The Impact of The Internet On Advertising Markets For New Media

Will the Internet Destroy the
News Media?
Susan Athey, Emilio Calvano & Joshua Gans
May 2011
Wednesday, 18 May 2011

The chart of doom
Newspaper Ad Revenue (1980 $m)
Source: Newspaper Associa0on of America

The chart of doom

$30,000
$22,500
$15,000
$7,500
$0
1960 1963 1966 1969 1972 1975 1978 1981 1984 1987 1990 1993 1996 1999 2002 2005 2008

The chart of doom

$30,000
$22,500
$15,000
$7,500
$0
1960 1963 1966 1969 1972 1975 1978 1981 1984 1987 1990 1993 1996 1999 2002 2005 2008

The chart of doom

$30,000
$22,500
$15,000
$7,500
$0
1960 1963 1966 1969 1972 1975 1978 1981 1984 1987 1990 1993 1996 1999 2002 2005 2008

Explaining the decline

Psychology (digital less effective)



Economic fundamentals

Attention is still scarce ...

Attention is still scarce ...
and advertisers still want to access that

attention.

Disruption

Disruption
The Internet has disrupted the traditional

business model ...

Disruption

business model ...
need to wait for experiments to sort it
out.

Disruption

business model ...
out.
Most experiments rely on the ‘waterbed
effect’ ...

Disruption

business model ...
out.
Most experiments rely on the ‘waterbed
effect’ ...
but based on assumption readers choose
content and advertisers sell access.

Only two facts

Only two facts
The Internet has facilitated consumer

switching between outlets

Only two facts
The Internet has facilitated consumer

switching between outlets
There is imperfect tracking between

outlets

Puzzles

Puzzles
Evidence that competition reduces ad prices

Puzzles
Outlets claim mergers will improve ad revenue

Puzzles
For-profit outlets object to lifting of ad

restrictions on public broadcasters

Puzzles
For-profit outlets object to lifting of ad

restrictions on public broadcasters
Larger outlets earn higher ad revenue per

consumer

Switching

Switching
Browsing

Switching
Browsing
Free content

Switching
Browsing
Free content
Aggregators, social
networks and
search

Switching
Browsing
Free content
Aggregators, social
networks and
search

Tracking

Tracking
Switching alone does not explain the

decline.

Tracking
Switching alone does not explain the

decline.
With no tracking or perfect tracking do

not care where consumers are.

Advertising supply

Advertising supply
Two attention periods.

Advertising supply
Two outlets with ad capacity per unit of attention, ai.

Advertising supply
If they have opportunity to choose, consumers select outlet i with

probability xi.

Advertising supply

probability xi.
In a given period, the probability that a consumer can choose is ρ.

Advertising supply

probability xi.
Dil = xi − xi (1 − xi )ρ
D s = 2 ρ x1 x2

Advertising supply

probability xi.
Dil = xi − xi (1 − xi )ρ
D s = 2 ρ x1 x2
i's ad inventory = Dil 2ai + D s ai

Advertiser demand

Advertiser demand
Advertisers want to impress each consumer once over

the two periods …

Advertiser demand
Advertisers want to impress each consumer once over

the two periods …
and have heterogeneous values (v) on impressing

consumers with distribution F(v); assumed to be U[0,1].

The impression game
Morning Afternoon
Outlet 1
Outlet 2

The impression game
Morning Afternoon
Outlet 1
Outlet 2

The impression game
Morning Afternoon
Outlet 1
Outlet 2

The impression game
Morning Afternoon
Outlet 1
Outlet 2

The impression game
Morning Afternoon
Outlet 1
Outlet 2

The impression game
Morning Afternoon
Outlet 1
Outlet 2

The impression game
Morning Afternoon
Outlet 1
Outlet 2

The impression game
Morning Afternoon
Outlet 1
Outlet 2

The impression game
Morning Afternoon
Outlet 1
Outlet 2
If Starbucks single-homes, it misses impressions.

If Starbucks multi-homes, it wastes impressions.

