Who Ate All Our Cookies? Investigating Publishers’ Challenges

Caused by Changes in Third-party Cookie Tracking

Valerio Stallone

, Aline Gägauf and Tania Kaya

Institute of Marketing Management, ZHAW School of Management and Law, Winterthur, Switzerland

Keywords: Third-party Cookies, First-party Cookies, Digital Advertising.

Abstract: This paper investigates the potential reactions of Swiss publishers, as actors with-in the digital advertising

ecosystem, to the forthcoming fundamental changes to user tracking in the world wide web. The results of

this mixed methodical study initiate the discussion on the future of cookie tracking by setting and then

answering to four hypotheses regarding first-party tracking, shared ID solutions, Google’s Privacy Sandbox,

and a national walled garden system. The results show a clear inclination of Swiss publishers towards first-

party tracking and shared ID solutions, neutral standing towards Google’s efforts to undo their harm

provoking with their upcoming change, and an aversion towards a nation-wide walled garden. These findings

intend to increase the volume of the discussion on the effects of BigTech’s changes on the digital advertising

ecosystem as a whole and therefore stimulate further research on the effects on single actors within this

ecosystem – beyond the publishers themselves.

1 INTRODUCTION

The online advertising industry is currently facing

significant challenges. With Apple’s Safari and

Mozilla’s Firefox disabling third-party cookies in

2020 and Google Chrome's announcement that third-

party cookies will be disabled by 2023 (Sparkes,

2022; Szabocsik, 2021), lately postponed to 2024

(Love, 2022), many applications of relevant

stakeholders of the online advertising ecosystem will

be eliminated (Szabocsik, 2021). The decision to

phase out tracking cookies was made to protect the

privacy of users. In this research, the aim is to initiate

an open discussion lead by the clarification of the

standpoint of publishers as suggested by the

Marketing Science Institute (2020) in their research

priorities. It is yet unclear, how disabling third-party

cookie tracking will affect publishers and their

business models.

A cookie is a text string that is stored in the

Internet user's browser when the user accesses a

particular website. The web cookie was invented in

1994 with the intention of maintaining status between

clients and servers (Cahn et al., 2016) and is used to

store and read different data. For example, products

that have been placed in a shopping cart on a website

still appear in the shopping cart when the website is

https://orcid.org/0000-0002-1014-1830

accessed later thanks to the cookies. Furthermore,

log-in data, details on personal information and other

data can also be stored by a cookie (Github, 2019).

If a cookie is set by the website on which an

Internet user is staying, it is called a first-party

cookie. These cookies are generally used to identify

users, remember the user's settings, or save the

shopping cart (Cookie-Script, 2021). A company that

owns first-party cookies can enter into a partnership

with another company. If the first-party cookies are

forwarded to the partner company, then the partner

company subsequently owns second-party cookies

(Cookie-Script, 2021). Cookies that are set for the

Internet user by domains other than the one that

appears in the URL line of the browser are called

third-party cookies.

Cookies are domain-related, meaning that a third-

party provider has a different ID stored for the

Internet user than another third-party provider.

Cookie synchronization provides a channel for

information exchange between different third-party

providers in the background, in order to obtain

information about Internet users and, for example, to

serve user-specific advertisements (Papadopoulos et

al., 2019).

Worldwide, different approaches to replace use of

cookies are already being developed. However, it is

Stallone, V., Gägauf, A. and Kaya, T.

Who Ate All Our Cookies? Investigating Publishers’ Challenges Caused by Changes in Third-party Cookie Tracking.

DOI: 10.5220/0011336400003318

In Proceedings of the 18th International Conference on Web Information Systems and Technologies (WEBIST 2022), pages 97-104

ISBN: 978-989-758-613-2; ISSN: 2184-3252

unclear whether one of these solution approaches is

sufficient for a publisher or whether a combination of

different solutions must be used to avoid a potential

drop in revenue. Also unclear is, whether these

changes will affect national publishers differently.

