Drift is the gradual, often unintended divergence of a system, document, or process from its original intended state, design, or specification. It rarely happens through a single event. It accumulates through small, disconnected changes over time. None of which individually seem significant enough to trigger a review or correction.
In documentation contexts, drift describes the process by which content becomes inaccurate over time. A product evolves incrementally. Terminology shifts. Flows get replaced. Architecture changes. The documentation doesn’t keep up, not because anyone decided to neglect it, but because no system existed to prevent neglect. The two are equivalent in outcome.
Drift is most common in evergreen content — foundational material like architecture overviews, onboarding guides, and concept documentation that was written during a period of stability and then became an orphan: no owner, no update trigger, no audit schedule.
Why This Matters for Technical Writers
Drift is one of the most consequential failure modes in a knowledge system, and one of the least visible. A stale parameter name fails loudly, such as when the API call breaks and the user knows something is wrong. A stale mental model fails silently, such as when the user builds their understanding around a version of the product that no longer exists and proceeds confidently in the wrong direction. By the time the error surfaces, it looks like a user mistake rather than a documentation failure. Recognizing drift as a structural problem, not a behavioral one, is the starting point for addressing it through governance, ownership, and audit processes rather than individual effort.
Common Confusion
Drift is often treated as a content quality problem, something to fix through better writing or more careful review at publication. In practice it is a systems problem. Content that is accurate at publication can drift without anyone making an error. The mechanism is the absence of a process to detect and respond to change over time, not a failure at the point of creation.
Related Terms
Evergreen content, content lifecycle, knowledge debt, DDLC