– ื–ื” ืื•ืชื• ืขื™ืงืจื•ืŸ ื›ืžื• ืืžื–ื•ืŸ ื•ื ื˜ืคืœื™ืงืก,
06:45
but now instead of deciding whether a TV show is going to be good or bad,
139
405480
4616
ืื‘ืœ ืขื›ืฉื™ื• ื‘ืžืงื•ื ืœื”ื—ืœื™ื˜ ืื ืชื•ื›ื ื™ืช ื˜ืœื•ื•ื™ื–ื™ื” ืชื”ื™ื” ื˜ื•ื‘ื” ืื• ื’ืจื•ืขื”,
06:50
you're deciding whether a person is going to be good or bad.
140
410120
2896
ืืชื ืžื—ืœื™ื˜ื™ื ืื ืื“ื ืขื•ืžื“ ืœื”ื™ื•ืช ื˜ื•ื‘ ืื• ืจืข.
06:53
And mediocre TV, 22 minutes, that can be pretty bad,
141
413040
5496
ื•ื˜ืœื•ื™ื–ื™ื” ื‘ื™ื ื•ื ื™ืช, 22 ื“ืงื•ืช, ืฉื™ื›ื•ืœื” ืœื”ื™ื•ืช ื“ื™ ื’ืจื•ืขื”,
06:58
but more years in prison, I guess, even worse.
142
418560
2640
ืื‘ืœ ื™ื•ืชืจ ืฉื ื™ื ื‘ื›ืœื, ืื ื™ ืžืชืืจ, ืืคื™ืœื• ื’ืจื•ืขื” ื™ื•ืชืจ.
07:02
And unfortunately, there is actually some evidence that this data analysis,
143
422360
4136
ื•ืœืžืจื‘ื” ื”ืฆืขืจ, ื™ืฉ ืœืžืขืฉื” ื›ืžื” ืขื“ื•ื™ื•ืช ืฉื ื™ืชื•ื— ื”ืžื™ื“ืข ื”ื–ื”,
07:06
despite having lots of data, does not always produce optimum results.
144
426520
4216
ืœืžืจื•ืช ืฉื™ืฉ ืœื ื• ื”ืจื‘ื” ืžื™ื“ืข, ืœื ืชืžื™ื“ ืžื™ื™ืฆืจ ืชื•ืฆืื•ืช ืื•ืคื˜ื™ืžืœื™ื•ืช.
07:10
And that's not because a company like Multi-Health Systems
145
430760
2722
ื•ื–ื” ืœื ื‘ื’ืœืœ ืฉื—ื‘ืจื” ื›ืžื• "ืžืขืจื›ื•ืช ืžื•ืœื˜ื™-ื”ืœื˜'"
07:13
doesn't know what to do with data.
146
433506
1627
ืœื ื™ื•ื“ืขืช ืžื” ืœืขืฉื•ืช ืขื ื”ืžื™ื“ืข.
07:15
Even the most data-savvy companies get it wrong.
147
435158
2298
ืืคื™ืœื• ื”ื—ื‘ืจื•ืช ื”ื›ื™ ืชืื‘ื•ืช ืžื™ื“ืข ื˜ื•ืขื•ืช.
07:17
Yes, even Google gets it wrong sometimes.
148
437480
2400
ื›ืŸ, ืืคื™ืœื• ื’ื•ื’ืœ ื˜ื•ืขื” ืœืคืขืžื™ื.
07:20
In 2009, Google announced that they were able, with data analysis,
149
440680
4496
ื‘ 2009, ื’ื•ื’ืœ ื”ื•ื“ื™ืขื” ืฉื”ื ืžืกื•ื’ืœื™ื, ืขื ื ื™ืชื•ื— ืžื™ื“ืข,
07:25
to predict outbreaks of influenza, the nasty kind of flu,
150
445200
4136
ืœื—ื–ื•ืช ื”ืชืคืจืฆื•ื™ื•ืช ืฉืœ ืฉืคืขืช, ื”ืกื•ื’ ื”ื ื‘ื–ื™ ืฉืœ ืฉืคืขืช,
07:29
by doing data analysis on their Google searches.