The solution approaches need to be investigated so

that a publisher can properly plan its strategy for

selling programmatic inventory. The goal of this

work is to present the different solutions and prepare

for discussion. The advantages and disadvantages of

these approaches should be explored, and its

implications investigated.

The focus of this work is only on publishers, in

order to being able to deliver specific

recommendations for action for this essential

participant of the digital advertising ecosystem

(Gusic & Stallone, 2020). Effects on other

participants of the online advertising industry eco-

system (such as advertisers, intermediaries, and

users) are not analyzed. Although being a worldwide

phenomenon, we geographically limit our work to

Switzerland. Mobile applications are not considered.

We start this paper by highlighting the online

advertising ecosystem narrowing down the focus of

the potential revenue loss due to the disablement of

Third-Party-Cookies. We then move on to our mixed

method procedure: We present (1) our hypotheses

deduced from interviews with experts and (2) show

the survey with Swiss publishers. After showing the

results of the survey, we discuss them and interpret

them, in order to derive a differentiated set of

implications.

2 RELATED WORK

2.1 Publishers and Tracking

McKinsey already examined the implications for US

publishers of disabling third-party cookies

(Brodherson et al., 2021). For this purpose, the

authors conducted 28 expert interviews in the US and

came to the conclusion that 80% of online advertising

activities on non-premium publishers is placed with

third-party cookies. For premium publishers, this

percentage is lower, because many of these publishers

already use first-party cookies. The aim of another

study was to find out the actual use of ad choices and

financial impact (Johnson et al., 2020). The results

showed that only a small part (0.26%) of advertising

inventory sold to publishers in Europe, had been

shown to users who opted out of behavioral targeting.

Further on, the authors found out that these

advertising content generated 52% less revenue than

advertising content with behavioral targeting

(Johnson et al., 2020). Similar arguments are made in

two other studies in this area way back in 2014 and

2013 respectively. 2014, authors concluded that the

revenue loss for impressions without cookies ranges

from 37.5% to 66% (Beales & Eisenach, 2014). 2013,

scholars noted in their publication a 30% loss of

revenue if the top 5% users do not allow web tracking

(Gill et al., 2013). Earlier on, authors concluded in

their study that a publisher can make twice as much

revenue when buying via programmatic with

behavioral targeting than compared to revenue via the

traditional sales channel without behavioral targeting

(Chen & Stallaert, 2010). In contrast to the studies

listed so far, which all predict a significant loss of

revenue for the publisher, scholars concluded that a

publisher should expect only a small loss, when

avoiding behavioral targeting as an offering (Marotta

et al., 2019).

2.2 Third-party Tracking Alternatives

2.2.1 First-party Data

First-party cookies enable a publisher to create its

own walled garden by aggregating first-party data.

An example of how first-party data can be linked is

shown by the solution from Meta, which offers

advertisers a solution on how target audiences can be

found again within their platforms with the help of the

advertiser's first-party data. With Meta's conversions

API, the advertiser has the possibility to share his

first-party data with Meta's server. Meta promises that

in combination of the Meta Pixel and the Conversions

API will improve the performance and measurement

of the advertising campaign.

The study of Diener et al. (2020) goes on to state

that first-party data can be used as a basis for

personalization and measurement if publishers and

advertisers collect the data of their users and process

it. Diener et al. (2020) have concerns about the fact

that the open Internet is becoming a more and more

proprietary, leading to smaller, national publishers

having no chance to successfully use their first-party

data when compared to the data held by the big walled

gardens Alphabet, Meta, Apple, Microsoft, and

Amazon (Diener et al., 2020).

2.2.2 Shared ID

With the elimination of third-party cookies, however,

this solution is gaining more importance. The shared

ID solution is based on first-party cookies. When a

website is visited, different first-party data is shared

with it. In the process we can talk about declared and

derived information (ID5, 2022; Papadopoulos et al.,

WEBIST 2022 - 18th International Conference on Web Information Systems and Technologies

2019). Declared information is the e-mail address

that is voluntarily entered by the user on the website,

also known as first-party data according to Hassler

(2021). When visiting a website, passive

identification signals such as the IP address or the

user agent of the of the device are exchanged via the

http-protocol, also called derived information. It can

be processed by algorithms to deduce the uniqueness

of the user. This information is considered personal

data and requires the consent of the user due to legal

regulations (e.g. GDPR and CCPA), also known as

zero-party data according to Hassler (2021).

