Google’s Unskippable Labs has run over 250 video ad experiments. Learn their approach

Google’s Unskippable Labs has run over 250 video ad experiments. Learn their approach

Published
March 2019

“When it comes to video ad creative, what works?” That’s the question Ben Jones, global creative director at Google’s Unskippable Labs, unpacks day in and day out.

“There is a lot of muscle memory when it comes to making ads. Some practices are backed in data, and some are gut instinct, bluster, or opinion,” says Jones. And that’s where Unskippable Labs finds space to experiment. “We find hypotheses we don’t have a definite answer to, and we say, ‘We don’t know if this works; you don’t know if this works. Let’s design an experiment to test it.’”

Unskippable Labs partners with brands to review about 1,000 video ads each month, and has run over 250 experiments globally to put some analytical rigor behind those gut instincts. Watch Jones explain his team’s four-pronged approach to video ad experimentation and find out how your brand can start to experiment too.

0:01

[MUSIC PLAYING]

0:02

BEN JONES: I am a relentlessly curious person,

0:06

and so what inspires me is everything,

0:08

the way that words and images come together

0:11

to move us in all of its forms.

0:12

I'm Ben Jones.

0:13

I'm a creative director, and I work

0:16

on a team called Unskippable Labs at Google.

0:18

We run experiments in video content.

0:19

We look at about a thousand ads a month

0:22

to say what are patterns of creative effectiveness

0:24

that advertisers can take advantage of.

0:27

When we partner with brands, all our experiments run in-market.

0:29

We don't do panels or focus groups or anything.

0:31

Everything is live.

0:32

Then we get the data and see what

0:34

conclusions we can come to.

0:37

Based on the experimentation, we see four primary opportunities

0:42

in storytelling-- visual language, narrative structure,

0:45

tuning for audience, and creative systems.

0:48

We see the opportunity to evolve the visual language that's

0:49

used in ads, and some of these are

0:52

very simple-- tighter framing, faster pacing,

0:56

use of large-type supers, brightness and contrast.

0:59

The second opportunity we see is around narrative structure,

1:00

and it's building for attention.

1:03

You need to give them, we say, dessert first,

1:06

or a punch in the face, a thing in the beginning that is like,

1:08

ah, there's a reason I should pay attention to this.

1:10

The third opportunity is tuning for audience.

1:12

Making creative choices to reflect who you are

1:15

and what you're interested in improves creative response

1:16

significantly.

1:18

The last is how are you sequencing ads

1:21

and what are the right mixtures of formats and forms,

1:24

and how do you best deploy creative and media

1:27

assets for maximum impact.

1:31

Google is in this space because if clients' videos are better,

1:33

then the whole platform is better.

1:34

But I think especially within Google,

1:36

there is this idea that with a big enough data

1:39

set we'll find enough patterns and we'll

1:42

unlock the ability to figure out how every ad should work.

1:43

The challenge with advertising is

1:46

as soon as you make an ad that follows all the rules,

1:47

all the rules change.

1:49

So can machine learning help us?

1:50

Absolutely.

1:53

Bringing some data in can help creatives defend good ideas

1:54

and make better work.

1:57

It can help clients put more effective work in market.

1:58

Then they can take more risks.

2:01

I think data is a tool like many tools,

2:03

and there is a lot that it can teach us

2:07

and a lot that it can show us, but it works with creativity,

2:09

not to replace creativity.

2:11

And so the only way I think to really

2:15

be breakthrough as a brand is to be experimenting all the

2:22

time.

What publishers need to know now about creating a better ad experience