151
449360
3776
ืขืœ ื™ื“ื™ ื ื™ืชื•ื— ืฉืœ ื—ื™ืคื•ืฉื™ ื’ื•ื’ืœ.
07:33
And it worked beautifully, and it made a big splash in the news,
152
453160
3856
ื•ื–ื” ืขื‘ื“ ื™ืคื”ืคื”, ื•ื–ื” ืขืฉื” ื’ืœ ื’ื“ื•ืœ ื‘ื—ื“ืฉื•ืช,
07:37
including the pinnacle of scientific success:
153
457040
2136
ื›ื•ืœืœ ื”ืคืกื’ื” ืฉืœ ื”ื”ืฆืœื—ื” ื”ืžื“ืขื™ืช:
07:39
a publication in the journal "Nature."
154
459200
2456
ืคืจืกื•ื ื‘ื™ืจื—ื•ืŸ "ื ื™ื™ืฆ'ืจ."
07:41
It worked beautifully for year after year after year,
155
461680
3616
ื–ื” ืขื‘ื“ ืžื•ืฉืœื ื‘ืžืฉืš ืฉื ื” ืื—ืจื™ ืฉื ื” ืื—ืจื™ ืฉื ื”,
07:45
until one year it failed.
156
465320
1656
ืขื“ ืฉืฉื ื” ืื—ืช ื–ื” ื ื›ืฉืœ.
07:47
And nobody could even tell exactly why.
157
467000
2256
ื•ืืฃ ืื—ื“ ืœื ื”ื™ื” ื™ื›ื•ืœ ืืคื™ืœื• ืœื”ืกื‘ื™ืจ ืœืžื” ื‘ื“ื™ื•ืง.
07:49
It just didn't work that year,
158
469280
1696
ื–ื” ืคืฉื•ื˜ ืœื ืขื‘ื“ ื‘ืื•ืชื” ืฉื ื”,
07:51
and of course that again made big news,
159
471000
1936
ื•ื›ืžื•ื‘ืŸ ืฉื–ื” ื”ื™ื” ืฉื•ื‘ ื—ื“ืฉื•ืช ื’ื“ื•ืœื•ืช,
07:52
including now a retraction
160
472960
1616
ื›ื•ืœืœ ื‘ื™ื˜ื•ืœ ืขื›ืฉื™ื•
07:54
of a publication from the journal "Nature."
161
474600
2840
ืฉืœ ื”ืคืจืกื•ื ื‘ืžื’ื–ื™ืŸ "ื ื™ื™ืฆ'ืจ."
07:58
So even the most data-savvy companies, Amazon and Google,
162
478480
3336
ืื– ืืคื™ืœื• ื”ื—ื‘ืจื•ืช ืฉื”ื›ื™ ืชืื‘ื•ืช ืžื™ื“ืข, ืืžื–ื•ืŸ ื•ื’ื•ื’ืœ,
08:01
they sometimes get it wrong.
163
481840
2136
ืœืคืขืžื™ื ื˜ื•ืขื•ืช.
08:04
And despite all those failures,
164
484000
2936
ื•ืœืžืจื•ืช ื”ื›ืฉืœื•ื ื•ืช ื”ืืœื•,
08:06
data is moving rapidly into real-life decision-making --
165
486960
3856
ืžื™ื“ืข ื ืข ื‘ืžื”ื™ืจื•ืช ืœื”ื—ืœื˜ื•ืช ื‘ื—ื™ื™ื ื”ืืžื™ืชื™ื™ื --
08:10
into the workplace,
166
490840
1816
ืœืชื•ืš ืžืงื•ื ื”ืขื‘ื•ื“ื”,
08:12
law enforcement,
167
492680
1816
ืœืื›ื™ืคืช ื”ื—ื•ืง,
08:14
medicine.
168
494520
1200
ืœืจืคื•ืื”.
08:16
So we should better make sure that data is helping.
169
496400
3336
ืื– ืื ื—ื ื• ืฆืจื™ื›ื™ื ืœื“ืื•ื’ ื™ื•ืชืจ ืฉื”ืžื™ื“ืข ืขื•ื–ืจ.