In order for the collected first-party data to be

further used, a publisher must work with an ID

provider. Via an API, the collected first-party cookies

are forwarded to the ID provider. If the required

consent of the user is available, it is the task of the ID

provider to create an ID for this user. This ID consists

of random sequences of numbers and letters. The API

then sends this ID back to the publisher so that the

publisher can store the ID in the user's first-party

cookie, which can be read by SSPs and DSPs (Davies,

2019). Thus, advertisers can recognize their target

audience on the publisher's website when using this

approach (Wakefield & Mussard, 2021).

2.2.3 Privacy Sandbox

In addition to the information that Google will block

third-party cookies in their browser, readers of the

post were also informed that Google has launched a

new initiative called Google Privacy Sandbox

(Cooper et al., 2022; Geradin et al., 2021). One

component of this solution is interest-based

advertising technology. Federated Learning of

Cohorts (FLoC) is an API that can be used to target

groups of web users with similar interests into

clusters. The FLoC API relies on a cohort assignment

algorithm, assigning web users to cohort ID based on

their browsing history. The browser updates the

cohort ID when web users are active. In order to

ensure privacy, the browser requires that this cohort

ID be shared by at least n different users (Turati,

2022), whereas n has to be above a certain predefined

threshold.

A cohort assignment algorithm is a trade-off

between privacy and utility: The more users share a

cohort ID, the harder it is to use this signal to infer an

individual user's behavior from the entire Web. On

the other hand, a large cohort is more likely to consist

of a large number of users, making it harder to use

this information to personalize ads (Turati, 2022).

According to Bindra (2021) and Diemert et al. (2022),

initial tests with simulations based on FLoC

suggestions from Chrome were successful. The result

was that advertisers can expect at least 95% of the

conversions per dollar spent when compared to

cookie-based advertising.

3 HYPOTHESES BUILDING

3.1 Interviews

In order to build our hypotheses, we relied on online

advertising experts. We planned to find experts

within the online advertising ecosystem of the online

advertising industry (see Gusic & Stallone, 2020 for

a thorough presentation of them). Individuals were

contacted via email or LinkedIn. We ended up with

six experts willing to support our research.

In order to be able to evaluate the six interviews,

we decided for this work to use content analysis

according to Mayring (2020). This method is used for

the systematic processing of texts and should help to

gain new insights. We evaluated the transcripts of the

interviews summarizing the content analysis. The

deductive approach pursues the goal of classifying

and utilizing the statements from the interviews based

on predefined categories.

3.2 Hypotheses

The experts see various advantages in first-party data.

Since the data comes directly from the publisher, they

can decide for themselves which data is shared with

whom. However, first-party data also requires the

publisher to install a login wall in order to access the

data. All experts recognize similar side effects with

first-party data. A lot of resources must be made

available by the company so that first-party data can

be built up. The experts assume that only larger

publishers will be able to provide these technical and

human resources at the beginning. It is not only the

resources that the experts consider problematic, but

also the density of the data. A large publisher will be

able to provide much more data than a small publisher

with generic content. A login wall could have a

deterrent effect on an Internet user, which could lead

to a publisher losing a website visitor because an

attempt is made to build up first- party data. From an

advertiser's perspective, the problem is that first-party

data cannot be used across publishers. Advertisers

would therefore have to plan a strategy per publisher.

Based on this evaluation, the following hypothesis

was made:

H1: Large publishers will invest resources to

strengthen their first-party data.

