Skip Navigation

Google said it was too hard to target ads to people

Came across this article and it got me thinking, are there any simple ways to defeat advanced tracking methods (fingerprinting, tracking pixels, etc.)?

Obviously you could go the Tor on a virtual machine route, or a non persistent set up like TAILS, but what about a browser that's able to give say, a 80% solution?

I work in the security industry and am always looking for the solution that is simple enough that its palatable to a client (not asking to change your whole lifestyle, just push this button) but also relatively effective.

59 comments
  • Meanwhile the linked website is full of intrusive ads and hundreds of "legitimate interest" tracking cookies. Oh the irony...

    • uBO exists.

      • You don’t get the point, do you? I know I can block those, yet it’s hypocritical to complain about privacy and tracking in an article while doing the same. It’s not even the fact they use cookies at all, I get they might need them for analytics and such. But this site is out of control

  • I personally like Mullvad's approach to something like this.

    First, use their browser and VPN together (browser was co-developed with the Tor Project folks),

    In the VPN, you want to turn on DAITA. It's an interesting concept and I hope more legit projects like Mullvad start doing these things.

    They're essentially adding bunk data to your VPN traffic to hide you from any AI analysis that might use only your throughput to identify you and your habits.

    • try creepjs with mullvad browser, 100% traceable. it will always know it was you even if you clean the identity and restart the browser. and I bet googles tracking is even more advanced.

      • Can you elaborate? When I'm using Mullvad Browser+VPN, have DAITA and Multi-hop on, it doesn't know who I am at all.

        Since this is a VPN, there are a ton of visits with this FP ID, and the FP ends up calculating differently (and I get different visits results, trust scores) whenever I refresh my session in the browser, or even just reconnect the VPN.

        The other data on the page are all completely generic guesses at my system, monitor size, etc. and maybe 10% of that info is accurate to my system. Even that info is not very useful. For instance it says I'm running "Linux x86_64"... they certainly nailed that information down...


        When I do this with only the VPN and Firefox, then the data is a lot more consistent between refreshes, incognito mode, etc. and the FP ID is pretty much the same every time in Firefox.

        The other data taking guesses at my system are also more accurate when using regular ol' Firefox. For instance, it actually adds to the "Linux x86_64" that I am using an AMD GPU (no additional info than brand). Still not all that damning if it wasn't for the FP ID in this scenario.


        I've read through the docs, and several other articles, that explain more about creepjs, but I culd be misunderstanding something somewhere I guess.

        ETA: I'm also noticing that in regular Firefox, the timezone data is all fairly accurate to the current servers my VPN is hopping through. In Mullvad Browser, though, the timezone data is all over the place and not at all accurate to what my VPN is set to, let alone where I actually am.

        ETA2: maybe my settings are more specific than you expect? Maybe your data about being 100% traceable is with 0 configuration of the browser or VPN?

        My setup:

        • Mullvad Browser + Mullvad VPN
        • DAITA turned on
        • Multi-hop turned on
        • Lockdown mode on
        • All DNS content blockers enabled
        • Extra steps to unify VPN+Browser DNS compatibility

        I could see if maybe you just installed Mullvad VPN and didn't use their browser (or didn't configure the browser for the VPN) that you'd be way more traceable.

  • i use uMatrix (by the same author as uBlock Origin), which essentially allows very granular control over what dynamic content to allow:

    per domain and subdomain you can allow script, xhr, media, frames, cookies, images, css, and other things

    so you can say, for example, on lemm.ee deny any scripts from google.com from loading and deny any xhr (so analytics can’t work even if the script is hosted on the sites own domain)

    this stops a lot of fingerprinting in its tracks (except when you need to allow eg reCAPTCHA), but it does break pretty much every website until you go and allow only known good things (like scripts and xhr to the sites own domain)

    there’s also server-side fingerprinting, which is harder again

    • This looks to be an excellent tool, thanks for sharing and have a good one.

      • you’re welcome! you too!

        it should be noted though that it hasn’t been updated since 2021, and its repo has been archived (i’m not sure of the reasons). it still works great, but it’s not going to get any updates

  • I will keep my adblocker until contextual ads are the norm again.

59 comments