Only about 4% of all browsing sessions happen in private or incognito mode, according to a Carnegie Mellon University study that analyzed 167,128 real browsing sessions contributed by 451 Windows computer users between October 2014 and December 2017 (published at USENIX SOUPS 2018). A 2023 real-world traffic sample of 45,000 site visitors puts the figure a little higher, at 5.8%, but the two independently collected numbers land in the same narrow band.
That gap between how rarely private mode is actually used and how often people assume it hides their activity from everyone is the real story here. Below is what the academic and traffic data show about how often incognito mode gets used, who uses it most, and how badly its protections are misunderstood.
How many browsing sessions are actually private?
Private browsing occurred in just 4% of the 167,128 browsing sessions Carnegie Mellon University researchers observed through the Security Behavior Observatory, a longitudinal panel that instruments participants' home computers to record real browsing activity rather than relying on self-reported survey answers. The study's authors concluded that people switch into private mode to complete a specific task rather than browsing the open web in it by default.
A separate, independently collected number lands close by. Quantable Analytics measured 45,000 website visitors in 2023 using a browser-detection script and found that 5.8% of all traffic came from private or incognito windows. The two figures were gathered years apart using completely different methods, one from instrumented desktop panels, one from live site traffic, which makes the agreement more convincing than either number alone.
| Source | Method | Sample | Private share |
|---|---|---|---|
| Carnegie Mellon University (SOUPS 2018) | Instrumented browsing panel, sessions counted directly | 451 users, 167,128 sessions | 4% |
| Quantable Analytics (2023) | Live traffic detection script across a real site | 45,000 visitors | 5.8% |
Figure 1: Two independently collected measures of private-browsing traffic share, four years apart. Source: Carnegie Mellon University SOUPS 2018 paper, Quantable Analytics 2023 traffic sample.
Which browsing segments see the most incognito traffic?
Private browsing is not evenly distributed across content or device types. Quantable Analytics' 2023 traffic sample found adult content sites recorded the highest private-mode share of any tracked segment, at 10.3% of traffic, nearly double the 5.8% web average measured across the same dataset. Logged-in users showed the lowest rate, at just 3.6%, which makes sense since a signed-in session already carries an identity the private window cannot hide.
Figure 2: Private-browsing traffic share across seven tracked segments from the same dataset. Source: Quantable Analytics, 2023 traffic sample of 45,000 visitors.
Desktop and iOS carry noticeably more private-mode traffic than mobile in general and Android specifically, a gap wide enough that any site-level privacy analysis needs to account for device mix before comparing incognito rates across properties. If your own analytics show a private-mode share far above or below these figures, treat it as a signal about your specific audience and content category rather than a measurement error, since the source dataset itself found wide swings by segment.
How many people have ever used private browsing?
46% of US adults say they have used private browsing on a desktop computer at least once, and 43% say they have used it on a mobile device, according to DuckDuckGo's 2017 survey of 5,710 US respondents recruited through SurveyMonkey. That "have you ever" figure sits well above the 4% to 5.8% session-level share, which makes sense: most people who have tried private mode use it occasionally for a specific task rather than as their default browsing state.
Figure 3: Share of DuckDuckGo's 2017 survey respondents who reported ever using private browsing on desktop. Source: DuckDuckGo, "A Study on Private Browsing," 2017, n=5,710.
An earlier 2017 Google Consumer Surveys study, run by Google security researcher Elie Bursztein with at least 200 US respondents per question, found a similar pattern from a different angle: 61% of respondents were aware private browsing existed (up from roughly 51% in 2012), but only about 35% of those aware of it actually used it. Awareness of the feature has grown faster than adoption of it, which tracks with how rarely private mode shows up as a share of actual traffic. That same survey found 77.7% of private-browsing use happened at home, versus 36.3% in public places and 22% at work, undercutting the assumption that incognito mode is mostly a workplace-discretion tool. If your site handles any of the account, health, or financial data these users are trying to keep private, a clear privacy policy generator that spells out what your own tracking scripts still capture in a private window is one of the few things actually within your control.
Do people understand what incognito mode actually protects?
Not well. A 2018 study from the University of Chicago and Leibniz Universitat Hannover tested 460 participants across 13 different versions of the in-browser disclosures that six major browsers show when a private window opens. The researchers found that these disclosures "fail to correct the majority of the misconceptions we tested," including beliefs that private browsing blocks geolocation, prevents ads, protects against viruses, and stops tracking by the websites visited or by network providers like an employer, school, or ISP. Some disclosure wording even increased specific misconceptions rather than reducing them.
Figure 4: What private browsing actually protects against versus what most users assume. Source: Wu et al., "Your Secrets Are Safe," WWW 2018 (460 participants).
DuckDuckGo's 2017 survey reached the same conclusion from the user side: two-thirds of respondents overestimated the privacy protections private browsing offers, most commonly believing it prevents tracking by websites and advertisers or hides searches from search engines. Neither is true. Private mode only clears what the browser stores locally; it does nothing to a website's own server-side logs, and it does nothing to a network operator watching traffic pass through. The average site still loads dozens of third-party tracking scripts inside a private window exactly as it would outside one, since most trackers do not check whether the tab is in private mode before firing.