08:19
Now, personally I've seen a lot of this struggle with data myself,
170
499760
3136
ืขื›ืฉื™ื•, ืื™ืฉื™ืช, ืื ื™ ืจืื™ืชื™ ื”ืจื‘ื” ืžื”ืžืื‘ืง ื”ื–ื” ื‘ืžื™ื“ืข ื‘ืขืฆืžื™,
08:22
because I work in computational genetics,
171
502920
1976
ืžืคื ื™ ืฉืื ื™ ืขื•ื‘ื“ ื‘ื’ื ื˜ื™ืงื” ื—ื™ืฉื•ื‘ื™ืช,
08:24
which is also a field where lots of very smart people
172
504920
2496
ืฉื”ื™ื ื’ื ืชื—ื•ื ื‘ื• ื”ืจื‘ื” ืื ืฉื™ื ืžืื•ื“ ื—ื›ืžื™ื
08:27
are using unimaginable amounts of data to make pretty serious decisions
173
507440
3656
ืžืฉืชืžืฉื™ื ื‘ื›ืžื•ื™ื•ืช ืฉืœื ื ื™ืชืŸ ืœืชืืจ ืฉืœ ืžื™ื“ืข ื›ื“ื™ ืœืขืฉื•ืช ื”ื—ืœื˜ื•ืช ื“ื™ ืจืฆื™ื ื™ื•ืช
08:31
like deciding on a cancer therapy or developing a drug.
174
511120
3560
ื›ืžื• ืœื”ื—ืœื™ื˜ ืขืœ ื˜ื™ืคื•ืœ ื‘ืกืจื˜ืŸ ืื• ืคื™ืชื•ื— ืชืจื•ืคื”.
08:35
And over the years, I've noticed a sort of pattern
175
515520
2376
ื•ื‘ืžื”ืœืš ื”ืฉื ื™ื, ื”ื‘ื—ื ืชื™ ื‘ืกื•ื’ ืฉืœ ืชื‘ื ื™ืช
08:37
or kind of rule, if you will, about the difference
176
517920
2456
ืื• ืกื•ื’ ืฉืœ ื—ื•ืง, ืื ืชืจืฆื•, ื‘ื ื•ื’ืข ืœื”ื‘ื“ืœ
08:40
between successful decision-making with data
177
520400
2696
ื‘ื™ืŸ ืงื‘ืœืช ื”ื—ืœื˜ื•ืช ืžื•ืฆืœื—ืช ืขื ืžื™ื“ืข
08:43
and unsuccessful decision-making,
178
523120
1616
ื•ืงื‘ืœืช ื”ื—ืœื˜ื•ืช ืœื ืžื•ืฆืœื—ืช.
08:44
and I find this a pattern worth sharing, and it goes something like this.
179
524760
3880
ื•ืื ื™ ืžื•ืฆื ืฉื–ื• ืชื‘ื ื™ืช ืฉืฉื•ื•ื” ืœื—ืœื•ืง, ื•ื”ื™ื ื”ื•ืœื›ืช ืžืฉื”ื• ื›ื–ื”.
08:50
So whenever you're solving a complex problem,
180
530520
2135
ืื– ื›ืœ ืคืขื ืฉืืชื ืคื•ืชืจื™ื ื‘ืขื™ื” ืกื‘ื•ื›ื”,
08:52
you're doing essentially two things.
181
532679
1737
ืืชื ืขื•ืฉื™ื ื‘ืขื™ืงืจื•ืŸ ืฉื ื™ ื“ื‘ืจื™ื.
08:54
The first one is, you take that problem apart into its bits and pieces
182
534440
3296
ื”ืจืืฉื•ืŸ ื”ื•ื, ืืชื ืžืคืจืงื™ื ืืช ื”ื‘ืขื™ื” ืœื—ืœืงื™ื ื”ืงื˜ื ื™ื ืฉืœื”
08:57
so that you can deeply analyze those bits and pieces,
183
537760
2496
ื›ืš ืฉืชื•ื›ืœื• ืœื ืชื— ืืช ื”ื—ืœืงื™ื ื”ืืœื”,
09:00
and then of course you do the second part.
184
540280
2016
ื•ืื– ื›ืžื•ื‘ืŸ ืืชื ืขื•ืฉื™ื ืืช ื”ื—ืœืง ื”ืฉื ื™.
