By Erwin SOTIRI
Jurisconsul law firm
“Just when I decided to be on a diet, everyone on the Internet talks about cookies. So I accept them all !”, ISMO
The EU Cookie Policy is primarily derived from the ePrivacy Directive (2002/58/EC) as amended by Directive 2009/136/EC. It requires websites to obtain informed consent from users before storing or retrieving information on a user's computer or mobile device via cookies. It is specifically focused on privacy issues related to electronic communications. Under this policy, websites must provide clear and comprehensive information about the purpose of the cookies or similar technologies they use and obtain the user's consent before placing cookies, except for those strictly necessary for the provision of a service explicitly requested by the user.
While the EU Cookie Policy and GDPR both seek to protect user privacy, their scope, application, and compliance requirements differ significantly. The Cookie Policy applies only to cookies and similar technologies and requires consent for their use, whereas the GDPR establishes a broader protection framework for personal data processing, emphasising individual rights and organisational responsibilities. Organizations operating in the EU must ensure adherence to both regulations in order to meet legal obligations and protect user privacy.
The exhausting experience of cookie consent banners
"Consent Fatigue" is one of the most notorious criticisms towards the cookie consent mechanism. Put it another way stated, the overabundance of consent requests has resulted in "consent fatigue" among users, which undermines the policy's objective of obtaining informed consent by encouraging them to click "accept" without thoroughly perusing or comprehending the ramifications. In fact, reading and managing dozens of pages of consent policies becomes quickly daunting and thus harming the objective of user’s fully informed consent sought by the ePrivacy Directive.
However, the objective of this article is not to venture into an analysis of numerous criticisms directed towards the ePrivacy Directive. Instead, it will focus on a single economic aspect: the global cost of the persistent reminders for users to click on the Cookie Consent button each time they visit a webpage.
The energy cost of EU cookie consent policy: a nearly tedious calculation method
To estimate the cost of the EU Cookie Consent Policy, we must proceed with a step-by-step analysis:
Determining the number of daily consulted websites
Recent data shows that there are more than 1.7 billion websites in the world. Statistics suggest that users visit around 130 web pages per day on average. However, the source information [source: https://kickstand.typepad.com/metamuse/2007/10/how-many-web-pa.html] indicates that this data only covers the USA and that the estimate is from 2011. We can only guess that this number has grown a lot since 2011. We did not find updated data that would help us determine the number of websites per day. Therefore, we decided to make the arbitrary assumption that only 10 websites per day were new and could potentially have cookies.
Estimating the number of websites subject to EU Cookie consent regulation
Not all websites are subject to the EU Cookie Consent Policy. However, a significant portion of the internet is affected by this regulation. Based on available statistics, we grossly assume that 50 % of websites utilise cookies. This means that roughly 10*50% = 5 websites per day are concerned with EU Cookie Consent Policy [Sources: https://cookiyes.com/blog/cookie-consent-statistics/, https://cookielawinfo.com/blog/cookie-consent-statistics-2023/, https://www.statista.com/statistics/1260352/global-cookie-consent-implementation-rate/].
Making a conservative assumption that an average internet user visits in 2024 approximately 10 new websites per day, of which 5 use cookies we can estimate that roughly 8.5 billion new website consultations occur daily that need cookie consent. [Source: https://www.internetlivestats.com/internet-users/].
Number daily internet users
The latest data affirms that globally, 5.47 billion people use the internet daily [source: https://earthweb.com/how-many-people-use-the-internet-daily/]
Calculating the cost of a single click
Determining the cost of a single click involves considering multiple factors, including device energy consumption, network infrastructure, and server power consumption.
Device energy consumption:
Assuming a desktop computer consumes approximately 100 watts per hour, and considering a brief time spent on a website for a click (rounded off to 10 seconds), the energy consumed by the user's device for a single click is estimated to be 0.000278 kilowatt-hours (kWh).
Network infrastructure and server consumption:
Estimating the energy used by network infrastructure and servers to handle a single click is complex and varies based on factors like server efficiency, data center location, and routing. For simplicity, we can assume a reduced network and server infrastructure consumption of 0.00008 kWh for a single click [according to Perplexity answer: https://bit.ly/perpenergyclick].
Total estimated energy consumption for a single website click
Combining the energy consumption of the user's device and the network infrastructure/server, we arrive at an estimated energy consumption of 0.000358 kWh per website click.