How long does a typical private session last, and when do people use it?
A median private browsing session lasts about 10 minutes, with a quarter of sessions ending inside 4.5 minutes and a quarter running past 22 minutes, according to Mozilla's Test Pilot research on Firefox 4 Beta users. That research, one of the earliest large-scale looks at real private-browsing behavior, also identified four recurring daily peaks in when people open a private window: around lunch, right after work or school, in the evening, and a smaller spike late at night.
Figure 5: Milestones in private-browsing adoption and the research measuring it. Source: see Where the Numbers Come From below.
The lunchtime and after-work spikes line up with the task-based pattern Carnegie Mellon's session data also found: people are not living inside private mode, they are stepping into it briefly for something specific, then closing the window. That short-session pattern is consistent across a 2010 dataset and a 2023 one, even though the browsers themselves have changed considerably in between.
The Bottom Line
Every dataset here points the same direction: private and incognito browsing is a narrow, task-based habit, not a default browsing state, and it protects far less than most people assume. Session-level measurement puts real usage at 4% to 5.8% of all traffic, while "have you ever tried it" self-reports run closer to half of US adults, a gap that only makes sense once you see how briefly and specifically people actually use the feature. The bigger risk sits in the understanding gap rather than the usage gap: two independent studies, four years apart, both found that browser disclosures do little to correct the belief that private mode hides activity from websites, advertisers, or network operators. Related questions worth tracking in this same cluster, including how many users actively opt out of ad tracking and what share of overall web traffic is automated bot activity rather than private human sessions, follow the same pattern of self-reported belief running ahead of measured reality.
Frequently Asked Questions
What percentage of browsing sessions happen in incognito mode? About 4% of all browsing sessions, according to a Carnegie Mellon University study that analyzed 167,128 real browsing sessions from 451 Windows computer users between October 2014 and December 2017. A separate 2023 traffic sample of 45,000 site visitors, run by Quantable Analytics, found a similar figure of 5.8%.
How many people have ever used private browsing? 46% of US adults say they have used private browsing on a desktop computer at least once, and 43% have used it on a mobile device, according to DuckDuckGo's 2017 survey of 5,710 US respondents. Two-thirds of those respondents overestimated the privacy protections private browsing actually provides.
Does incognito mode hide your activity from your employer or internet provider? No. Incognito and private browsing modes only stop the browser itself from storing history, cookies, and form data on the local device after the window closes. A 2018 University of Chicago and Leibniz Universitat Hannover study of 460 participants found that browsers' own in-app explanations fail to correct most users' misconceptions about geolocation blocking, ad prevention, and network-level tracking by an employer, school, or internet service provider.
Is incognito mode usage higher on desktop or mobile? Desktop. Quantable Analytics' 2023 traffic sample found private-mode sessions made up 8.2% of desktop traffic compared with 5% of mobile traffic, and 7.7% of iOS traffic compared with just 1.9% of Android traffic.
Where the Numbers Come From
- Habib, H. et al., Carnegie Mellon University. (2018). "Away From Prying Eyes: Analyzing Usage and Understanding of Private Browsing." USENIX SOUPS 2018. Private browsing occurred in 4% of 167,128 sessions from 451 participants, data collected October 2014 to December 2017.
- Quantable Analytics. (2023). "How Many People Use Incognito Mode?" Traffic sample of 45,000 site visitors, October 2023. Overall 5.8% incognito share; desktop 8.2%, mobile 5%; iOS 7.7%, Android 1.9%; NSFW content 10.3%; logged-in users 3.6%.
- DuckDuckGo. (2017). "A Study on Private Browsing: Consumer Usage, Knowledge, and Thoughts." Survey of 5,710 US respondents. 46% had used private browsing on desktop, 43% on mobile; two-thirds overestimated its privacy protections.
- Bursztein, E., Google. (2017). "Understanding How People Use Private Browsing." Google Consumer Surveys, at least 200 US respondents per question. 61% awareness (up from about 51% in 2012); 35% of aware respondents used it; 77.7% of use occurred at home.
- Wu, Y., Gupta, P., Wei, M., Acar, Y., Fahl, S., Ur, B., University of Chicago and Leibniz Universitat Hannover. (2018). "Your Secrets Are Safe: How Browsers' Explanations Impact Misconceptions About Private Browsing Mode." WWW 2018. 460 participants across 13 browser disclosure versions; disclosures failed to correct most tested misconceptions about geolocation, ads, viruses, and network-level tracking.
- Mozilla. (2010). "Understanding Private Browsing." Test Pilot "A Week in the Life" study of Firefox 4 Beta users. Median private session length about 10 minutes; usage peaks at lunch, after work, evening, and late night.
Note: All figures verified as of September 2026. The Carnegie Mellon and Quantable figures are the two primary, independently sourced measures of private-browsing traffic share; the 2010 Mozilla data on session timing is older but remains the most-cited public dataset on private-browsing session duration and has no widely available public replacement. This post will be refreshed if a newer large-scale traffic or panel study on private-browsing share is published.