The advertiser’s dilemma
Custom analysis of data provided to authors by

ComScore of 30 recent large, cross-outlet campaigns

The advertiser’s dilemma
Wasted
Impressions
Custom analysis of data provided to authors by

ComScore of 30 recent large, cross-outlet campaigns

Solving the dilemma

Solving the dilemma
No switching

Solving the dilemma
No switching
No tracking

Solving the dilemma
No switching
No tracking
Coordination in time

Solving the dilemma
No switching
No tracking
Pay per click

Solving the dilemma
No switching
No tracking
Pay per click
Perfect tracking

Missed & wasted impressions
Expected Unique Impressions

Single-Home on i

Single-Home on i θ i = Dil + 12 D s

Multi-Home

Multi-Home θ12 = D1l + D2l + 43 D s

Multi-Home (2 on i)

Multi-Home (2 on i) θ12′ = 1

Multi-Home (2 on i) θ12′ = 1
θ12 ≤ θ1 + θ 2 = θ12′ for ρ > 0

No switching
Price
Multi-homers
(1 impression)
Quantity (Impressions)

No switching
Price Supply
Multi-homers
(1 impression)
2a

No switching
Price Supply
v=1
Multi-homers
(1 impression)
2a

No switching
Price Supply
v=1
Multi-homers
(1 impression)
v=p
2a

No switching
Price Supply
v=1
Multi-homers
(1 impression)
v=p
2a

No switching
Price Supply
v=1
Multi-homers
(1 impression) p
v=p
2a

More switchers
Price Supply
v=1
Multi-homers
(1 impression)
Single- Single-
home on 1 home on 2
v=p
2a

More switchers
Price Supply
v=1
Multi-homers
(1 impression)
p
Single- Single-
home on 1 home on 2
v=p
2a

More switchers
Price Supply
v=1
↑ D s ⇒↓ p
Multi-homers
(1 impression)
p
Single- Single-
home on 1 home on 2
v=p
2a

High switchers
Price Supply
v=1
MH 2 on 1 MH 2 on 2
Multi-homers
(1 impression)
p
Single- Single-
home on 1 home on 2
v=p
2a

High switchers
Price Supply
v=1
MH 2 on 1 MH 2 on 2
Multi-homers
(1 impression)
p
Single- Single-
home on 1 home on 2
v=p
2a

High switchers
Price Supply
v=1
MH 2 on 1 MH 2 on 2
↑ D s ⇒↑ p
Multi-homers
(1 impression)
p
Single- Single-
home on 1 home on 2
v=p
2a

High switchers
Price Supply
v=1
MH 2 on 1 MH 2 on 2
↑ D s ⇒↑ p
Multi-homers if a high
(1 impression)
p
Single- Single-
home on 1 home on 2
v=p
2a

Incentives to adopt perfect tracking
Low ad capacity (a)

Profits
Ds

Low ad capacity (a)

Profits
Imperfect
Tracking
Ds

Low ad capacity (a)

Profits
Perfect
Tracking
Imperfect
Tracking
Ds

High ad capacity (a) Imperfect

Profits Tracking
Perfect
Tracking
Ds


Profits Tracking
Perfect
Tracking
Ds


Profits Tracking
Perfect
Tracking
Ds

Public broadcasters & blogs
Price
Supply

Price
Supply
Blogs and other non-ad
content …

Price
Supply
content …
decrease available ad
capacity in the market …

Price
Supply
content …

Price
Supply
content …
and reduce adverse effect

of switching …

Price
Supply
content …

of switching …

Price
Supply
content …

of switching …
Causing impression prices

to rise.

Price
Supply
content …

of switching …
Causing impression prices

to rise.

