06:30
TED2010

Sebastian Wernicke: Lies, damned lies and statistics (about TEDTalks)

Filmed:

In a brilliantly tongue-in-cheek analysis, Sebastian Wernicke turns the tools of statistical analysis on TEDTalks, to come up with a metric for creating "the optimum TEDTalk" based on user ratings. How do you rate it? "Jaw-dropping"? "Unconvincing"? Or just plain "Funny"?

- Data scientist
After making a splash in the field of bioinformatics, Sebastian Wernicke moved on to the corporate sphere, where he motivates and manages multidimensional projects. Full bio

If you go on the TED website,
00:16
you can currently find there
00:19
over a full week of TEDTalk videos,
00:21
over 1.3 million
00:24
words of transcripts
00:26
and millions of user ratings.
00:28
And that's a huge amount of data.
00:30
And it got me wondering:
00:33
If you took all this data
00:35
and put it through statistical analysis,
00:37
could you reverse engineer a TEDTalk?
00:39
Could you create
00:41
the ultimate TEDTalk?
00:43
(Laughter) (Applause)
00:45
And also, could you create
00:47
the worst possible TEDTalk
00:49
that they would still let you get away with?
00:51
To find this out, I looked at three things:
00:53
I looked at the topic that you should choose,
00:55
I looked at how you should deliver it
00:57
and the visuals onstage.
01:00
Now, with the topic: There's a whole range of topics you can choose,
01:02
but you should choose wisely,
01:05
because your topic strongly correlates
01:07
with how users will react to your talk.
01:09
Now, to make this more concrete,
01:12
let's look at the list of top 10 words
01:14
that statistically stick out
01:17
in the most favorite TEDTalks
01:19
and in the least favorite TEDTalks.
01:21
So if you came here
01:24
to talk about how French coffee
01:26
will spread happiness in our brains,
01:28
that's a go.
01:31
(Laughter) (Applause)
01:33
Whereas, if you wanted to talk about
01:35
your project involving
01:37
oxygen, girls, aircraft --
01:39
actually, I would like to hear that talk, (Laughter)
01:41
but statistics say it's not so good.
01:43
Oh, well.
01:45
If you generalize this,
01:47
the most favorite TEDTalks are those
01:49
that feature topics we can connect with,
01:51
both easily and deeply,
01:54
such as happiness, our own body,
01:56
food, emotions.
01:58
And the more technical topics,
02:00
such as architecture, materials and, strangely enough, men,
02:02
those are not good topics to talk about.
02:05
How should you deliver your talk?
02:08
TED is famous for keeping
02:10
a very sharp eye on the clock,
02:12
so they're going to hate me
02:14
for revealing this, because, actually,
02:16
you should talk as long as they will let you. (Laughter)
02:18
Because the most favorite TEDTalks
02:20
are, on average, over 50 percent longer
02:23
than the least favorite ones.
02:25
And this holds true for all ranking lists on TED.com
02:28
except if you want to have a talk
02:30
that's beautiful, inspiring or funny.
02:32
Then, you should be brief. (Laughter) But other than that,
02:34
talk until they drag you off the stage.
02:36
(Laughter)
02:39
Now, while ...
02:41
(Applause)
02:43
While you're pushing the clock, there's a few rules to obey.
02:49
I found these rules out by comparing the statistics
02:52
of four-word phrases
02:54
that appear more often in the most favorite TEDTalks
02:56
as opposed to the least favorite TEDTalks.
02:59
I'll give you three examples.
03:01
First of all, I must, as a speaker,
03:03
provide a service to the audience and talk about what I will give you,
03:05
instead of saying what I can't have.
03:08
Secondly, it's imperative
03:10
that you do not cite The New York Times.
03:12
(Laughter)
03:14
And finally, it's okay for the speaker -- that's the good news --
03:16
to fake intellectual capacity.
03:19
If I don't understand something, I can just say, "etc., etc."
03:21
You'll all stay with me.
03:24
It's perfectly fine.
03:26
(Applause)
03:28
Now, let's go to the visuals.
03:32
The most obvious visual thing on stage is the speaker.
03:34
And analysis shows if you want to be
03:37
among the most favorite TED speakers,
03:39
you should let your hair grow a little bit longer than average,
03:41
make sure you wear your glasses and be slightly more dressed-up
03:44
than the average TED speaker.
03:47
Slides are okay, though you might consider going for props.
03:49
And now the most important thing,
03:52
that is the mood onstage.
03:54
Color plays a very important role.
03:56
Color closely correlates
03:58
with the ratings that talks get on the website.
04:00
(Applause)
04:03
For example, fascinating talks
04:05
contain a statistically high amount
04:07
of exactly this blue color, (Laughter)
04:09
much more than the average TEDTalk.
04:11
Ingenious TEDTalks, much more this green color,
04:13
etc., et.
04:15
(Laughter) (Applause)
04:17
Now, personally, I think
04:24
I'm not the first one who has done this analysis,
04:26
but I'll leave this
04:28
to your good judgment.
04:30
So, now it's time to put it all together
04:33
and design the ultimate TEDTalk.
04:35
Now, since this is TEDActive,
04:37
and I learned from my analysis
04:39
that I should actually give you something,
04:41
I will not impose the ultimate
04:43
or worst TEDTalk on you,
04:45
but rather give you a tool to create your own.
04:47
And I call this tool the TEDPad.
04:49
(Laughter)
04:52
And the TEDPad is a matrix
04:55
of 100 specifically selected,
04:57
highly curated sentences
04:59
that you can easily piece together to get your own TEDTalk.
05:02
You only have to make one decision,
05:07
and that is: Are you going to use the white version
05:09
for very good TEDTalks,
05:11
about creativity, human genius?
05:13
Or are you going to go with a black version,
05:15
which will allow you to create really bad TEDTalks,
05:17
mostly about blogs,
05:19
politics and stuff?
05:21
So, download it and have fun with it.
05:23
Now I hope you enjoy the session.
05:25
I hope you enjoy designing your own
05:28
ultimate and worst possible TEDTalks.
05:30
And I hope some of you will be inspired for next year
05:32
to create this, which I really want to see.
05:35
Thank you very much.
05:38
(Applause) Thanks.
05:40

▲Back to top

About the Speaker:

Sebastian Wernicke - Data scientist
After making a splash in the field of bioinformatics, Sebastian Wernicke moved on to the corporate sphere, where he motivates and manages multidimensional projects.

Why you should listen

Dr. Sebastian Wernicke heads the data science department at Solon, a Munich-based consultancy supporting companies and investors in media, entertainment, telecoms, and technology industries. Wernicke originally studied bioinformatics and previously led the strategy and growth of Seven Bridges Genomics, a Cambridge-based startup that builds platforms for genetic analysis.

Before his career in statistics began, Wernicke worked stints as both a paramedic and successful short animated filmmaker. He's also the author of the TEDPad app, an irreverent tool for creating an infinite number of "amazing and really bad" and mostly completely meaningless talks. He's the author of the statistically authoritative and yet completely ridiculous "How to Give the Perfect TEDTalk."

More profile about the speaker
Sebastian Wernicke | Speaker | TED.com