With regard to the shared ID, the experts see the

advantage of standardization. The ID created via a

Who Ate All Our Cookies? Investigating Publishers’ Challenges Caused by Changes in Third-party Cookie Tracking

shared ID. The ID can be understood by all parties

affiliated with this provider, although login data being

required. The experts strongly advise against working

with shared ID providers who still rely on

fingerprinting. Different shared ID providers give rise

to the problem that the IDs are not understood if

different providers are used by the parties. Another

concern expressed by the experts is the long-term

nature of this solution. This is because individual user

data is shared between the parties. Some experts fear

that for this reason the solution will be restricted again

by law in a few years. The evaluation therefore raises

a second hypothesis:

H2: Smaller publishers who market themselves

will work with shared ID providers.

Products such as contextual or geo-targeting not

only offer new ways to track a user in the industry,

but could also focus more on the environment a user

is in. These targeting methods have existed for some

time and are likely to be used forever. With

contextual targeting, the advertiser buys specifically

on an inventory. Despite the side effects, experts

argue that contextual targeting is essential for a

publisher. This is also because this targeting method

represents a long-term solution.

H3: All publishers already know about

contextual targeting.

The experts were asked whether they considered

a Swiss walled garden to be possible. The experts

were of very similar opinion on this question. A great

potential is seen in this approach. It would make it

easier for advertisers to plan their advertising buying,

as it would be possible to buy from several publishers

with the same strategy. But it would also bring

advantages for Swiss publishers, as the publishers

would then be more competitive against the big

companies Meta, Alphabet, Apple, Microsoft and

Amazon. Publishers also benefit from shared data.

For example, Publisher 1 only knows the age and

Publisher 2 only the gender of the Internet user.

Through shared data, this results in an Internet user

where age and gender are known. However, a Swiss

walled garden requires that all publishers participate.

Some experts express the fear that certain publishers

would not participate at the beginning, but that these

publishers will join at a later stage. What would need

to be considered in a Swiss Walled Garden is a way

to share the data between the parties. After this

evaluation, a fourth hypothesis is made.

H4: All publishers are ready to build a Swiss

Walled Garden.

Google's Privacy Sandbox was also mentioned as

another approach to solving the problem of

eliminating third-party cookies. Various advantages

are noted here. On the one hand, this approach allows

access to data, even to parties who otherwise do not

have any data. Another important aspect of this

solution is the fact that an Internet user is no longer

tracked individually. Rather, the Internet user is

hidden in a group with similar interests. The reach

that can be achieved with this solution is also listed as

an advantage. From the Internet user's point of view,

it is further assumed that they feel less tracked with

this solution. Especially because Google builds the

solution on the W3C-standard, it would be optimal for

the online advertising industry if other browsers

would also implement this approach. However, there

are also some concerns expressed about Google's

solution. Why an Internet user is assigned to a cohort

is no longer comprehensible. The solution is therefore

a black box for Internet users, but also for online

advertising industry participants. Experts see another

side effects in the even greater market power that

Google will gain as a result. Two experts expressed

the wish that such a solution should have come from

an independent organization such as the IAB.

4 ASKING THE PUBLISHERS

4.1 Methodological Approach

A total of 12 questions on various topics were asked

in the survey. The questions were mainly closed

questions, which respondents could answer with

single or multiple choice. The operationalization of

the survey is shown in Table 1. In some cases, hybrid

questions were asked. For these questions,

respondents had the option of choosing between

either predefined answers or their own answer under

"other". Four open-ended questions were also asked.

These questions always involved a statement of

reasons for an answer previously given. At the

beginning of the questionnaire, six questions were

asked about the company. They were served to find

out differences between different attributes. If it was

stated that no programmatic inventory was offered,

then the respondents had to answer whether there

were plans to offer the programmatic sales channel in

the future and via which channel in order to be

forwarded to the questions for publishers. The

condition for this diversion was also that the

inventory is marketed by the publisher.

WEBIST 2022 - 18th International Conference on Web Information Systems and Technologies

100

Table 1: Operationalization.