Positional advantages
in advertising markets

Content quality

Content quality Paywalls

Content quality Paywalls Magnet content

Asymmetric Quality
x1 = x2
Low Ds High Ds
v=1 v=1
MH 2 on 1 MH 2 on 2
Multi-homers
(1 impression)
Multi-homers
(1 impression)
Single- Single-
Single- Single-
home on 1 home on 2
home on 1 home on 2
v=p v=p

Asymmetric Quality
x1 > x2
Low Ds High Ds
v=1 v=1
MH 2 on 1 MH 2 on 2
Multi-homers
(1 impression)
Multi-homers
(1 impression)
Single- Single-
Single- Single-
home on 1 home on 2
home on 1 home on 2
v=p v=p

Asymmetric Quality
x1 > x2
Low Ds High Ds
v=1 v=1
MH 2 on 2
Multi-homers
(1 impression)
Multi-homers
(1 impression)
Single-home on 1
v = p1 Single-home on 1
Single-home on 2 v = p1
Single-home on 2
v = p2 v = p2

Asymmetric Quality
x1 > x2
Low Ds High Ds
v=1 v=1
MH 2 on 2
Multi-homers
(1 impression)
Multi-homers
(1 impression)
Single-home on 1
v = p1 Single-home on 1
Single-home on 2
v = p2 v = p2
Outlet 1 gains a positional advantage

in the advertising market
Paywalls

Paywalls
“No [it’s not a two way street with Google sending

traffic] What’s the point of someone coming
occasionally who likes a headline they see on Google?
Sure we go out and say we have so many millions of
visitors. The fact is that there is not enough advertising
in the world to go around to make all the websites
profitable. We’d rather have fewer people coming to
our websites but paying. They don’t suddenly become
loyal readers of our websites.” (Rupert Murdoch)

Paywalls

Paywalls
D = x1 − x1 x2 ρ12
l
1

Paywalls
D = x1 − x1 x2 ρ12
l
1
D2l = x2 − x2 x1ρ21

Paywalls
D = x1 − x1 x2 ρ12
l
1
D2l = x2 − x2 x1ρ21
D s = ( ρ12 + ρ21 )x1 x2

Paywalls
D = x1 − x1 x2 ρ12
l
1
D2l = x2 − x2 x1ρ21
D s = ( ρ12 + ρ21 )x1 x2
Micropayments

Paywalls
D = x1 − x1 x2 ρ12
l
1
D2l = x2 − x2 x1ρ21
D s = ( ρ12 + ρ21 )x1 x2
Micropayments ↑ ρ12 ,↓ ρ21

Paywalls
D = x1 − x1 x2 ρ12
l
1
D2l = x2 − x2 x1ρ21
D s = ( ρ12 + ρ21 )x1 x2
Subscriptions

Paywalls
D = x1 − x1 x2 ρ12
l
1
D2l = x2 − x2 x1ρ21
D s = ( ρ12 + ρ21 )x1 x2
Subscriptions ↓ ρ21

Paywalls
D = x1 − x1 x2 ρ12
l
1
D2l = x2 − x2 x1ρ21
D s = ( ρ12 + ρ21 )x1 x2
Limited paywall

Paywalls
D = x1 − x1 x2 ρ12
l
1
D2l = x2 − x2 x1ρ21
D s = ( ρ12 + ρ21 )x1 x2
Limited paywall ↑ ρ12

Paywalls
D = x1 − x1 x2 ρ12
l
1
D2l = x2 − x2 x1ρ21
D s = ( ρ12 + ρ21 )x1 x2
Subscriptions ↓ ρ21 D <D

l
1
l
2

Paywalls
D = x1 − x1 x2 ρ12
l
1
D2l = x2 − x2 x1ρ21
D s = ( ρ12 + ρ21 )x1 x2
Subscriptions ↓ ρ21 D <D

l
1
l
2
Paywalls may cause a positional disadvantage in advertising markets

One more thing ...

Magnet Content
Single-home on 1
Multi-Home 2 on 2 (2 impressions)
Single-home on 2
Single-home on 1
(2 impressions)
Single-home on 2
(2 impressions) Single-home on 1

Magnet Content
Suppose that Outlet 2 hasSingle-home on 1

Multi-Home 2 on 2limited content (2 impressions)
Single-home on 2
Single-home on 1
(2 impressions)
Single-home on 2

Magnet Content

... and so can only serveSingle-home

Single-home on 1 one on 2
(2 impressions)
period
Single-home on 2
of attention.

Magnet Content
ρ =1


(2 impressions)
period
Single-home on 2
of attention.

