Campaign Audits
Programmatic campaign audits
Compare a programmatic campaign with its brief, buying settings, delivery and outcomes, then assign decisions for material findings.
A programmatic campaign audit compares the approved brief with the settings that governed buying, the delivery that resulted and the evidence behind reported outcomes. It identifies material differences and decides what to correct, investigate or change in the next brief.
Establish the record
Collect the approved brief and amendments, buying IDs, dated settings records, delivery exports, change history and outcome report. Record the currency, cost definition, flight dates, report extraction time and named time zone. For a campaign spanning Australian states, “Australian time” is too vague for daily reconciliation.
Separate fixed requirements from preferences the trader was authorised to adjust. Record who approved each material change. If the relevant saved state is missing, a campaign name or current setting cannot establish what applied throughout the flight.
Interpret platform state
In Display & Video 360, campaign status is relevant evidence about serving eligibility: active campaigns are eligible, while paused campaigns are not, even when their insertion orders and line items are active. Archived campaigns are automatically paused. A campaign marked completed has reached its end date, but may still serve if its insertion orders and line items remain active.
When reviewing a discrepancy, compare status records at campaign, insertion-order and line-item level for the affected period. A completed label alone does not establish that serving stopped, and campaign flight dates do not control serving.
Work through four questions
| Question | Compare | Record |
|---|---|---|
| Was the buy configured as approved? | Brief, dated settings and changes by buying ID | Matches, authorised changes, mismatches and gaps |
| Where did the budget go? | Spend by buying ID, source and period | Material spend requiring review |
| Did delivery meet the plan? | Comparable planned and delivered measures | Variance and evidence for possible causes |
| What follows? | Findings, limits and available options | Decision, owner and review trigger |
Inspect the settings that actually govern serving. In Display & Video 360, campaign-level planned spend and dates are planning fields; they do not limit serving. Insertion-order budgets and dates apply to their line items. Other buying platforms may organise these controls differently.
Separate observations from explanations
If spend is low, check supply, targeting, creative, frequency, budget and pacing before considering bids and auction losses. Display & Video 360’s Troubleshooter can help analyse why a deal is not spending or a line item is not winning available inventory. Detailed data may take longer to become available.
If spend is on plan but the outcome is weak, inspect the placement and creative mix, destination, event definition and reporting maturity. An attributed conversion is an association under specified rules, not proof that the ad caused the action. Do not label all spend on a weak-looking placement waste without a defensible outcome measure and enough observations for the decision.
For a reach variance, compare like definitions and periods. Impressions are not people. Use a combined reach estimate for a supported period where available, and label missing or recent figures accordingly.
Top 5 Reasons for Low Spend in Programmatic Campaigns (Australia)
- Supply constraintsLimited available inventory in target geographies or formats
- Targeting misconfigurationsOverly narrow audience segments or incorrect geo-targeting (e.g., excluding major cities like Sydney or Melbourne)
- Creative issuesNon-compliant or low-performing creatives rejected by publishers
- Frequency cappingToo aggressive frequency limits reducing overall reach
- Bid strategy inefficienciesBids too low relative to market rates, especially in premium placements
Test report evidence before drawing conclusions
In Display & Video 360’s Troubleshooter, total available requests, impressions won and win rate data are available within one hour; detailed data such as targeting and creatives may take longer. The tool presents data in the advertiser’s time zone, so compare that setting with the time zone recorded in the audit before matching daily figures.
The Troubleshooter can show quantified reasons for lost impression availability at three levels. Its filtered-impression reasons include pre-bid issues, invalid traffic, unauthorised sellers and domains that could not be crawled for brand safety. Treat these as diagnostic evidence for the selected inventory and period, not as a complete account of every delivery variance.
Native creatives are not supported by the Troubleshooter, and detailed breakdowns may be delayed or unavailable. Record the tool’s date range and inventory selection alongside any finding.
Pros and Cons of Using Display & Video 360’s Troubleshooter for Audits
- ProsProvides real-time insights on win rate, available requests, and impressions won within one hour; includes diagnostic reasons for lost inventory at pre-bid, invalid traffic, and brand safety levels.
- ConsDoes not support native creatives; detailed breakdowns may be delayed or unavailable; filtered-impression reasons are not exhaustive across all platforms.
Validate the outcome definition
For Floodlight outcomes, record the conversion window and counting method used in the report. Floodlight can count every conversion, the first conversion for a unique user during each 24-hour day, one conversion per user per session, transactions and revenue, or items sold; totals are not comparable unless the counting rules align.
Check whether reported conversions are click-through or view-through and whether the applicable conversion window is consistent with the reporting period. Floodlight gives clicks precedence over impressions to avoid double-counting when a user clicked within the click conversion window. Different platforms can report different conversion totals because their attribution models and tracking methods differ.
Check reach measurement availability
Display & Video 360 unique reach uses statistical models to estimate people across browsers, devices, formats and networks, including co-viewing on connected TV. It measures estimated unique users rather than cookies, so label it as modelled reach rather than a direct count of impressions or devices.
Most reach and frequency metrics can only be reported for date ranges of 92 days or less. Data may also be unavailable because of country coverage or minimum thresholds for impressions and unique users; record those limits before treating a missing figure as a delivery failure.
Key Metrics and Limits in Display & Video 360 Reach Reporting
- Reach Measurement Type
- Modelled unique users (not cookies)
- Maximum Date Range
- 92 days or less
- Minimum Thresholds
- Impressions and unique users required for data availability
- Coverage
- Includes connected TV and cross-device co-viewing
Close with decisions
For each material issue, state the requirement, observed record, affected period and amount or delivery, best-supported explanation and remaining uncertainty. Assign an action and the evidence that would show it was completed. A missing required exclusion may call for a settings correction and approval review. Low delivery under a required constraint may instead call for a revised spend expectation.
Keep the final decision register short: keep, correct, investigate or change the next brief. Preserve unresolved questions as questions; a before-and-after pattern alone does not establish causation.
Where required evidence is missing, or diagnostic detail is delayed, unsupported or absent, mark the issue unverified and set a review trigger rather than presenting an inference as a confirmed cause.
In this guide
- Checking a campaign's settings against the approved briefCompare approved campaign requirements with dated buying settings, inherited controls and change history.
- Reviewing wasted spend before the next campaignClassify possible wasted programmatic spend, check its cause and make a bounded decision for the next campaign.
- Identifying unexplained differences between planned and delivered reachReconcile planned and delivered programmatic reach by checking scope, bought impressions, frequency and reporting limits.
- Writing an actionable programmatic campaign post-mortemTurn campaign results, setting changes and unresolved findings into a concise post-mortem with owners and closure checks.
