4 ms·
The article hypothesizes that analytics services will probably move their tracking code from client-side JavaScript to the server-side, so that the publisher's
by majewsky 11y ago
The article hypothesizes that analytics services will probably move their tracking code from client-side JavaScript to the server-side, so that the publisher's server will forward the analytics data to the analytics server.
There is an interesting irony in this: By using client-side blocking more widely, we (the users) force evolution of analytics towards methods that are less introspectable and less blockable from our side.
- kuschku 11y agoIn many countries, we already made that illegal, and in the rest, NoScript will be the only solution.
- manigandham 11y agoWhat part is illegal? And if the data is captured server-side, how will NoScript help?
- kuschku 11y agoThe "tracking users without direct approval". And without JS tracking is very hard to do at all – a large part of the current tracking from Google depends on JS and/or Flash for detecting unique identifiers of the user.
- manigandham 11y agoDo you have a link to "direct approval" regulation? I'm curious as I haven't seen outright legislation regarding how server-side analytics would be limited. As far as JS tracking, you're right that it does allow for more fidelity in the information retrieved but the server also receives plenty of data that it can use. Many of the device fingerprinting techniques are actually workarounds due to the blocking of 3rd party cookies, but if the site you're visiting is collecting information server-side, then they can just set 1st party cookies which aren't blocked and track you that way. No need for fingerprinting. Even easier if sites have a registration system.
- kuschku 11y agoOne result of the regulation is the cookie law: You can only start tracking users after you have approval from them. You may not even set a cookie or store their IP before you have that approval. Direct approval is needed, and, in some cases (like insurances, banks, etc) your site even has to work without tracking.
- manigandham 11y agoThis is already happening. Google Analytics has a server-side API [1] that's used by some big sites now. The Intercept, a publisher focused on privacy, even setup a proxying system for their analytics (powered by parse.ly) recently. [2] 1. https://developers.google.com/analytics/devguides/collection/protocol/v1/ https://developers.google.com/analytics/devguides/collection... 2. https://theintercept.com/2015/11/04/what-the-intercepts-new-audience-measurement-system-means-for-reader-privacy/ https://theintercept.com/2015/11/04/what-the-intercepts-new-...
- rdtsc 11y ago> The article hypothesizes that analytics services will probably move their tracking code from client-side JavaScript to the server-side, From talking to several companies that specialize in this (via a friend privately, and by interviewing at some) that is already happening. Basically using all tricks you heard from https://panopticlick.eff.org/ https://panopticlick.eff.org/ are used, fonts, canvas, webgl fingerprints -- everything.
- manigandham 11y agoThat would be more related to device fingerprinting rather than just server-side analytics. You would still need client-side javascript to accomplish most of that, but yes it's also on the rise due to the (in my view somewhat unfair) collapse of the 3rd party browser cookie.
- whyever 11y agoAt least the performance of the websites will improve.
- RodgerTheGreat 11y agoServer-side tracking doesn't sound like that bad of an endgame to push towards. If users are behind a layer or two of NAT and browsers unilaterally stop leaking information in their user-agent strings and other optional HTTP fields (which are basically only used for tracking people and nasty brittle feature-detection hacks), server side tracking is easily crippled, and nothing of value is lost.
- manigandham 11y agoTracking (at least by most ad networks) isn't about knowing your exact identity online but just to separate different requests into uniques as best as possible. There's an exponential cost and very diminishing returns trying to get everyone's actual real identity - and this is actually why Google and Facebook are the biggest adtech players, because everyone has a login with them. The use of all these headers and user-agent info is for device fingerprinting which has matured because of the loss of the 3rd party cookie. If sites start using server-side analytics then they can just set 1st party cookies and get around the whole need for fingerprinting.