Term Variable Indicator Expression

Company Size

Number of

Employees

1-9

10-49

50-249

250 or more

Program-

matic

inventory

Usage Application

Yes

Accep-

tance

Importance

5: Very

important

4: Important

3: Neutral

2: Not

important

1: Not

important at

all

Promo-

tion

Selling

inventory

Self

Interme-

diary

Approach

Familia-

rity

Shared ID

solutions

Familiar or

not

First-party data

Own walled

garden

Google Privacy

Sandbox

Proba-

bility of

occurren

Shared ID

solutions

First-party data

Own walled

garden

Google Privacy

Sandbox

5: Very

probable

4: Probable

3: Neutral

2: Not

probable

1: Not

probable at

all

Swiss

Walled

Garden

Willngness to

join

Yes

In order to check whether the survey functioned

correctly, a pretest was first carried out. People from

the authors' environment filled out the survey and

checked whether the branchings were correct and

whether there were any spelling mistakes. The

respondents could fill out the survey in German or

English. The survey took about five minutes to be

completed. The survey with the Swiss publishers was

conducted between May 10 and May 23, 2021, one

year after the first announcement of phasing out third-

party cookies by 2022, one month before the

announcement of delaying it to 2023 and almost a

year before the second postponement to 2024 (Love,

2022). The participants were deliberately written to.

Only people who work in the programmatic area for

Swiss publishers and advertisers were contacted. We

contacted a total of 20 Swiss publishers.

4.2 Results

We excluded survey that were not completed

resulting in ten response sets to be considered. All of

the respondents confirmed, they were selling

programmatic inventory. We depict the results of all

the questions in Table 2.

Table 2: Results.

Term Variable Indicator Results

Company Size

Number of

Employees

1-9 = 1

10-49 = 3

50-249 = 4

250 or more

= 2

Program-

matic

inventory

Usage Application 10 yes

Accep-

tance

Importance avg. 3.8

Promo-

tion

Selling

inventory

Self = 7

Interme-

diary = 3

Approach

Familia-

rity

Shared ID

solutions

Familiar = 4

First-party data Familiar = 7

Own walled

garden

Familiar = 7

Google Privacy

Sandbox

Familiar = 7

Proba-

bility of

occurren

Shared ID

solutions

avg. 3.3

First-party data avg. 4

Own walled

garden

avg. 1.3

Google Privacy

Sandbox

avg. 3.4

Swiss

Walled

Garden

Willngness to

join

Yes = 4

No = 4

No answer

= 2

5 FINDINGS

The analysis of the individual hypotheses is based on

the size of the companies. They are subdivided as

follows: Large companies (250 or more employees),

Who Ate All Our Cookies? Investigating Publishers’ Challenges Caused by Changes in Third-party Cookie Tracking

101

medium companies (50-249 employees) and small

companies (1-49 employees).

H1: Large publishers will invest resources to

strengthen their first-party data.

To evaluate the first hypothesis, the responses on

the login wall for first-party data, as well as the

information on the probability of building up first-

party data, were compared with the size of the

companies. Out of the three large companies overall,

all would install login walls at websites to collect

first-party data (100%). In comparison, five of the

respondents who do not work for a large company say

they would not build a login wall, while three of the

respondents in this category would consider a login

wall. Regardless of whether a login wall is installed,

the inclusion of first-party data in the strategy of both

large and small companies is considered very likely.

Based on the results of this analysis, the first

hypothesis is confirmed.

H2: Smaller publishers who market themselves

will work with shared ID providers.

For the evaluation of the second hypothesis, the

company type Publisher was taken into account.

Larger collaborations consider cooperation with a

shared ID provider. A collaboration with a shared ID

provider is indicated as "very probable" or "probable"

by three respondents. All these three individuals work

for medium or large companies. Respondents from

smaller companies, indicate that collaboration with a

shared ID provider is "not probable" or "not probable

at all". According to this analysis, the second

hypothesis must be falsified. Currently, it looks more

like medium-sized and large companies will

cooperate with a shared ID provider. This result is

hardly surprising insofar as this solution approach has

generally received little approval.

H3: All publishers already know about

contextual targeting.

The third hypothesis’ statement cannot be

confirmed: Only 7 out of 10 said they were familiar

with this potential alternative. For this reason, the

third hypothesis must be falsified.

H4: All publishers are ready to build a Swiss

Walled Garden.