09:02
You put all of these bits and pieces back together again
185
542320
2656
ืืชื ืžื—ื‘ืจื™ื ื—ื–ืจื” ืืช ื›ืœ ื”ื—ืœืงื™ื
09:05
to come to your conclusion.
186
545000
1336
ื›ื“ื™ ืœื”ื’ื™ืข ืœืžืกืงื ื” ืฉืœื›ื.
09:06
And sometimes you have to do it over again,
187
546360
2336
ื•ืœืคืขืžื™ื ืืชื ืฆืจื™ื›ื™ื ืœืขืฉื•ืช ืืช ื–ื” ืฉื•ื‘,
09:08
but it's always those two things:
188
548720
1656
ืื‘ืœ ื–ื” ืชืžื™ื“ ืฉื ื™ ื”ื“ื‘ืจื™ื ื”ืืœื”:
09:10
taking apart and putting back together again.
189
550400
2320
ืœืคืจืง ื•ืœื”ืจื›ื™ื‘ ื‘ื—ื–ืจื”.
09:14
And now the crucial thing is
190
554280
1616
ื•ืขื›ืฉื™ื• ื”ื“ื‘ืจ ื”ืงืจื™ื˜ื™ ื”ื•ื
09:15
that data and data analysis
191
555920
2896
ืฉื”ืžื™ื“ืข ื•ื ื™ืชื•ื— ื”ืžื™ื“ืข
09:18
is only good for the first part.
192
558840
2496
ื”ื•ื ื˜ื•ื‘ ืจืง ืœื—ืœืง ื”ืจืืฉื•ืŸ.
09:21
Data and data analysis, no matter how powerful,
193
561360
2216
ืžื™ื“ืข ื•ื ื™ืชื•ื— ืžื™ื“ืข, ืœื ืžืฉื ื” ื›ืžื” ื—ื–ืงื™ื,
09:23
can only help you taking a problem apart and understanding its pieces.
194
563600
4456
ื™ื›ื•ืœื ืจืง ืœืขื–ื•ืจ ืœื›ื ืœืคืจืง ืืช ื”ื‘ืขื™ื” ื•ืœื”ื‘ื™ืŸ ืืช ื”ืคื™ืกื•ืช.
09:28
It's not suited to put those pieces back together again
195
568080
3496
ื–ื” ืœื ืžืชืื™ื ืœื—ื™ื‘ื•ืจ ื”ืคื™ืกื•ืช ื”ืืœื• ื—ื–ืจื” ื™ื—ื“
09:31
and then to come to a conclusion.
196
571600
1896
ื•ืื– ืœื”ื’ื™ืข ืœืžืกืงื ื”.
09:33
There's another tool that can do that, and we all have it,
197
573520
2736
ื™ืฉ ื›ืœื™ ื ื•ืกืฃ ืฉื™ื›ื•ืœ ืœืขืฉื•ืช ืืช ื–ื”, ื•ืœื›ื•ืœื ื• ื™ืฉ ืื•ืชื•,
09:36
and that tool is the brain.
198
576280
1296
ื•ื”ื›ืœื™ ื”ื–ื” ื”ื•ื ื”ืžื•ื— ืฉืœื ื•.
09:37
If there's one thing a brain is good at,
199
577600
1936
ืื ื™ืฉ ืžืฉื”ื• ืื—ื“ ืฉื”ืžื•ื— ื˜ื•ื‘ ื‘ื•,
09:39
it's taking bits and pieces back together again,
200
579560
2256
ื–ื” ืœืงื—ืช ืคื™ืกื•ืช ืžื™ื“ืข ื•ืœื—ื‘ืจ ืื•ืชืŸ ื™ื—ื“,
09:41
even when you have incomplete information,
201
581840
2016
ืืคื™ืœื• ื›ืฉื™ืฉ ืœื›ื ืคื™ืกื•ืช ืžื™ื“ืข ืœื ืฉืœืžื•ืช,
09:43
and coming to a good conclusion,
202
583880
1576
ื•ืœื”ื’ื™ืข ืœืžืกืงื ื”,
09:45
especially if it's the brain of an expert.
