digital living room

1

ADVERTISING ATTENTION IN THE WILD – A COMPARISON OF ONLINE AND TELEVISED VIDEO ADVERTISING

Created in partnership with

YuMe Online Video NetworkBy

IPG Media LabApril 2011

Advertising Attention In The Wild

A Comparison of online and Televised Video advertising

Created in partnership with

YuMe By IPG Media Lab

April 2011

2

Questions we set out to answer

1. How much more ad avoidance happens beyond active ad skipping?

2. What is the relative attention level to video advertising in a lean forward PC experience vs. a lean back TV experience?

3. What behaviors most distract attention to video ads?

3

Methodology•March 2011•Los Angeles•Recreated normal viewing choices•Respondents brought companion media•30 minutes in office/30 minutes in living room•Post survey on ad recall

4

Sample: N=48Gender Employment Status

Household Income

Female 48% Full-time56%

$100,000-$200,000 13%

Male 52% Part-time31% $75,000-$100,000 19%

Retired 6% $50,000-$75,000 33%Age Student 4% $25,000-$50,000 25%18-24 15% Unemployed 2% Less than $25,000 10%

25-29 15%

30-34 10% EducationChildren <18 in Household

35-39 10% High school/GED 8% No 77.08%

40-44 15% Some college27% Yes 22.92%

45-49 13% Associate's degree 6%

50-55 10% Bachelor's degree48%

56-60 6% Master's degree 6%

65-69 6% Doctorate degree 2%

Trade or other technical school degree 2%

• Recruited from LA metro area• Must watch online video

5

Attention scores explained

1 to 0.9Full attention

0.9 and 0.4Partial attention

0.4 to -1 No attention

Frame by frame, second by second.

6

Scale of TV ad Fast Forwarding

35% US DVR HH penetration

10% of DVR HH viewing time shifted

x 65% of ads skipped in time shifted viewing

2% of total TV impressions skipped

Source: Magna Global

7

Smart phones are the most common distraction media

OL Mobile Phone - Call

OL Other

OL Read Book/Magazine

OL Do Work

OL IM/Chat/Email

No OL Distractions

OL Mobile Phone - Data

0% 10% 20% 30% 40% 50%

6.3%

8.3%

10.4%

12.5%

16.7%

27.1%

45.8%

Online: % of Sample Using Distraction

TV Play Game

No TV Distractions

TV Mobile Phone - Call

TV Other

TV Do Work

TV Read Book/Magazine

TV Use Laptop

TV DVR

TV Mobile Phone - Data

0% 10%20%30%40%50%60%70%

4.2%

6.0%

8.3%

8.3%

12.5%

12.5%

33.3%

45.8%

60.4%

TV: % of Sample Using Distraction

8

Persona 1: Cathy the Ad-Ignorer

9

Persona 2: Michie the Multi-tasker

10

Persona 3: Steve the Vegged-Out Relaxer

11

Finding #1: Not all distractions are equal

No OL Distractions

OL IM/Chat/Email

OL Mobile Phone - Call

OL Mobile Phone - Data

OL Other

OL Do Work

OL Read Book/Magazine

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

0.60

0.48

0.47

0.47

0.38

0.34

0.13

Online Ad Attention Level

TV Play Game

TV DVR

TV Use Laptop

TV Do Work

TV Mobile Phone - Data

No TV Distractions

TV Read Book/Magazine

TV Mobile Phone - Call

TV Other

0 0.2 0.4 0.6 0.8

0.54

0.52

0.52

0.47

0.46

0.44

0.43

0.33

0.26

TV Ad Attention LevelWorst

Best

12

Finding #1 (cont.) : The more distractions, the lower ad attention

0 1 2 3

-1.00

-0.80

-0.60

-0.40

-0.20

0.00

0.20

0.40

0.60

0.80

1.00

0.44 0.53

0.44 0.37

0.60

0.45 0.40

Ad Attention vs. # of DistractionsTV Ad Attention OnlineVideo Ad Attention

Count of Distraction Media During Viewing Session

Avera

ge A

ttenti

on S

core

13

Finding #2: TV 2x video clutter; Ubiquitous banners

OL TV

Video 5.5 9.5

Banner/ Bug 21.6 0.7

Total 27.1 10.3

14

Finding #3: Online video content +8.5% more attention

% Full Attention During Content Time0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

60.1%51.6%

OL TV

% o

f S

eco

nds R

eci

evin

g F

ull

Att

enti

on

15

Finding #4: TV has 3x drop in attention from content to ad

OL TV0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

60.1%51.6%55.2%

36.9%

% Full Attention During Content Time

% Full Attention During Video Ad Time

% o

f S

eco

nd

s R

eci

evin

g F

ull

At-

ten

tio

n

Decrease in Attention From Program to Ad

OL = ∆ - 4.8% TV = ∆ - 14.7%

16

Finding #5: Online video ads +18.3% more attention than TV

• 63% of TV impressions were ignored.• DVR fast forwarding is estimated to lead to 2% ad skipping

% Full Attention During Video Ad Time0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

