A dataset is a snapshot, not a promise that the world stopped changing

Every dataset has an invisible timestamp, even when there is no date column on the screen.

Businesses move. Websites change. Products disappear. Prices shift. People change roles. Public records are updated on their own schedules.

That means a dataset can be accurate when it is assembled and still become less accurate over time without anyone making a mistake.

This is easy to forget because spreadsheets feel static. Once the rows are organized and the columns are clean, the file can look finished.

The world behind it is not finished.

For research work, recency should be part of the meaning of the data. A source-checked date, retrieval date, reporting period, or another clear time reference can tell the next user what kind of snapshot they are looking at.

That becomes especially important when several sources are combined. One field may come from a recently updated company page while another comes from an older public record. Both may be valid within their own context, but they do not necessarily describe the same moment.

The goal is not to make every dataset real-time. Many useful datasets never need to be.

The goal is to avoid confusing “accurate when observed” with “true forever.”

Good data carries enough time context for the next person to know the difference.