203
585480
2936
ื‘ืขื™ืงืจ ืื ื–ื” ื”ืžื•ื— ืฉืœ ืžื•ืžื—ื”.
09:48
And that's why I believe that Netflix was so successful,
204
588440
2656
ื•ืœื›ืŸ ืื ื™ ืžืืžื™ืŸ ืฉื ื˜ืคืœื™ืงืก ื”ื™ื• ื›ืœ ื›ืš ืžืฆืœื™ื—ื™ื,
09:51
because they used data and brains where they belong in the process.
205
591120
3576
ื‘ื’ืœืœ ืฉื”ื ื”ืฉืชืžืฉื• ื‘ืžื™ื“ืข ื•ื‘ืžื•ื— ื‘ืžืงื•ื ื‘ื• ื”ื ื”ื™ื• ืฉื™ื™ื›ื™ื ืœืชื”ืœื™ืš.
09:54
They use data to first understand lots of pieces about their audience
206
594720
3536
ื”ื ื”ืฉืชืžืฉื• ื‘ืžื™ื“ืข ืจืืฉื™ืช ื›ื“ื™ ืœื”ื‘ื™ืŸ ื”ืจื‘ื” ื—ืœืงื™ื ื‘ื ื•ื’ืข ืœืงื”ืœ ืฉืœื”ื
09:58
that they otherwise wouldn't have been able to understand at that depth,
207
598280
3416
ืฉืื—ืจืช ื”ื ืœื ื”ื™ื• ืžืกื•ื’ืœื™ื ืœื”ื‘ื™ืŸ ื‘ืขื•ืžืง ื”ื–ื”,
10:01
but then the decision to take all these bits and pieces
208
601720
2616
ืื‘ืœ ืื– ื”ื”ื—ืœื˜ื” ืœืงื—ืช ืืช ื›ืœ ื”ืคื™ืกื•ืช ื”ืืœื•
10:04
and put them back together again and make a show like "House of Cards,"
209
604360
3336
ื•ืœื—ื‘ืจ ืื•ืชืŸ ื™ื—ื“ ืฉื•ื‘ ื•ืœืขืฉื•ืช ืชื•ื›ื ื™ืช ื›ืžื• "ื‘ื™ืช ื”ืงืœืคื™ื,"
10:07
that was nowhere in the data.
210
607720
1416
ื–ื” ืœื ื”ื™ื” ื‘ืฉื•ื ืžืงื•ื ื‘ืžื™ื“ืข.
10:09
Ted Sarandos and his team made that decision to license that show,
211
609160
3976
ื˜ื“ ืกืจื ื“ื•ืก ื•ื”ืฆื•ื•ืช ืฉืœื• ืขืฉื• ืืช ื”ื”ื—ืœื˜ื” ื”ื–ื• ื›ื“ื™ ืœืคืชื— ืืช ื”ืชื•ื›ื ื™ืช,
10:13
which also meant, by the way, that they were taking
212
613160
2381
ืžื” ืฉื’ื ืืžืจ, ื“ืจืš ืื’ื‘, ืฉื”ื ืœืงื—ื•
10:15
a pretty big personal risk with that decision.
213
615565
2851
ืกื™ื›ื•ืŸ ืื™ืฉื™ ื“ื™ ื’ื“ื•ืœ ืขื ื”ื”ื—ืœื˜ื” ื”ื–ื•.
10:18
And Amazon, on the other hand, they did it the wrong way around.
214
618440
3016
ื•ืืžื–ื•ืŸ, ืžืฆื“ ืฉื ื™, ื”ื ืขืฉื• ืืช ื–ื” ื‘ื“ืจืš ื”ื”ืคื•ื›ื”.
10:21
They used data all the way to drive their decision-making,
215
621480
2736
ื”ื ื”ืฉืชืžืฉื• ื‘ืžื™ื“ืข ื›ืœ ื”ื“ืจืš ื›ื“ื™ ืœื”ื ื™ืข ืืช ืชื”ืœื™ืš ื”ื”ืœื—ื˜ื•ืช ืฉืœื”ื,
10:24
first when they held their competition of TV ideas,
216
624240
2416
ืจืืฉื™ืช ื”ื ืขืฉื• ืืช ื”ืชื—ืจื•ืช ืฉืœื”ื ืฉืœ ืจืขื™ื•ื ื•ืช ืœืชื•ื›ื ื™ื•ืช,
10:26
then when they selected "Alpha House" to make as a show.