Total daily energy consumption
To determine the total cost of the EU Cookie Consent Policy, we multiply the estimated energy consumption per website click (0.000358 kWh) by the number of daily internet users (5.47 billion) and of consultations that are subject to the regulation (5).
Daily Energy Consumption = Energy per click x Daily users x Consent banners
Daily Energy Consumption = 0.000358 x 5,470,000,000 x 5
This calculation yields a total power consumption of approximately 9,791,300 kWh.
Daily and annual costs
Assuming an average cost of 0.20 EUR per kilowatt-hour (kWh), we can calculate the daily and annual costs associated with the EU Cookie Consent Policy:
Daily Cost:
§ Daily Cost = Total Power Consumption × Cost per kWh
§ Daily Cost = 9,791,300 kWh × 0.20 EUR
§ Daily Cost ≈ 1,958,260 EUR
Annual Cost:
§ Annual Cost = Daily Cost × Number of Days in a Year
§ Annual Cost = 1,958,260 EUR × 365 days
§ Annual Cost ≈ 714,764,900 EUR
Therefore, the estimated daily worldwide energy spent for running the EU Cookie consent Policy would amount roughly to 10 million kWh with an estimated cost of 2 million EUR, and the estimated annual infrastructure cost of 715 million EUR.
It is important to note that these cost estimates are based on several assumptions and may vary depending on factors such as the efficiency of devices and servers, the location of data centres, and fluctuations in energy prices.
Economic impact of the EU cookie consent policy
This is indeed a very rough calculation that makes several arbitrary assumptions and leaves many relevant data out of the picture.
However, should this analysis bear some significance, the EU Cookie Consent Policy evokes a non-negligible energy consumption and the corresponding expenses, thereby exercising a substantial economic influence. Although the policy's primary objective is to safeguard user privacy, it concurrently places website operators under obligations of compliance and involves expenses. The economic ramifications of the policy and the protection of user privacy continue to be obstacles for policymakers and website owners.
Important context:This article began as a Sunday morning thought experiment. The original idea was to show why language models cannot solve legal problems in the way professional lawyers do. Some language models, such as GPT-4, can pass the bar exam, but that does not mean that they can “reason” in a legal environment, which involves solving issues that require multiple levels of abstraction and where data is not directly available but needs to be inferred from certain contexts. The challenge was to demonstrate it in a short article that does not confuse the reader with a maze of legal complexities. A simple math problem where some data might be missing could perhaps serve as an analogy. During my research, I had to deal with many consent banners, which gave me the idea to ask LLMs to estimate the global energy used in those clicks, which could also be a basis for discussing the “Consent Fatigue” that the EU Cookie Consent Policy unintentionally creates. The outcome was rather disappointing: none of the models were able to infer data and instead kept referring to random or useless data. Sometimes the math was also wrong. One amusing fact was that one model kept estimating at nearly 42% the websites that have cookies without further explanation. Maybe a GROK joke? After a few hours, I lost patience and ended up fixing everything manually, as I usually do with documents where chatbots are asked for facts or provisions. The result is this article, which aims at a truth that is probably much more complex than LLM step-by-step reasoning can handle. AGI, we are not there yet! |
Disclaimer:
Please be advised that the author of this article lacks expertise in statistics and the methods and assumptions utilised in the analysis may be flawed. The data included in this article was acquired through inquiries with different search engines including chatbots and should be regarded as approximations. The costs linked to the EU Cookie Consent Policy may substantially deviate in practise from the approximations presented in this article.
While the author has attempted to present information in this article in an accurate manner, they cannot guarantee that no errors or omissions have occurred. Any actions taken or decisions made in reliance on the information presented in this article are beyond the author's control.
Readers are advised to consult with qualified professionals before making any decisions or taking any actions related to the EU Cookie Consent Policy or its economic implications.
Sources:
The data and information presented in this article were obtained through interrogating the following search engines and chatbots:
ChatGPT
Perplexity
Google Bard
Bing AI
The reader acknowledges that these chatbots are still under development and their responses may not always be accurate or reliable. The author has attempted to verify the information provided by the chatbots through other sources, but cannot guarantee its accuracy.
Readers are advised to consult with qualified professionals before making any decisions or taking any actions related to the EU Cookie Consent Policy or its economic implications.