"Can you talk about the privacy and anonymity levels of these 4 scenarios?" Your thinking is going in the right direction here, but you should consider some de

"Can you talk about the privacy and anonymity levels of these 4 scenarios?" Your thinking is going in the right direction here, but you should consider some deeper analysis of your behavior ... In your example, 1.2B people access gmail - so far, so good. But how many people access gmail and HN ? How about gmail and HN and MeFi ? All during US timezone waking hours ? You can see where I am going with this. Now let's look at a different angle - you only visit gmail, so you're safely back in the 1.2B herd, right ? But how many people visit gmail with OSX ? With OSX vX.Y ? With your screen resolution, your enabled fonts, and your patch level? You might be very surprised at how unique your web hit really is.[1] The point being made here is that if you funnel all of your usage (traffic ?) everywhere, you create a usage "fingerprint" that can then be noticed or logged or tracked or correlated anywhere else you go. It's not hopeless - you could spin up a throwaway VM that you would only use for gmail and nothing else, but again - there are opsec considerations there that can sink the whole operation, just like above, if you aren't very, very careful. [1] https://panopticlick.eff.org/ https://panopticlick.eff.org/