Webcast Prepared for Members of The Advertising Research Foundation

Prepared by: Joel Rubinson, Chief Research Officer, The ARF

January 14, 2009

Joelrubinson@gmail.com

Follow me on twitter @joelrubinson

Empirical Generalizations Regarding the Effectiveness of TV Advertising Over Time

Can TV still work? A prima facie case against…

• DVR household penetration in 2008 has reached 25 - 30% (Eggerton, 2008; Goetzel, 2008; Steinberg, 2008) and 50% of DVR owners typically fast forward (Morrissey, 2008).

• Increase in TV commercial clutter. Over the last 50 years, “non-program content” in a 60 minute prime time show has increased from 11 to 18 minutes (Papazian, 2007).

• Shift in media consumption patterns that have led 50%+ of TV viewers to multi-task, presumably paying less attention to TV as they are simultaneously on the web, using their phones, reading, etc. (MRI, 2008; Papper, Holmes, Popovich, & Bloxham, 2005).

Can TV still work? It’s even getting a little ugly…

• A sense that there is something wrong about push marketing that attempts to “interrupt” viewers and intrude

• “What would you rather have, 5 million impressions or 10 quality relationships?” ANA conference speaker

• “What is the future of a business where the consumer hates your product — in this case, advertising?”  Greg Stuart, former head of the IAB, in AdWeek

Can TV still work? Some empirical results in favor…

• Evidence was published in 2007 by professors Hu, Lodish, and Krieger from IRI Behaviorscan testing using experimental design that the average level of effectiveness of TV advertising, using test vs. control methods, has actually INCREASED.

• Evidence that people are in heightened states of awareness while fast forwarding and still absorb messages (especially those they have seen before).

Anecdote: NYU marketing class

» I asked students to write down their favorite brand and then to comment on how they became “introduced” to that brand

» I called on 5 students and no one said that TV was the way they became aware of the brand or even had much influence on them

» One student said her favorite brand was Apple…here is the conversation that ensued…

» Can people fully retrieve such information?

The key questions

» Has the effectiveness of TV impressions changed over time and if so, has it increased or decreased?

» Is TV a reasonable option in the media mix or has a decline in effectiveness made it an undesirable way to spend your money?

ARF conducted a meta-analysis of seven US databases, 388 cases

» The ARF wanted to factually inform this issue via a meta-analysis…let’s harvest what the industry has learned!

» To integrate the results across databases, we used an indexing approach

• Choose a “dependent variable” for each database (some measure of impact vs. weight on an “impressions” rather than dollar basis)

• Average all the cases across time in a given database on this dependent measure and call that an index of 100

• Analyze the index of cases over time (trending up or down) and/or compare the index of effectiveness across different media platforms

The databases that the ARF had access to

» Sales lift analysis databases

• Represent a broad range of CPG product categories– IRI – (n = 125) controlled and matched market testing where the level

of spending was increased substantially from a non-0 level– PM Group (n =37) and Dratfield Analytics (n = 27)—marketing Mix

modeling– ARS® – (n = 112) modeling wearout factors, and GRP advertising

pressure vs. the sales impact associated with advertising.

» Cross Media comparison databases

• Represent a broad range of products and services (not limited to CPG)– Marketing Evolution (n = 40) and Dynamic Logic (n = 47) –

respondent level, post-hoc experimental design methods– PointLogic/Compose – Media planning tool for analyzing the relative

impact of various advertising platforms

TV Advertising does NOT appear to be declining in effectiveness, and perhaps is even increasing

TV Advertising does NOT appear to be declining in effectiveness, and perhaps is even increasing

figure 1IRI Behaviorscan Data 1990-2002

0

20

40

60

80

100

120

140

160

90-94 95-99 00-02

Year

Ind

ex Index of Elasticity

TV Advertising does NOT appear to be declining in effectiveness, and perhaps is even increasing

figure 2PM Group Data 2003-2008

0.00

20.00

40.00

60.00

80.00

100.00

120.00

2003-2005 2006-2008

Year

Ind

ex o

f T

V E

ffecti

ven

ess

Average

TV Advertising does NOT appear to be declining in effectiveness, and perhaps is even increasing

Dratfield Analytics

0

20

40

60

80

100

120

140

2005 2006 2007 2008

Year

TV

eff

ecti

ven

ess

Effectiveness index

TV Advertising does NOT appear to be declining in effectiveness, and perhaps is even increasing

•Even when controlling for copy quality and wearout, TV advertising effectiveness appears to be increasing.