D1l = x1 − x1 x2 Magnet Content
ρ =1


(2 impressions)
period
Single-home on 2
of attention.

ρ =1 D2l = 0


(2 impressions)
period
Single-home on 2
of attention.

ρ =1 D2l = 0
D s = x2 (1 + x1 )


(2 impressions)
period
Single-home on 2
of attention.

ρ =1 D2l = 0
D s = x2 (1 + x1 )
Low x2
Single-home on 1
Single-home on 2
Single-home on 1
(2 impressions)
Single-home on 2

ρ =1 D2l = 0
D s = x2 (1 + x1 )
Low x2
v=1
Single-home on 1
Single-home on 2
Single-home on 1
(2 impressions)
Single-home on 2

ρ =1 D2l = 0
D s = x2 (1 + x1 )
Low x2
v=1
Single-home on 1
Single-home on 2
Single-home on 1
(2 impressions)
Single-home on 2

ρ =1 D2l = 0
D s = x2 (1 + x1 )
Low x2
v=1
Single-home on 1
Single-home on 2
Single-home on 1
(2 impressions)
Single-home on 2

ρ =1 D2l = 0
D s = x2 (1 + x1 )
Low x2
v=1
Single-home on 1
Single-home on 2
Single-home on 1
(2 impressions)
Single-home on 2

ρ =1 D2l = 0
D s = x2 (1 + x1 )
Low x2
v=1
Single-home on 1
Single-home on 2
Single-home on 1
(2 impressions)
v = p1
Single-home on 2

ρ =1 D2l = 0
D s = x2 (1 + x1 )
Low x2
v=1
Single-home on 1
Single-home on 2
Single-home on 1
(2 impressions)
v = p1
Single-home on 2
v = p2

ρ =1 D2l = 0
D s = x2 (1 + x1 )
Low x2 High x2
v=1
Single-home on 1
Single-home on 2
Single-home on 1
(2 impressions)
v = p1
Single-home on 2
v = p2

ρ =1 D2l = 0
D s = x2 (1 + x1 )
Low x2 High x2
v=1 v=1
Single-home on 1
Single-home on 2
Single-home on 1
(2 impressions)
v = p1
Single-home on 2
v = p2

ρ =1 D2l = 0
D s = x2 (1 + x1 )
Low x2 High x2
v=1 v=1
Single-home on 1
Single-home on 2
Single-home on 1
(2 impressions)
v = p1
Single-home on 2
v = p2

ρ =1 D2l = 0
D s = x2 (1 + x1 )
Low x2 High x2
v=1 v=1
Single-home on 1
Single-home on 2
Single-home on 1
(2 impressions)
v = p1
Single-home on 2
v = p2

ρ =1 D2l = 0
D s = x2 (1 + x1 )
Low x2 High x2
v=1 v=1
Single-home on 1
Single-home on 2
Single-home on 1
(2 impressions)
v = p1
Single-home on 2
v = p2

ρ =1 D2l = 0
D s = x2 (1 + x1 )
Low x2 High x2
v=1 v=1
Single-home on 1
Single-home on 2
Single-home on 1
(2 impressions)
v = p1
v = p2

ρ =1 D2l = 0
D s = x2 (1 + x1 )
Low x2 High x2
v=1 v=1
Single-home on 1
Single-home on 2
Single-home on 1
(2 impressions)
v = p1
v = p2 v = p1

Magnet Content
Profits
0.20
Full content provider
0.15
0.10
Imperfect Tracking
0.05 Limited content provider
0.2 0.4 0.6 0.8 1.0

x2

Magnet Content
Profits
0.20
0.15
0.10
Imperfect Tracking
0.2 0.4 0.6 0.8 1.0

x2
As limited content provider’s share rises, industry profits rise.

Magnet Content
Profits
0.20
0.15
0.10
Imperfect Tracking
0.2 0.4 0.6 0.8 1.0

x2
Twice the return to magnet content than full content.

Magnet Content
Profits
0.20
0.15
0.10
Imperfect Tracking
0.2 0.4 0.6 0.8 1.0

x2
Higher return per reader due to positional advantage.