In order to be able to answer the fourth

hypothesis, the question about the Swiss walled

garden was evaluated. Since an initial analysis

already showed that less than the half of the

respondents would be willing to join a Swiss walled

garden, the question was examined in more detail to

identify any patterns. We encountered that

willingness is highest among medium-sized

companies, at 60%. In a second comparison, where

the willingness was put in relation to the type of

company, a higher agreement can be seen among

publishers (50%). Not all publishers and marketers

are willing to set up a Swiss walled garden. For this

reason, the fourth hypothesis must be falsified.

Although the establishment of a joint Swiss walled

garden cannot yet be given any real chance, half of

the publishers are open to such a project. The fourth

hypothesis has therefore not been confirmed. It will

take time and experience to win over more companies

for a joint Swiss walled garden.

The result of the analysis shows that there is

currently no consensus among Swiss publishers

regarding the right solution. The survey was able to

confirm only one of the four hypotheses. Agreement

can be seen with the approaches of first-party data and

the Google Privacy Sandbox.

6 DISCUSSION

The aim was to find out, from the publishers' point of

view, which approaches could represent a possible

solution. The results of the survey clearly show that

publishers are generally very familiar with the various

solutions. The solution approach of building up first-

party data itself achieves the highest probability of

being included in the strategy of the companies. There

is also a high level of agreement with the Google

Privacy Sandbox. Shared IDs were exclusively

indicated by large companies as a probable solution.

What does not represent a solution approach for

publishers is the establishment of their own walled

garden. The participants in the survey also do not see

a Switzerland-wide walled garden as an optimal

approach. At least the importance of the topic was

recognized. This is because the programmatic sales

channel is relevant for many of the publishers, which

consequently means that all of these publishers must

expect a drop in revenue if third-party cookies are

deactivated.

6.1 Implications

Many publishers see the creation of first-party data as

a possible solution. In order to implement this

solution, the publisher must first clarify important

strategic questions. Decisions must be made about

how to access the data, how to interpret the data, what

data and also how and with whom the data will be

shared. Some experts have expressed fears of a waste

WEBIST 2022 - 18th International Conference on Web Information Systems and Technologies

102

of resources on this point. In this work, it was

assumed that only the large companies would build

up first-party data due to the resources required and

the high effort involved. However, regardless of the

company size, it has been shown that this approach is

followed in all the companies surveyed. The actual

effort of each individual company should not be

underestimated. After all, SSPs and DSPs cannot read

first-party data from publishers without additional

effort. If all publishers collect their own first-party

data and offer it as targeting, this means that an

advertiser must increasingly buy from individual

publishers and can no longer apply a unified buying

strategy.

A smaller effort exists for the publisher in a

cooperation with a shared ID provider. The

assumption that smaller publishers use this approach

is related to the fact that these publishers have fewer

human and financial resources. However, the survey

showed that large publishers are considering such

collaboration. Shared ID providers have been around

for a while, but the awareness of this solution

approach is the lowest. The fact that there are already

many different providers could make it even more

difficult to understand the solution approach. It could

be that due to the lack of understanding, there is a

reluctant attitude towards this solution approach. The

acceptance of the Google Privacy Sandbox by Swiss

publishers should also be viewed with caution when

it comes to weakening Alphabet's monopoly position:

This could intensify the mechanism of "digging one's

own grave".