Figure 4ARS Persuasion Points Delivered vs. Incremental Sales

(From Marketing Mix Modeling) 1997-2005

0

20

40

60

80

100

120

140

1997-1999 2000-2002 2003-2005

Inde

x Index of Inremental Days of Sales/ Persuasion Points DeliveredIndex of Incremental Days of Sales

/ Persuasion Points Delivered

TV Advertising does NOT appear to be declining in effectiveness, and perhaps is even increasing

figure 1IRI Behaviorscan Data 1990-2002

0

20

40

60

80

100

120

140

160

90-94 95-99 00-02

Year

Ind

ex Index of Elasticity

figure 2PM Group Data 2003-2008

0.00

20.00

40.00

60.00

80.00

100.00

120.00

2003-2005 2006-2008

Year

Ind

ex o

f T

V E

ffecti

ven

ess

Average

Dratfield Analytics

0

20

40

60

80

100

120

140

2005 2006 2007 2008

Year

TV

eff

ecti

ven

ess

Effectiveness index

Figure 4ARS Persuasion Points Delivered vs. Incremental Sales

(From Marketing Mix Modeling) 1997-2005

0

20

40

60

80

100

120

140

1997-1999 2000-2002 2003-2005

Inde

x Index of Inremental Days of Sales/ Persuasion Points Delivered

Index of Incremental Days of Sales

/ Persuasion Points Delivered

TV can be an effective platform

TV can be an effective platform

Figure 5Marketing Evolution Data 2004 -2007

0

20

40

60

80

100

120

140

160

180

Online Print TV

Media

Ind

ex o

f P

eop

le I

mp

acte

d p

er $

1k

Familarity

Purchase Intent

Awareness

TV can be an effective platform

Figure 7aDynamic Logic 2004-2008

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

No Yes

Opportunity to See

Aid

ed B

rand

Aw

aren

ess

tv

online

print

Figure 7bDynamic Logic 2004-2008

0%

10%

20%

30%

40%

50%

No Yes

Opportunity to See

Pos

itive

Pur

chas

e In

tent

tv

online

print

TV can be an effective platformF ig ure 6a

C ompos e T ouc hpoint R anking s 2005-2008

0

5

10

15

20

25

30

35

tou

ch

po

int

ran

k

C ompos e 200506

C ompos e 2007

C ompos e 2008

F ig ure 6bC ompos e 2005-2008 Awarenes s R anking s

0

5

10

15

20

25

30

35

Television

Home s hopping televis

ion

P roduct place

ment (TV/m

ovie)

Movie thea tre

R adioNewspapers

Maga zines

P roduct article

s

F ree cus tomer m

agaz ines

Loose in

s erts

Interne t advertisin

g

Interne t search

C ompany webs ite

s

E mails

Outdoor

A irport

P ublic tr

ansporta

tion

Doctor's

office

800 numbers

Direct

ma il

C e lebrity endors ements

S ports s ponso

rship

E vent s

pons ors hip

In-s tore advertising

E xamining in

-store

P rice promotions

P roduct sa

mples

P ers ona l reco

mmendation

P rofes sional re

commendation

To

uch

po

int

Ran

kin

g

C ompos e 200506

C ompos e 2007

C ompos e 2008

Conclusions

» TV advertising does not appear to be declining in effectiveness, and perhaps is even increasing

» TV can be an effective platform

» TV appears to work best at generating brand awareness

» Marketers must be alert to the dynamic nature of TV as long-form video becomes more available via other platforms and as TV becomes interactive and targetable