Magnet Content
Profits
0.20
0.15
0.10
Imperfect Tracking
0.2 0.4 0.6 0.8 1.0

x2
Evolution away from frequency and towards reach.

Profits
0.20
0.15
0.10
Imperfect Tracking
0.2 0.4 0.6 0.8 1.0

x2

150 million US accounts
Profits
0.20
0.15
0.10
Imperfect Tracking
0.2 0.4 0.6 0.8 1.0

x2

95% users visit once every 2 days

Profits
0.20
0.15
0.10
Imperfect Tracking
0.2 0.4 0.6 0.8 1.0

x2


Profits
10% of 0.20
US site visits
0.15
0.10
Imperfect Tracking
0.2 0.4 0.6 0.8 1.0

x2


Profits
10% of 0.20
US site visits
30% of 0.15
US display ads
0.10
Imperfect Tracking
0.2 0.4 0.6 0.8 1.0

x2

Conclusions

Conclusions
The efficient operation of advertising markets are dependent

upon consumer behavior.

Conclusions

A combination of consumer switching and imperfect tracking

can explain competitive effects in advertising markets for
news media.

Conclusions


news media.
Tracking technologies may assist in efficiency but adoption

faces impediments.

Conclusions


news media.

faces impediments.
In addition to deterring consumers, paywalls may cause outlets

to have a positional disadvantage in advertising markets.

Conclusions


news media.

faces impediments.
In addition to deterring consumers, paywalls may cause outlets

to have a positional disadvantage in advertising markets.
Incentives to provide content move from full line provision to

magnet but limited content provision.

The Impact of The Internet On Advertising Markets For New Media

Încărcat de

Informații document

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

The Impact of The Internet On Advertising Markets For New Media

Încărcat de

Drepturi de autor:

Formate disponibile

Will the Internet Destroy the

Wednesday, 18 May 2011

Newspaper Ad Revenue (1980 $m)

Source: Newspaper Associa0on of America

Wednesday, 18 May 2011

Newspaper Ad Revenue (1980 $m)

Source: Newspaper Associa0on of America

Wednesday, 18 May 2011

Newspaper Ad Revenue (1980 $m)

Source: Newspaper Associa0on of America

Wednesday, 18 May 2011

Newspaper Ad Revenue (1980 $m)

Source: Newspaper Associa0on of America

Wednesday, 18 May 2011

Wednesday, 18 May 2011

Psychology (digital less effective)

Wednesday, 18 May 2011

Psychology (digital less effective)

Wednesday, 18 May 2011

Psychology (digital less effective)

Wednesday, 18 May 2011

Wednesday, 18 May 2011

Attention is still scarce ...

Wednesday, 18 May 2011

Attention is still scarce ...

and advertisers still want to access that

Wednesday, 18 May 2011

Wednesday, 18 May 2011

The Internet has disrupted the traditional

Wednesday, 18 May 2011

The Internet has disrupted the traditional

Wednesday, 18 May 2011

The Internet has disrupted the traditional

Wednesday, 18 May 2011

The Internet has disrupted the traditional

Wednesday, 18 May 2011

Wednesday, 18 May 2011

The Internet has facilitated consumer

Wednesday, 18 May 2011

The Internet has facilitated consumer

There is imperfect tracking between

Wednesday, 18 May 2011

Wednesday, 18 May 2011

Evidence that competition reduces ad prices

Wednesday, 18 May 2011

Evidence that competition reduces ad prices

Outlets claim mergers will improve ad revenue

Wednesday, 18 May 2011

Evidence that competition reduces ad prices

Outlets claim mergers will improve ad revenue

For-profit outlets object to lifting of ad

Wednesday, 18 May 2011

Evidence that competition reduces ad prices

Outlets claim mergers will improve ad revenue

For-profit outlets object to lifting of ad

Larger outlets earn higher ad revenue per

Wednesday, 18 May 2011

Wednesday, 18 May 2011

Wednesday, 18 May 2011

Wednesday, 18 May 2011

Wednesday, 18 May 2011

Wednesday, 18 May 2011

Wednesday, 18 May 2011