Not all types of targeting are affected by the

deactivation of third-party cookies. Contextual

targeting is already used in campaigns. The solution

approach is very well known among publishers, but

not as well as we originally thought. Some experts

also pointed out the advantage that this type of

targeting will never be affected by data protection

laws and can always be used. This fact led to the

assumption that this solution approach is followed by

all publishers. However, based on the survey results,

this assumption had to be falsified. From the author's

point of view, this approach represents a sustainable

solution. However, it was probably rejected because

the question was formulated imprecisely. The survey

should have been supplemented with the probability

of cookieless solutions. The fact that this type of

targeting is already in use perhaps means that

contextual targeting is not perceived as a solution to

this problem, but more as simply targeting that still

works. Disabling third-party cookies might lead to the

risk of shifting the competitive advantage to the big

foreign companies such as Alphabet, Meta, Apple,

Amazon, and Microsoft. This is because the solutions

from these companies also work without third-party

cookies. To counteract this, a Switzerland-wide

solution seems sensible and appropriate. However,

the participants disagreed on the establishment of a

Swiss walled garden. Such a project requires

technical know-how as well as the firm will of all

participants in the online advertising industry. The

experts considered a nationwide solution across all

market participants to be the best way to achieve this

goal. To be able to do this, the participation of the

large media houses and marketers is particularly

needed. However, the survey shows that approval

from these companies is rather low. Various points

are cited as reasons for their rejection. For example,

there are fears that Switzerland is not technically

ready for such a project, that closed systems limit

innovation and that data protection could suffer as a

result. In particular, little attention is paid to walled

gardens because they are closed systems.

Consequently, solutions such as the development of

open systems would increase the interest of market

participants.

6.2 Limitations and Future Research

In this paper, we undertook an investigation into the

world of the online advertising ecosystem and the

perspective of publishers on it. Since the number of

publishers relevant to this ecosystem in Switzerland

is not large, we ended up with a reduced number of

respondents. Because the topic is also relevant for

other countries, the same study should be extended to

the surrounding German-speaking countries like

Germany and Austria. With this further research,

differences and similarities between the countries

could be identified and it could be found out whether,

for example, other data protection laws influence the

assessment of the solution approaches. Another

recommendation is to repeat the study when the

solutions are in use.

REFERENCES

Beales, H., & Eisenach, J. A. (2014). An Empirical Analysis

of the Value of Information Sharing in the Market for

Online Content (SSRN Scholarly Paper No. 2421405).

Social Science Research Network.

https://doi.org/10.2139/ssrn.2421405

Bindra, C. (2021, January 25). Building a privacy-first

future for web advertising. Google.

https://blog.google/products/ads-commerce/2021-01-

privacy-sandbox/

Brodherson, M., Broitman, A., Macdonald, C., & Royaux,

S. (2021). The demise of third-party cookies and

identifiers. McKinsey.

https://www.mckinsey.com/business-

functions/marketing-and-sales/our-insights/the-

demise-of-third-party-cookies-and-identifiers

Who Ate All Our Cookies? Investigating Publishers’ Challenges Caused by Changes in Third-party Cookie Tracking

103

Cahn, A., Alfeld, S., Barford, P., & Muthukrishnan, S.

(2016). An Empirical Study of Web Cookies.

Proceedings of the 25th International Conference on

World Wide Web, 891–901.

https://doi.org/10.1145/2872427.2882991

Chen, J., & Stallaert, J. (2010). An Economic Analysis of

Online Advertising Using Behavioral Targeting (SSRN

Scholarly Paper No. 1787608). Social Science

Research Network.

https://doi.org/10.2139/ssrn.1787608

Cookie-Script. (2021). All you need to know about third-

party cookies. https://cookie-script.com/all-you-need-

to-know-about-third-party-cookies.html

Cooper, D. A., Yalcin, T., Nistor, C., Macrini, M., &

Pehlivan, E. (2022). Privacy considerations for online

advertising: A stakeholder’s perspective to

programmatic advertising. Journal of Consumer

Marketing, ahead-of-print(ahead-of-print).

https://doi.org/10.1108/JCM-04-2021-4577

Davies, J. (2019, September 23). WTF are shared identity

solutions? Digiday. https://digiday.com/media/what-

are-shared-identity-solutions-and-can-they-really-

replace-cookies/

Diemert, E., Fabre, R., Gilotte, A., Jia, F., Leparmentier, B.,

Mary, J., Qu, Z., Tanielian, U., & Yang, H. (2022).

Lessons from the AdKDD’21 Privacy-Preserving ML

Challenge. Proceedings of the ACM Web Conference

2022, 2026–2035.

https://doi.org/10.1145/3485447.3512076

Diener, C., Treidler, A., von Wersch, O., & Stapelbroek, J.

(2020). Online Advertising in the Post-Cookie Era.