217
626680
3696
ืื– ื›ืฉื”ื ื‘ื—ืจื• ืืช "ื‘ื™ืช ืืœืคื" ื›ื“ื™ ืœื™ืฆื•ืจ ื›ืชื•ื›ื ื™ืช.
10:30
Which of course was a very safe decision for them,
218
630400
2496
ืฉื›ืžื•ื‘ืŸ ื”ื™ืชื” ื”ื—ืœื˜ื” ืžืื•ื“ ื‘ื˜ื•ื—ื” ื‘ืฉื‘ื™ืœื,
10:32
because they could always point at the data, saying,
219
632920
2456
ืžืคื ื™ ืฉื”ื ืชืžื™ื“ ื™ื›ืœื• ืœื”ืฆื‘ื™ืข ืขืœ ื”ืžื™ื“ืข, ื•ืœื”ื’ื™ื“.
10:35
"This is what the data tells us."
220
635400
1696
"ื–ื” ืžื” ืฉื”ืžื™ื“ืข ืื•ืžืจ ืœื ื•."
10:37
But it didn't lead to the exceptional results that they were hoping for.
221
637120
4240
ืื‘ืœ ื–ื” ืœื ื”ื•ื‘ื™ืœ ืœืชื•ืฆืื” ื™ื•ืฆืืช ื“ื•ืคืŸ ืฉื”ื ืงื™ื•ื• ืœื”.
10:42
So data is of course a massively useful tool to make better decisions,
222
642120
4976
ืื– ืžื™ื“ืข ื”ื•ื ื›ืžื•ื‘ืŸ ื›ืœื™ ืžืื•ื“ ืžื•ืขื™ืœ ืœืขืฉื•ืช ื”ื—ืœื˜ื•ืช ื˜ื•ื‘ื•ืช ื™ื•ืชืจ,
10:47
but I believe that things go wrong
223
647120
2376
ืื‘ืœ ืื ื™ ืžืืžื™ืŸ ืฉื“ื‘ืจื™ื ืžืฉืชื‘ืฉื™ื
10:49
when data is starting to drive those decisions.
224
649520
2576
ื›ืฉืžื™ื“ืข ืžืชื—ื™ืœ ืœื”ื ื™ืข ืืช ื”ื”ื—ืœื˜ื•ืช ื”ืืœื•.
10:52
No matter how powerful, data is just a tool,
225
652120
3776
ืœื ืžืฉื ื” ื›ืžื” ื—ื–ืง, ืžื™ื“ืข ื”ื•ื ืจืง ื›ืœื™,
10:55
and to keep that in mind, I find this device here quite useful.
226
655920
3336
ื•ื›ื“ื™ ืœื–ื›ื•ืจ ืืช ื–ื”, ืื ื™ ืžื•ืฆื ืืช ื”ืžื›ืฉื™ืจ ื”ื–ื” ืคื” ืžืžืฉ ืžื•ืขื™ืœ.
10:59
Many of you will ...
227
659280
1216
ืจื‘ื™ื ืžื›ื...
11:00
(Laughter)
228
660520
1216
(ืฆื—ื•ืง)
11:01
Before there was data,
229
661760
1216
ืœืคื ื™ ืฉื”ื™ื” ืžื™ื“ืข,
11:03
this was the decision-making device to use.
230
663000
2856
ื–ื” ื”ื™ื” ืžื›ืฉื™ืจ ืงื‘ืœืช ื”ื”ื—ืœื˜ื•ืช ื‘ื• ื”ืฉืชืžืฉื•.
11:05
(Laughter)
231
665880
1256
(ืฆื—ื•ืง)
11:07
Many of you will know this.
232
667160
1336
ื”ืจื‘ื” ืžื›ื ื™ื“ืขื• ืืช ื–ื”.
