Methodology

Key concepts and definitions

Exit

An exit is the act of one user leaving one platform. It is a discrete event, recorded once, with a sequential ID and timestamp.

Each exit record includes:

  • • The platform being exited
  • • The reason for exiting (from standardized taxonomy)
  • • The replacement (what the user is using instead, or “Nothing”)
  • • A sequential ID (append-only, no gaps except for deletions)
  • • A timestamp

Exits are permanent and cannot be edited. Only a full account deletion can remove an exit from the ledger.

Replacement

A replacement is what the user adopts after exiting a platform. It can be another digital tool, an analog solution, or nothing at all.

Examples:

  • • Exited Slack → Replaced with Discord
  • • Exited Notion → Replaced with Obsidian
  • • Exited Instagram → Replaced with Nothing
  • • Exited Netflix → Replaced with Books

“Nothing” is a valid replacement. It signals full abandonment of that category of tool.

Stack

A user's stack is the set of platforms they are currently “in” — the tools they actively use and have not exited.

The stack is tracked over time. When a user exits a platform, it is removed from their stack. Replacements are optionally added.

The stack serves as a mirror: users can see their digital footprint shrink (or grow) over time. Nightly snapshots preserve the state of the stack at regular intervals.

This creates a personal audit trail of software usage.

Immutability

EXIT operates on an append-only model. Records cannot be edited or deleted by users.

Why:

  • • Preserves data integrity for researchers and analysts
  • • Prevents retroactive manipulation of migration trends
  • • Creates a permanent cultural archive
  • • Ensures sequential IDs remain meaningful

If an exit record contains an error (e.g., wrong platform), corrections are handled through annotation or flagging, not deletion. Administrators can merge duplicate platforms or add aliases to improve data quality.

Users can delete their entire account, which removes all associated exit records. This creates gaps in the sequential ID sequence, which is expected and documented.

Deduplication and Aliases

Users may refer to the same platform by different names:

  • • “Gmail” vs “Google Mail”
  • • “Twitter” vs “X”
  • • “Adobe Photoshop” vs “Photoshop”

Administrators can merge duplicate entries and assign aliases to canonical platform names. This improves data consistency without altering historical records.

All platform pages support aliases, so searches for “Twitter” and “X” return the same aggregated data.

Append-Only Corrections

If a user reports an error in their exit record, the original record remains unchanged. Instead, a correction note is appended.

Example:

Exit #00247: User exited Slack → Discord

Correction: Replacement should be “Teams” not “Discord”

This preserves the historical record while acknowledging the error. Think of it like accounting: you do not erase ledger entries, you add correcting entries.

Data Quality

EXIT maintains data quality through:

  • • Rate limits to prevent spam
  • • Bot traps during submission
  • • Admin review of flagged records
  • • Platform merging and alias management
  • • Public visibility (community oversight)

All exit records are public by default. This transparency creates natural accountability and allows researchers to filter or exclude low-quality data if needed.

Pseudonymity

Users are pseudonymous by default. Exit records are associated with a username, not a real name or email address.

Users can optionally add a display name or link their profile to external identities, but this is not required.

All data published in aggregated reports is fully anonymized. Individual exit records remain pseudonymous (username only).

These definitions ensure EXIT remains a reliable, citation-ready dataset for individuals, researchers, and businesses.