German Association for the Digital Economy.

https://www.bvdw.org/veroeffentlichungen/publikatio

nen/detail/artikel/online-advertising-in-the-post-

cookie-era/

Geradin, D., Katsifis, D., & Karanikioti, T. (2021). Google

as a de facto privacy regulator: Analysing the Privacy

Sandbox from an antitrust perspective. European

Competition Journal, 0(0), 1–65.

https://doi.org/10.1080/17441056.2021.1930450

Gill, P., Erramilli, V., Chaintreau, A., Krishnamurthy, B.,

Papagiannaki, K., & Rodriguez, P. (2013). Best paper -

- Follow the money: Understanding economics of

online aggregation and advertising. Proceedings of the

2013 Conference on Internet Measurement Conference,

141–148. https://doi.org/10.1145/2504730.2504768

Github. (2019). Modul: Trackingmethoden und

Gegenmaßnahmen. GitHub.

https://github.com/datenklaus-app/datenklaus_webapp

Gusic, N., & Stallone, V. (2020). The Digital Advertising

Ecosystem—Status Quo, Challenges & Trends. 36–42.

http://www.iadisportal.org/digital-library/the-digital-

advertising-ecosystem-%C2%96-status-quo-

challenges-and-trends

Hassler, M. (2021). Von Data-driven zu People-based

Marketing: Erfolgreiche Digital Marketing Strategien

in einer Privacy First Ära (2021st ed.). mitp.

ID5. (2022). The Advertising Identity Guide.

https://advertisingidentity.guide

Johnson, G. A., Shriver, S. K., & Du, S. (2020). Consumer

Privacy Choice in Online Advertising: Who Opts Out

and at What Cost to Industry? Marketing Science,

39(1), 33–51. https://doi.org/10.1287/mksc.2019.1198

Love, J. (2022, July 27). Google Delays Phasing Out Ad

Cookies on Chrome Until 2024. Bloomberg.Com.

https://www.bloomberg.com/news/articles/2022-07-

27/google-delays-phasing-out-ad-cookies-on-chrome-

until-2024

Marketing Science Institute. (2020, May 7). MSI

Announces 2020-22 Research Priorities. Marketing

Science Institute. https://www.msi.org/articles/2020-

22-msi-research-priorities-outline-marketers-top-

concerns/

Marotta, V., Abhishek, V., & Acquisti, A. (2019). Online

Tracking and Publishers’ Revenues: An Empirical

Analysis. 35.

Mayring, P. (2020). Qualitative Forschungsdesigns. In G.

Mey & K. Mruck (Eds.), Handbuch Qualitative

Forschung in der Psychologie: Band 2: Designs und

Verfahren (pp. 3–17). Springer Fachmedien.

https://doi.org/10.1007/978-3-658-26887-9_18

Papadopoulos, P., Kourtellis, N., & Markatos, E. P. (2019).

Cookie Synchronization: Everything You Always

Wanted to Know But Were Afraid to Ask. The World

Wide Web Conference on - WWW ’19, 1432–1442.

https://doi.org/10.1145/3308558.3313542

Sparkes, M. (2022). Google moves to drop cookies. New

Scientist, 253(3372), 15.

https://doi.org/10.1016/S0262-4079(22)00177-4

Szabocsik, M. (2021, May 4). How Marketing Can Move

Forward As Third-Party Cookies Are Phased Out.

Forbes.

https://www.forbes.com/sites/forbescommunicationsc

ouncil/2021/05/04/how-marketing-can-move-forward-

as-third-party-cookies-are-phased-out/

Turati, F. (2022). Analysing and exploiting Google’s FLoC

advertising proposal. https://doi.org/10.3929/ethz-b-

000539945

Wakefield, L., & Mussard, H. (2021). IAB Europe Guide to

the Post Third-Party Cookie Era. IAB Europe.

https://iabeurope.eu/knowledge-hub/iab-europe-guide-

to-the-post-third-party-cookie-era-updated-in-

february-2020/

WEBIST 2022 - 18th International Conference on Web Information Systems and Technologies

104