11:08
This toy here is called the Magic 8 Ball,
233
668520
1953
ื”ืžื›ืฉื™ืจ ื”ื–ื” ื ืงืจื ื›ื“ื•ืจ 8 ืงืกื•ื,
11:10
and it's really amazing,
234
670497
1199
ื•ื”ื•ื ื‘ืืžืช ืžื“ื”ื™ื,
11:11
because if you have a decision to make, a yes or no question,
235
671720
2896
ืžืคื ื™ ืฉืื ื™ืฉ ืœื›ื ื”ื—ืœื˜ื” ืœืขืฉื•ืช, ืฉืืœื” ืฉืœ ื›ืŸ ื•ืœื,
11:14
all you have to do is you shake the ball, and then you get an answer --
236
674640
3736
ื›ืœ ืžื” ืฉืืชื ืฆืจื™ื›ื™ื ืœืขืฉื•ืช ื–ื” ืœื ืขืจ ืืช ื”ื›ื“ื•ืจ, ื•ืื– ืืชื ืžืงื‘ืœื™ื ืชืฉื•ื‘ื” --
11:18
"Most Likely" -- right here in this window in real time.
237
678400
2816
"ื”ืกื™ื›ื•ื™ื™ื ื’ื“ื•ืœื™ื" -- ืžืžืฉ ืคื” ื‘ื—ืœื•ืŸ ื”ื–ื” ื‘ื–ืžืŸ ืืžืช.
11:21
I'll have it out later for tech demos.
238
681240
2096
ืื ื™ ืื•ืฆื™ื ืื•ืชื• ืื—ืจ ื›ืš ืœื”ื“ื’ืžื•ืช.
11:23
(Laughter)
239
683360
1216
(ืฆื—ื•ืง)
11:24
Now, the thing is, of course -- so I've made some decisions in my life
240
684600
3576
ืขื›ืฉื™ื•, ื”ืขื ื™ื™ืŸ ื”ื•ื, ื›ืžื•ื‘ืŸ -- ืื– ืขืฉื™ืชื™ ื›ืžื” ื”ื—ืœื˜ื•ืช ื‘ื—ื™ื™
11:28
where, in hindsight, I should have just listened to the ball.
241
688200
2896
ื‘ื”ืŸ, ื‘ืžื—ืฉื‘ื” ืื—ื•ืจื”, ื”ื™ื™ืชื™ ืฆืจื™ืš ืคืฉื•ื˜ ืœื”ืงืฉื™ื‘ ืœื›ื“ื•ืจ.
11:31
But, you know, of course, if you have the data available,
242
691120
3336
ืื‘ืœ, ืืชื ื™ื•ื“ืขื™ื, ื›ืžื•ื‘ืŸ, ืื ื™ืฉ ืœื›ื ืืช ื”ืžื™ื“ืข ื–ืžื™ืŸ,
11:34
you want to replace this with something much more sophisticated,
243
694480
3056
ืืชื ืจื•ืฆื™ื ืœื”ื—ืœื™ืฃ ืืช ื–ื” ืขื ืžืฉื”ื• ื”ืจื‘ื” ื™ื•ืชืจ ืžืชื•ื—ื›ื,
11:37
like data analysis to come to a better decision.
244
697560
3616
ื›ืžื• ื ื™ืชื•ื— ืžื™ื“ืข ืœื”ื’ื™ืข ืœื”ื—ืœื˜ื” ื˜ื•ื‘ื” ื™ื•ืชืจ.
11:41
But that does not change the basic setup.
245
701200
2616
ืื‘ืœ ื–ื” ืœื ืžืฉื ื” ืืช ื”ื—ืœืง ื”ื‘ืกื™ืกื™.
11:43
So the ball may get smarter and smarter and smarter,
246
703840
3176
ืื– ื”ื›ื“ื•ืจ ืื•ืœื™ ื™ื”ื™ื” ื—ื›ื ื™ื•ืชืจ ื•ื™ื•ืชืจ ื•ื™ื•ืชืจ,
11:47
but I believe it's still on us to make the decisions
247
707040
2816
ืื‘ืœ ืื ื™ ืžืืžื™ืŸ ืฉื–ื” ืขื“ื™ื™ืŸ ืชืœื•ื™ ื‘ื ื• ืœืขืฉื•ืช ืืช ื”ื”ื—ืœื˜ื•ืช
11:49
if we want to achieve something extraordinary,
248
709880
3016
ืื ืื ื—ื ื• ืจื•ืฆื™ื ืœื”ืฉื™ื’ ืžืฉื”ื• ื™ื•ืฆื ื“ื•ืคืŸ,
11:52
on the right end of the curve.
