Thanks for taking the time to drop by my little part of the ‘net. 🙂
I wanted to create a quick study that I could knock out over Spring Break with little fuss with enough reliable, trustworthy data to inform your decisions on anti-tracking extensions.
The methodology was simple: Take a “swath of the Internet” and sample it with various browsers on various devices (with various extensions.)
My initial scope was a bit extreme — nine browsers, three operating systems, and (roughly) three solutions per platform. Originally, I wanted to test Firefox with uBlock and bog-standard Safari on MacOS, as well as Firefox Focus on iOS, to see if they had any noticeable impact. However, shortly into the testing suite, I realized just how much of a daunting task this would be.
(It took roughly one hour to run the full suite of tests — per browser.)
In the interest of having a Spring Break, I pulled the MacOS tests and brought things back to a more manageable level: Three Windows 10 browsers and two iOS browsers.
Test Data
Full Suite (Excel)
Individual Tests (CSV)
Test Equipment
While decidedly “unscientific,” I do, in the interest of fairness, want to show the specs of the machines I used. This is mildly important for page load times — machines with more CPU power, better Internet connections, and more RAM tend to perform better overall.
Or, as the old Netizens love to say, YMMV.
Edge, Vivaldi, Vivaldi + uBO, Firefox + uBO, Hardened Firefox:
- System: Dell Optiplex 3020
- OS: Windows 10 Pro
- CPU: Intel i5-4590t @ 2.00 GHz
- RAM: 8GB DDR3 @ 1600 MHz
- Storage: 256GB SATA SSD
- Displays: 1920×1080 (Ads and Trackers), 1680×1050 (Load Time)
- Internet Connection: ~50 Mbps Wi-Fi download (variable)
Safari, Safari + ABP
- System: Apple iPhone 13 Mini
- Internet Connection: ~50 Mbps Wi-Fi download (variable)
Hardened Firefox?
I had to call it something. 🙂
I went into Firefox’s Privacy and Security settings (about:preferences#privacy) and did the following:
- Enabled Strict Tracking Prevention
- Checked “Tell websites not to sell or share my data” and “Send websites a ‘Do Not Track’ request”
- Unchecked all boxes in “Firefox Data Collection and Use”
- Enabled HTTPS-only mode in all windows
- Set DNS over HTTPS to “Max Protection”
While this is not everything one can do to bulk up Firefox’s protections, they’re a few clicks that help to improve the privacy experience. (And I’ll leave it there. I won’t start the Cloudflare debate today.)
General Test Methodology
Upon visiting a site, I would scroll to the bottom of the page, then back to the top of the page, leaving enough time for the site to contact ad services and trackers. (This was usually fifteen to twenty seconds.)
However, a few sites had special procedures that differed from the ordinary “visit the site, scroll down, scroll back up” pattern.
- GoComics: Visit gocomics.com/peanuts. Click “Read Now.”
- YouTube: Using a browser that has not yet logged into the site, search for “pie” and click the first video.
- Spotify: Using a browser that is logged into a personal, non-Premium Spotify account, listen to music for thirty minutes and see if advertisement audio plays.
- Flickr: Visit flickr.com. Search for “pie.” Scroll for ten seconds.
- Reddit: Visit reddit.com. Scroll for ten seconds.
- Bing: Visit bing.com. Scroll for ten seconds.
- Google Search: Search Google for “pie.” Scroll for ten seconds.
- JetPhotos: Visit jetphotos.com. Scroll for ten seconds.
- Archive.org: Visit archive.org. Scroll for ten seconds.
(I like pie quite a bit.)
Test One (Trackers Blocked) Methodology
Edge: I entered the Settings menu via edge://settings/privacy/blockedTrackers and cleared the list for each new site visited. Any tracker that appeared in this list upon loading the site was counted toward Edge’s total.
Vivaldi: I used Vivaldi’s homescreen “Ads Blocked” and “Trackers Blocked” counters and pressed “Reset Statistics” for each new site visited. Any tracker that appeared in this list upon loading the site was counted toward Vivaldi’s total.
Hardened Firefox (Firefox + uBO + Privacy Settings): I used uBlock Origin’s built-in tracker counter, which automatically resets upon reloading a page or visiting a new page. The number it reported was counted toward Firefox’s total.
Safari: I used Safari’s built-in tracker counter (enabled on the New Tab page) and cleared the list for each new site visited. Any tracker that appeared in this list upon loading a new site was counted toward Safari’s total.
Safari + ABP: Same as Safari.
Test Two (Ads Shown) Methodology
All browsers: Any image or link with the word “Sponsored,” “Promoted,” “Advertisement,” or “Ad” counted to the total. Any page with “infinite scrolling” (no defined spot to end scrolling) was scrolled for ten seconds and then all passed ads were counted and added to the total.
Notes:
Desktop Bing bakes advertisements into the infinitely-scrolling newsfeed, while Mobile Bing provides a more static page.
DuckDuckGo clearly labels promoted links as [AD], while other search engines did not appear to do so. Both ABP and uBO removed these links.
YouTube and Spotify were tested for video and audio ads only.
Test Three (Load Time) Methodology
Safari for iOS does not have an easily-accessible Inspect mode to test load times. Because of this, load times were only tested on desktop browsers.
Edge: Press Ctrl+Shift+I to enter Inspect mode. Navigate to “Network.” Load tested page. Record time shown after “Time:” in seconds.
Vivaldi: Press Ctrl+Shift+I to enter Inspect mode. Navigate to “Network.” Load tested page. Record time shown after “Time:” in seconds.
Firefox: Load tested page. Right-click page. Select “inspect” from the menu. Navigate to “Network.” Re-load tested page. Record time shown after “Time:” in seconds.
Test Four Methodology
All browsers: Visit coveryourtracks.eff.org. Click “Test your Browser” with “Test with a real tracking company” checked. Record results.
Off Track – A Study © 2024 by Coviate is licensed under CC BY 4.0