The Tracker Tax: How Third Party Trackers Impact Website Speed in the United States
Personal consumer data has become the oil of the 21st century. American companies are expected to spend close to $20 billion on consumer data in 2018 (New York Times 2018). Data trackers are the main way advertisers gather this valuable consumer data.
Trackers are small JavaScripts used by advertisers like Google and Facebook to monitor consumers while they browse websites. In addition to privacy concerns, each tracker on a webpage costs webpage speed. Researchers at Ghostery and CliqZ conducted a study to see the performance effects of tracking of the top 500 websites.
The Prevalance of Tracking
Using a custom crawler, the researchers set out to collect the number of trackers and page load times for the top 500 websites in the US as determined by Alexa. Using Ghostery, an ad and tracker blocking browser extension, and some performance-oriented JavaScript the researchers found that just over 10% of the top websites are tracker free. In other words nearly 90% on the top 500 websites use one or more trackers.
- Nearly 90% of top 500 websites have at least one tracker
- 65% have at least 10 trackers
- Over 20% have 50 or more trackers
- Some of the top sites have well over 100 trackers
A previous study of 144 million page loads in 12 countries found that 77.45% of page loads contain trackers (Macbeth 2017).
The Cost of Tracking
On average websites take more than twice as long to load with trackers than without (see Figure 1). The average top 500 website loaded in 19.3 seconds with trackers. With trackers blocked the load time dropped by 55% to 8.6 seconds.
Figure 1: Mean Load Time of Top 500 Websites With and Without Trackers
The distribution of page load times of the top 500 websites is shown in Figure 2. Only 17% of the pages tested loaded within 5 seconds. Nearly 60% of pages loaded in more than 10 seconds, 18% of pages loaded in more than 30 seconds, and nearly 5% of pages took more than one minute to load.
Figure 2: Load Time Distribution of Top 500 Websites With Trackers
The study showed a direct relationship between the number of trackers and page load time. The more trackers the slower the page load time. The closest quadradic model of the relationship between trackers and page load time found a compounding effect for trackers. Each additional tracker adds about 2.5% to page load time.In addition, the researchers found a "piggybacking" effect, where trackers load other trackers not originally on the site, further compounding the slow page load times.
The 10 Most Tracked Websites
The most tracked websites in the study showed the cumulative effect of trackers. Seven of the top 500 websites had over 100 trackers present (see Figure 3).
Figure 3: Most Tracked Domains With and Without Trackers
The slowest domains were on average 10 times faster when trackers were blocked and users saved an average of 84 seconds per page load (see Figure 4).
Figure 4: Slowest Top 500 Websites With and Without Trackers
Ad and Tracker Blocking Software
Raymond.cc conducted a test of 10 ad blocking extensions on performance (Raymond.cc 2016). The fastest ad blocking software improved page speed by more than 3 times (uBlock Origin). Overall, the three best performing ad blocking extensions were uBlock Origin, Ghostery, and Adguard. uBlock won in raw page speed (see Figure 5) and CPU usage. Ghostery won in peak memory used in Chrome.
Figure 5: Three Fastest Ad & Tracker Blocking Software
Conclusion
The use of tracking software has a significant compounding effect on website performance. Each additional tracking widget slows web page load time by 2.5%. For the average top 500 website, page speed was greater than 2 times faster by blocking tracking. The 10 slowest websites tested were 10 times faster without tracking. Reducing the number of trackers on your website can give a significant boost to web page speed.
Further Reading
- Dance, G., LaForgia, M. & Confessore. N. New York Times, Dec. 19, 2018 - "Facebook Offered Users Privacy Wall, Then Let Tech Giants Around It."
- Front page article on Facebook personal data sharing practices. American companies are expected to spend close to $20 billion by the end of 2018 to acquire and process consumer data, according to the Interactive Advertising Bureau.
- Raymond.cc, 2016 - 10 Ad Blocking Extensions Tested for Best Performance
- Raymond.cc tested 10 ad and tracking blocking extensions in Firefox and Chrome. Using 10 popular websites (including CNN.com) Raymond tested the extensions for average load time, peack CPU usage, and peak memory usage with and without ad blocking software. uBlock proved best in average load time and peak CPU usage. Ghostery tested best in peak memory usage.
- S. Macbeth, 2017 - "Tracking the Trackers: Analysing the global tracking land- scape with GhostRank."
- In this analysis of the Ghostrank data, CliqZ found that Google and Facebook dominate the tracking universe. Google tracks 64.4% of all web pages tested, and Facebook tracks 28.8%.
By website optimization on 19 Dec 2018 PM