249
712920
1936
ื‘ืฆื“ ื”ื ื›ื•ืŸ ืฉืœ ื”ืขืงื•ืžื”.
11:54
And I find that a very encouraging message, in fact,
250
714880
4496
ื•ืื ื™ ืจื•ืื” ืืช ื–ื” ืžืื•ื“ ืžืขื•ื“ื“, ืœืžืขืฉื”,
11:59
that even in the face of huge amounts of data,
251
719400
3976
ืฉืืคื™ืœื• ืžื•ืœ ื›ืžื•ื™ื•ืช ืขืฆื•ืžื•ืช ืฉืœ ืžื™ื“ืข,
12:03
it still pays off to make decisions,
252
723400
4096
ื–ื” ืขื“ื™ื™ืŸ ืžืฉืชืœื ืœืขืฉื•ืช ื”ื—ืœื˜ื•ืช,
12:07
to be an expert in what you're doing
253
727520
2656
ืœื”ื™ื•ืช ืžื•ืžื—ื™ื ื‘ืžื” ืฉืืชื ืขื•ืฉื™ื
12:10
and take risks.
254
730200
2096
ื•ืœืงื—ืช ืกื™ื›ื•ื ื™ื.
12:12
Because in the end, it's not data,
255
732320
2776
ืžืคื ื™ ืฉื‘ืกื•ืคื• ืฉืœ ื“ื‘ืจ, ื–ื” ืœื ืžื™ื“ืข,
12:15
it's risks that will land you on the right end of the curve.
256
735120
3960
ื–ื” ืกื™ื›ื•ื ื™ื ืฉื™ื‘ื™ืื• ืืชื›ื ืœืฆื“ ื”ื ื›ื•ืŸ ืฉืœ ื”ืขืงื•ืžื”.
12:19
Thank you.
257
739840
1216
ืชื•ื“ื” ืœื›ื.
12:21
(Applause)
258
741080
3680
(ืžื—ื™ืื•ืช ื›ืคื™ื™ื)
ืขืœ ืืชืจ ื–ื”

ืืชืจ ื–ื” ื™ืฆื™ื’ ื‘ืคื ื™ื›ื ืกืจื˜ื•ื ื™ YouTube ื”ืžื•ืขื™ืœื™ื ืœืœื™ืžื•ื“ ืื ื’ืœื™ืช. ืชื•ื›ืœื• ืœืจืื•ืช ืฉื™ืขื•ืจื™ ืื ื’ืœื™ืช ื”ืžื•ืขื‘ืจื™ื ืขืœ ื™ื“ื™ ืžื•ืจื™ื ืžื”ืฉื•ืจื” ื”ืจืืฉื•ื ื” ืžืจื—ื‘ื™ ื”ืขื•ืœื. ืœื—ืฅ ืคืขืžื™ื™ื ืขืœ ื”ื›ืชื•ื‘ื™ื•ืช ื‘ืื ื’ืœื™ืช ื”ืžื•ืฆื’ื•ืช ื‘ื›ืœ ื“ืฃ ื•ื™ื“ืื• ื›ื“ื™ ืœื”ืคืขื™ืœ ืืช ื”ืกืจื˜ื•ืŸ ืžืฉื. ื”ื›ืชื•ื‘ื™ื•ืช ื’ื•ืœืœื•ืช ื‘ืกื ื›ืจื•ืŸ ืขื ื”ืคืขืœืช ื”ื•ื•ื™ื“ืื•. ืื ื™ืฉ ืœืš ื”ืขืจื•ืช ืื• ื‘ืงืฉื•ืช, ืื ื ืฆื•ืจ ืื™ืชื ื• ืงืฉืจ ื‘ืืžืฆืขื•ืช ื˜ื•ืคืก ื™ืฆื™ืจืช ืงืฉืจ ื–ื”.

https://forms.gle/WvT1wiN1qDtmnspy7