55.2%

36.9%

OL TV

% o

f Seconds R

ecie

vin

g F

ull A

ttenti

on

17

Finding #6: Attention is correlated with recall

Online TV

-1.00

-0.80

-0.60

-0.40

-0.20

0.00

0.20

0.40

0.60

0.80

1.00

0.30

0.44

0.61

0.28

0.60

0.44

0.64

0.49

Unremembered Ads

Correctly Recalled Ads, Aided

Correctly Recalled Ads, Unaided

Average Attention

DVR fast-forwarding arti-ficially increased unre-membered ad attention score

18

Finding #7: Online ads have 1.8x the aided recall and 1.5x the unaided recall

Aided Recall is statistically significant at 90% level of confidence

Aided Unaided0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

28% 25%

50%

38%

% of Sample Who Correctly Identified the Brand in a Video Ad Seen

TV Online

19

Finding #8: Gender attention is even, Women more likely to recall video ads

Average of On-lineVideo Ad At-

tention

Average of TV Ad Attention

-1.00

-0.80

-0.60

-0.40

-0.20

0.00

0.20

0.40

0.60

0.80

1.00

0.44 0.48 0.51 0.48

Ad Attention by GenderFemale Male

TV Aided TV Unaided OL Aided OL Unaided0%

10%

20%

30%

40%

50%

60%

19%

35%

56%

43%42%

16%

42%

30%

Ad Recall by Gender

Female Male

20

Finding #9: Ad attention drops off with time on screen

0 15 30 45 60 75 90 105

120

135

150

165

180

195

210

225

240

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

TV

Loga-rithmic (TV)OL

Loga-rithmic (OL)

Length of Video Ad Exposure in Seconds

Avera

ge A

ttenti

on L

evel W

hile W

atc

hin

g

Ad

21

Finding #10: Ad Fast-Fowarders have high attention levels…

% of time paying attention while an ad is on sc...0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

47%

35%

% of Ad Time Paying Full Attention to Screen

DVR FF No DVR

22

Finding #10 (cont.) : Fast-Fowarders have low recall levels

DVR FF No DVR0%

5%

10%

15%

20%

25%

30%

35%

40%

45%

50%

18%

29%

20%

32%

Unaided Recall Aided Recall

23

Finding #11: Attention is1.4x higher for TV “bugs” than video ads

Total % Full Attention % Full Attention During Video Ad

% Full Attention During "Other" Ads

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

60%55%

62%

49%

37%

50%

OL TV

24

Conclusions1. Ad fast forwarding accounts for a sliver of wasted

ad impressions

2. Smart phones are a persistent companion to video content

3. Online video ads have 20% more attentive impressions.

4. The familiar cadence of TV content increases drop off to ads vs. online

5. Attention is even but women more likely to recall video ads than men

6. Fast forwarded video ads have little recall

7. The commercial “layer” gets more attention than the commercial break.

25

THANK YOU!Travis@yume.com

Brian.Monahan@ipglab.com

26

How we made sense of it:While respondents were relaxing alone in our living room and office environments, we had three cameras with simultaneous, matched time-stamps recording:

1. Their face• This video was sent to Affectiva, which used a facial tracking

algorithm to determine on a second-by-second level basis whether the person was paying attention to the screen

2. The environment• This video was used to understand what distraction media

was being used and when

3. Their screen• This video allowed us to record against the time-line

– When an ad was on-screen– What brand the ad was for– When an individual fast-forwarded an ad– What an individual was viewing when spikes in

bio-metrics occurred

Additionally, a subset of respondents were given a bio-metric monitoring bracelet which told us when increases in cognition, excitement, and stress occurred.

All of these various data streams were then matched to a single time-line, which allowed the data to be aggregated, analyzed, and mined for meaning.

27

THE EFFECTS OF DISTRACTION

28

ATTENTION AND ADVERTISING

29

How to read a bio-metric chart

Source: Affectiva

EDA measures moments of intellectual engagement or relaxation on the green line

A “2” blue line means a video ad is on screen, and when it dips to 1.5 that means it’s being fast forwarded

Attention (the red line) is measured on a scale of 1 to -1. Anything over 0.9 is considered full attention on the screen

Any rapid changes in body temperature indicate stress or emotional engagement (purple Celsius line)

The time-line is used to sync all of these inputs to recorded video

30

Cathy the Ad-Ignorer: OnlineSimilar to TV, her attention is split between magazines or her phone screen and the computer screen. She consistently tunes out commercials with distraction media. Though nervous at first, she quickly calms down.

31

Michie the Multi-tasker: TVThough exposed twice to the same lengthy ad, her computer and phone are constant distractions as Michie flips nonchalantly from screen to screen. Moments of high engagement occur while she talks with friends, and the TV serves only as a pacifying distraction.

32

Steve the Vegged-Out Relaxer: OnlineIn place of a DVR, Steve switches rapidly between types of online video. He shows spikes of engagement watching footage of the Japanese earthquake aftermath, but quickly settles in to passively watch a long-form show, during which he maintains interest in ad breaks and correctly recalls Campbells soup as a sponsor.