A baseline is a documented starting condition
Record definitions, collection scope and a completed reporting period before judging improvement. Without a stable baseline, changed targets and changed data can be mistaken for changed performance.
A useful baseline fixes the metric definition, source, eligible population and time period. It also records what could not be measured. Without that agreement, a later team may celebrate a change caused by a new filter, repaired integration or larger measurement scope rather than better marketing performance.
A worked example
In a hypothetical baseline month, a supplier measures 800 relevant organic clicks and 18 unique inbound inquiries. The following month adds visitor-identification data. The new visitor records expand the available intelligence, but they must not be added to the inquiry count as though they were equivalent requests for contact.
How to apply it
- Save a completed-period export or dated report with definitions and exclusions.
- Record known tracking gaps, tests and changes in business capacity.
- Separate future collection repairs, target changes and raw performance movements in the comparison.
What the result can establish
The baseline is accepted when another reviewer can reproduce its count and explain its boundaries. In SharkBlade, retain the distinction between raw KPI values, category assessments and evidence coverage. Downstream sales or revenue conclusions need their own verified records.
Reference: Google: Search Console performance report.
Supporting documentation
Google: Search Console performance reportGoogle Analytics: recommended eventsGoogle: helpful, reliable content