Private Browsing Doesn’t Work That Way

I’m not certain that many people really understand what private browsing it – at least based on my interpretation data published by Mozilla Labs. The study is awesome and you should check it out.

I’ve been over what private browsing really is. It isn’t what I think what the general population thinks it is. (I don’t know what people are really thinking – but I can attempt to infer it by observing aggregate behavior.) Come with me and dive in.

The Mozilla Labs data indicates that private browsing activations spike at noon, with subsequent spikes around 5pm, 9pm, and shortly after midnight. The median duration of staying in private browsing is 10 minutes, with 50% of the cases falling between 4.5 and 22 minutes.

The data is absolutely fascinating. So just what’s going on? Why is it happening?

I think people are going into private browsing for such short bursts in an effort to evade surveillance.

Surveillance from who?

Employers, schools, and possibly family (in households with one limited number of computers). The noon spike certainly suggests some of that going on. And yet, so far as I know, private browsing doesn’t prevent most types of surveillance that are employed by IT departments at work and at school. Whatever is going on – even if it isn’t recorded in the browser – is certainly being recorded by the website (log files) and possibly by surveillance software installed by people who wish to surveil usage.

I’ve previously referred to private browsing as being porn mode. I think this holds true during the 10pm and post-midnight spikes. I certainly don’t think this holds at noon. (I hope not!).

I think there’s a false sense of security there.

The second point is whether or not free email service providers, subscription, and ecommerce sites experience massive increases in ‘first time visitors’ at these key times. In effect, with cookies in full suppression, it would appear to the web analyst that huge influxes of first time visitors are arriving. If you’re such an analyst – you might want to examine the period of time on your own site and see if they coincide.