Measurement
Part of Programmatic campaign measurement
Evaluating programmatic reach without double counting users
Use campaign-level deduplicated reach, understand overlap across report rows, and state identity, reporting-window and modelling limits.
Use a deduplicated unique-reach estimate for the combined campaign and a supported reporting period. Do not add reach across line items, devices or days: the same person can appear in several rows. Record the inventory and identity method covered; if the report does not disclose the identity basis, mark it as undisclosed rather than infer it. Unique reach is a modelled estimate within a measurement system, not a complete count of Australians exposed.
Separate exposures from people
Impressions count delivered ad opportunities and can generally be added across non-overlapping report rows that use the same definition. Reach estimates distinct people. If one person encounters two line items, each item may include that person in its reach. A combined reach estimate should count the person once.
Consider an illustrative calculation: line item A reaches 40,000 people and B reaches 30,000. If a reliable common method identifies 10,000 people in both groups, their union is 40,000 + 30,000 − 10,000 = 60,000. Without a trustworthy overlap estimate, the two line-item figures cannot establish that total. These numbers are arithmetic, not a measured campaign.
Impressions vs. Reach: Key Differences in Programmatic Measurement
- Impressions
- Count delivered ad opportunities; can be added across non-overlapping report rows.
- Reach
- Counts distinct people; must deduplicate to avoid double counting the same user.
Request a report at the decision level
Request reach for the combined campaign, intended geography, included inventory and relevant period. State clearly which impression definition the metric uses. Use matching dates when reading it beside impressions and spend. Include the reporting interval in the request rather than assuming daily reach figures can be added.
Display & Video 360 Help has pages titled “Reach reports” (https://support.google.com/displayvideo/answer/6170584?hl=en) and “Measuring reach and frequency” (https://support.google.com/displayvideo/answer/15272050?hl=en). Use them to check how reach is defined and which intervals the relevant reporting method supports.
Ask for campaign-level unique reach, but verify the exact metric label in the report: the page titles do not establish a UI field name or report template. The information available here does not specify a deduplication method, identity source or fixed interval length, so do not supply any of these by assumption.
Do not add estimates from separate reports unless a validated method supports that calculation. If the flight exceeds the period the reporting method supports, present the supported periods separately and leave the full-flight unique total unknown unless another validated method supplies it.
How to Request a Valid Campaign-Level Reach Report in Display & Video 360
- Step 1Request reach for the combined campaign, specifying geography, inventory, and reporting period.
- Step 2Confirm the impression definition used in the metric (e.g., viewable impressions).
- Step 3Use matching dates when comparing reach with impressions and spend.
- Step 4Verify the exact metric label in the report—do not assume it matches page titles.
- Step 5Do not sum daily or line-item reach unless a validated method supports it.
State the estimate's limits
Display & Video 360 Help provides reach reporting documentation. Review the relevant page for how a reach estimate is produced and for the identity and modelling inputs it uses.
Reach estimates may not be available in every report. Availability can depend on the reporting method and the data available for the request.
A missing estimate is not zero people. Consider a broader reporting slice if it still answers the brief. Do not fill the gap by summing daily or line-item reach.
Do not assume a fixed number of supported days: use the interval shown for the report. If the interval is not shown, keep supported periods separate and leave the full-flight unique total unknown unless another validated method supplies it.
If several platforms are involved and there is no validated method to deduplicate them, report each platform's reach separately and label combined unique reach unknown.
Key Limits of Reach Estimates in Programmatic Reporting
- Identity Source
- May be undisclosed—do not infer without evidence.
- Reporting Interval
- Not fixed—use the interval shown in the report.
- Platform Overlap
- If no validated deduplication method exists, combined unique reach is unknown.
- Missing Estimate
- Not zero—consider broader reporting slices instead.
Use the figure for a decision
Read estimated combined reach beside impressions and the campaign outcome. If impressions grow while new reach slows, inspect a comparable frequency report before deciding whether exposure is concentrating. Keep frequency-cap choices with the site's frequency guidance. Record the model, dates, included inventory and unavailable slices so the next reader can interpret the reach figure.
Best Practices for Using Reach Figures in Campaign Decisions
- Compare reach with impressions and campaign outcomes.Look for trends: rising impressions but flat new reach may indicate concentration.
- Review frequency reports before adjusting frequency caps.Align with site’s frequency guidance to avoid overexposure.
- Document model, dates, inventory, and unavailable data slices.Ensure transparency for future interpretation by team members.
