Illustrative data: measurability is 80,000 / 100,000 = 80%. Viewability is 48,000 / 80,000 = 60%.
A low viewability percentage can reflect placement behavior, loading, or the population being measured. A low measurability percentage is a different signal: the system could not obtain the necessary measurement for part of the eligible inventory.
Use the standard relevant to the environment and the vendor's exact metric definition. Do not apply one desktop threshold to every video, mobile-app, large-format, or CTV situation.
Separate served, measurable, and viewable
A served or rendered event records a point in delivery according to the reporting system. Measurable means the measurement information was obtained. Viewable means the applicable visibility and duration requirements were met.
Viewability is not proof of attention, recall, a click, or a sale. Treat those as separate outcomes.
Calculate two different rates
Using Active View's definitions, viewability is viewable impressions divided by measurable impressions. Measurability is measurable impressions divided by eligible impressions. Multiply either result by 100 to express a percentage.
Illustrative example: 100,000 eligible impressions, 80,000 measurable, and 48,000 viewable produce 80% measurability and 60% viewability. Dividing 48,000 by 100,000 produces 48%, which describes a different ratio.
Interpret thresholds in their proper scope
The MRC desktop guideline uses minimum area and continuous-time requirements, with different treatment for certain formats. Google's Active View documentation describes common display and in-stream video criteria. Refer to the linked sources for the relevant conditions rather than treating a short definition as a complete specification.
A threshold is a counting rule. Improving the percentage should not involve obscuring content or creating placements that degrade the user's experience.
Investigate the affected segment
- Align format, device, placement, date range, and vendor definition.
- Check whether measurability changed before assessing viewability.
- Inspect rendering and measurement integration errors.
- Compare loading and scrolling behavior in the affected environment.
- Use a comparable observation window after an authorized change.
Distinguish two operational cases
Case A: measurability drops after an integration change while viewability among measured impressions stays stable. Investigate measurement coverage first.
Case B: measurability is stable but a newly introduced placement has low viewability. Inspect its actual rendering and visibility behavior. Do not infer that every unmeasured impression was non-viewable.
Report counts with the rates
Include eligible, measurable, and viewable counts alongside the percentages. This prevents an apparent improvement from hiding a smaller measured population.
If the vendor cannot measure a segment, record that limitation instead of assigning it an invented zero or extrapolating a precise result without justification.