Privacy

• Pages visited, and which Archive wing they belong to
• Investigations opened and research depths selected
• Documents opened, scroll milestones reached, and approximate reading engagement
• PDF downloads
• Searches made within the Archive, and how many results they returned
• Use of Random File
• Movement between related investigations
• A general traffic-source category and, where available, the referring website
• Whether the device is broadly mobile or desktop

These measurements show which investigations interest readers, which depths they choose, how far documents are read, which searches succeed or find nothing, how readers discover and move through the Archive, and where gaps may be worth investigating in the future.

Terms entered into the Archive's own search may be retained as part of its anonymous analytics. They reveal what readers are looking for, which searches return nothing, and which subjects might deserve a future investigation.

Stored search terms are limited in length, and apparent email addresses, web addresses, and long numerical identifiers are redacted. No automatic process can recognise every form of personal information, so please do not enter personal or sensitive details into Archive search.

The Archive estimates approximate daily unique visitors using a one-way identifier derived from limited connection and browser information — including the visitor's IP address, browser type, and the date. The IP address itself is not stored as analytics data. The identifier changes each day, is used only to estimate aggregate daily readership, and is not intended to identify a person or follow anyone across days. The count is an estimate, not an exact figure.

A visit is tracked within the current browser tab and ends after about 30 minutes of inactivity. Its identifier is held within that tab, not in an analytics cookie.

The Archive also measures, in aggregate, how far readers progress through documents and approximately how long they spend actively reading. Inactive or background tabs are not intended to count as reading. This is used to understand engagement with the research, not to identify readers.

The analytics system does not intentionally collect or keep:

• Names or email addresses
• Visitor accounts
• Advertising identifiers or advertising profiles
• Precise geographic location
• Keystroke histories
• Cross-site browsing profiles
• Stored IP addresses as visitor records
• Browser fingerprints intended to identify people

The Archive does not use its analytics to build profiles of individual readers.

The Archive's analytics do not use cookies, advertising trackers, or third-party analytics services. Analytics are stored in the Archive's own database. They are not collected for advertising and are not provided to advertising or analytics companies.

The analytics respect the Do Not Track and Global Privacy Control browser signals. Visitors whose browser sends either signal are excluded from Archive analytics.

The Archive may note in general terms whether a visit arrived directly, from Google or another search engine, from social media, or from another website. This is used only to understand how readers find the Archive; it does not reveal the search queries people typed into Google.

Separately, the Archive uses Google Search Console to understand how its public pages appear and perform in Google Search. Search Console is independent of the Archive's own reader analytics.

Archive analytics are available only through the private administrative system and are used to understand aggregate readership and improve the Archive. There is currently no fixed automatic deletion period; analytics records are kept until they are removed by the Archive. This page will be updated if that changes.