
Measurement
Part of Programmatic campaign audits
Identifying unexplained differences between planned and delivered reach
Reconcile planned and delivered programmatic reach by checking scope, bought impressions, frequency and reporting limits.
To investigate a reach gap, first confirm the plan and delivery report cover the same population and metric. Match market, dates, included inventory and reach definition. Then check bought impressions, repeat exposure, changes to the buy and reporting limits. Two different headline figures alone do not prove a targeting failure.
Rebuild the plan and report
Keep the original plan, its date and assumptions: Australian geography, flight, budget, formats, supply, audience, frequency and planning tool. Check whether the brief recorded forecast unique reach, potential impressions or a target. These quantities are not interchangeable.
For delivery, request a combined reach estimate for the relevant buying scope and supported period. Record the exact metric: impression reach and viewable-impression reach answer different questions. Put bought impressions, spend and average frequency beside it.
Display & Video 360 provides reach and frequency metrics. Reported reach depends on the reporting period and data availability, so check freshness before comparing a recent extract with the plan.
How to Reconcile a Reach Gap
- Preserve the original plan detailsInclude Australian geography, flight dates, budget, formats, supply sources, audience targeting, and planning tool used.
- Clarify the planned metricConfirm whether the plan was based on forecast unique reach, potential impressions, or another target.
- Request combined delivery reportUse Display & Video 360 to get a unified reach estimate for the supported period and recording method.
- Record key delivery metricsInclude bought impressions, spend, average frequency, and exact reach metric (e.g., impression reach).
Trace the gap
- Scope:List markets, dates, buying IDs and environments included on each side. Identify additions, pauses and exclusions.
- Bought volume:Check whether fewer impressions were bought than the plan assumed. If so, investigate eligibility and auction losses where data permits.
- Exposure concentration:If impression volume is similar but reach is lower, inspect comparable frequency data. More repeat exposure can mean fewer newly reached people.
- Measurement:Check the model, supported date range, reporting delay and any unavailable slice. Treat an absent estimate as unknown, not zero.
Forecast figures are estimates based on their inputs and available market data, and should be treated as such rather than as observed outcomes.
Planned vs Delivered Reach: Key Factors to Compare
- Scope AlignmentMatch markets, dates, buying IDs, and environments in both plan and delivery reports.
- Bought ImpressionsVerify if fewer impressions were bought than planned; investigate auction losses or eligibility issues.
- Exposure ConcentrationCompare frequency data—higher repeat exposure may reduce unique reach.
- Measurement ConsistencyEnsure the same metric (impression reach vs viewable-impression reach) and reporting period are used.
Avoid a false total
Do not add ordinary daily or line-item unique reach figures: people can occur in several rows. Request a combined report where supported. Display & Video 360 offers overlap reporting, but that does not make arbitrary exports additive.
Check the actual report's accepted date range. For a longer flight, present supported periods separately unless a validated common method supplies full-flight deduplication.
State the planned figure, delivered estimate, aligned scope and remaining gap. Quantify the contribution of lower bought impressions or changed buying scope where the data supports it.
Keep frequency and model uncertainty visible, and assign the next evidence request rather than forcing the unexplained remainder into one cause.
Key Metrics for Reach Reconciliation
- Planned Unique Reach
- Forecast value from original brief
- Delivered Reach Estimate
- Reported figure from Display & Video 360
- Bought Impressions
- Actual impressions purchased
- Reporting Period
- Supported date range in report

