Geospatial metadata is the structured "data about the data" that describes a spatial dataset: what it represents, where and when it was collected, its coordinate reference system, resolution, accuracy, lineage, and the meaning of each attribute field. It is what lets someone else (or future-you) use a dataset correctly without guessing.
Why it matters
A spatial file with no metadata is a liability. Without a documented CRS you cannot reproject it safely; without lineage you cannot judge whether it is fit for purpose; without attribute definitions the columns are just cryptic codes. Good metadata is the difference between a reusable data product and a one-off file that nobody trusts six months later. It also underpins discoverability in catalogues and data portals.
A concrete example
The international standard is ISO 19115 (with ISO 19139 as its older XML encoding), which structures metadata into sections such as identification, data quality, spatial reference, and distribution. In the US, the older FGDC Content Standard (CSDGM) is still encountered. At the file level, a GeoTIFF embeds CRS and extent in its header (readable with gdalinfo), and a shapefile stores its CRS in the sidecar .prj file—but neither captures lineage or attribute definitions, which belong in a separate metadata record.
Common pitfall
Confusing file-embedded spatial reference information (CRS, extent, pixel size) with full metadata. The .prj file or GeoTIFF header tells software how to place the data, but it says nothing about who made it, how accurate it is, or what the fields mean. A missing or wrong .prj is also a frequent cause of layers landing in the wrong place entirely.