Question
What did the Google "Incognito mode" case reveal about user expectations versus tracking reality?
Answer
Google was found to collect data from users even in the supposedly private Incognito mode, and a court deemed Google's explanations misleading.
This landmark case exposed the gap between what users believe and what actually happens. Many people assume "private/incognito" browsing means they aren't tracked at all — but incognito mode only stops the local browser from saving history, cookies, and form data. It does not stop websites, advertisers, or the network from observing you.
What incognito mode does NOT hide:
- Your IP address (visible to every site and your ISP)
- Server-side logs and analytics
- Tracking via fingerprinting (which needs no stored state)
- Your employer's or school's network monitoring
Tip: Incognito is about local privacy from other people using your device, not anonymity from the websites you visit. The court's ruling hinged on this discrepancy being misrepresented.
Go deeper:
Evercookie (Wikipedia) — persistent tracking that survives deletion, why private browsing != anonymity.
Note saved — thanks!
Question
What is the difference between deterministic ("eindeutiges") and probabilistic ("wahrscheinliches") tracking?
Answer
Deterministic tracking identifies a user explicitly (via login/account), while probabilistic tracking infers identity from signals like IP, browser data, and behaviour using probability models.
| Deterministic | Probabilistic | |
|---|---|---|
| Identification | Explicit (login, account ID) | Inferred via probability models |
| Login needed? | Yes | No |
| Signals | Account IDs linked across devices | IP, browser data, surfing behaviour |
| Accuracy | Exact | High, improved by machine learning |
Deterministic is rock-solid because the user authenticated — e.g. logging into the same account on your phone and laptop lets a company link both devices to one person ("personalized offers through account IDs").
Probabilistic never needs a login: it correlates fingerprints, IP ranges, and behaviour patterns, then uses ML to decide "these two sessions are probably the same person." It's the backbone of cookieless tracking.
Note saved — thanks!