Campaign Setup
Programmatic campaign optimisation
Review programmatic results, diagnose delivery constraints, compare segments and record changes against the agreed campaign outcome.
Optimise a live programmatic campaign by comparing results with the agreed outcome, finding the cause of any gap and making a reviewable change. Extra spend or cheaper impressions help only if they serve the campaign goal.
Start with the decision
At each review, put the primary outcome beside spend, delivery and the remaining flight. Use the approved event definition, cost basis, date range and a named reporting time zone. For an Australian campaign, “Australian time” is too vague. Label recent conversion figures provisional if actions or reporting may still arrive.
Ask what the evidence could change. A delivery shortfall calls for an eligibility or auction check. Weaker attributed outcomes with steady delivery may call for a placement, creative, landing-page or measurement review. A promising inventory segment may justify a limited allocation change once its results are comparable and sufficiently informative.
| Finding | First question | Possible decision |
|---|---|---|
| Spend is behind plan | Where are opportunities being filtered or lost? | Address the identified constraint or revise the spend expectation. |
| Spend is on plan but outcomes weaken | What changed in the bought traffic or measurement? | Investigate the supported cause before buying more. |
| One segment appears stronger | Are definitions, conditions and outcome counts comparable? | Observe longer or make a bounded shift. |
| Retargeting looks efficient but adds little reach | Can its eligible pool use more budget well? | Review the next spend in both audience roles. |
Follow the delivery path
Separate incoming requests from impressions the campaign can buy. In Display & Video 360, consult the help guidance for troubleshooting deals and line items when investigating delivery issues.
A higher bid can address valuable eligible impressions lost on price only when their expected value supports the cost. It will not repair an ineligible creative, a required exclusion or a missing measurement event. If an automated target restricts delivery, relaxing it changes the efficiency trade-off; record that decision.
Compare outcomes on a common basis
For each meaningful segment, show spend, bought impressions, the chosen outcome count and its denominator. Keep format, market, dates, event and attribution rules visible.
A low CPM can conceal weak downstream results; a high conversion rate from very few actions can shift sharply with one more action. Choose a review threshold for the decision and outcome volume rather than applying a universal impression or conversion minimum.
An operating comparison can guide allocation without proving that the segment label caused a difference. Audiences, placements and auction conditions may vary. If those differences or sparse outcomes prevent a confident choice, keep any shift limited and name the evidence needed at the next review.
Key Metrics for Programmatic Campaign Review
- Spend
- Track against plan with defined cost basis
- Bought Impressions
- Measure actual impressions delivered by line item
- Outcome Count
- Attributed conversions or other KPIs per campaign goal
- Denominator
- Total eligible actions or impressions for rate calculation
- Review Threshold
- Set volume-based criteria to avoid misleading signals
Allocate with guardrails
Give prospecting and retargeting distinct roles, then compare the likely value of additional spend in each. Their pools, messages and paths to action differ, so a lower reported retargeting cost per attributed conversion does not establish that it deserves all available budget.
Check eligible pool size, repeated exposure, remaining opportunity and the campaign's required exclusions and spend limits.
Budget controls and any automatic allocation can work differently across platforms and settings. Check the active mode and saved settings before treating a planned split as fixed.
Make the change traceable
Before editing, record the affected IDs, current settings, reason, expected effect, owner and review time. Save a baseline with a named time zone and stable cost and outcome definitions.
Record when the edit was saved, then check its observed setting and subsequent delivery; a saved edit need not affect delivery instantly. Note simultaneous changes.
Where available, review platform change records alongside performance data. Timing alone does not prove an edit caused a better result. At the next review, compare suitable periods after allowing for outcome and reporting delay.
Check the intended result and side effects such as lost reach, higher cost or a less suitable placement mix. Keep, revise or reverse the change according to the evidence; label an inconclusive result clearly.
In this guide
- Comparing inventory segments at a useful sample sizeJudge whether inventory-segment delivery and outcomes can support a budget decision while keeping small samples and buying differences visible.
- Diagnosing performance changes before raising bidsCheck reporting, eligibility, auction losses and outcome quality before deciding whether a higher programmatic bid is justified.
- Balancing prospecting and retargeting budgetsReview prospecting and retargeting spend using audience opportunity, mature outcomes and the value of additional delivery.
- Recording the effect of a trader's optimisation changeLog a trader's reason, saved setting and baseline, then review delivery and mature outcomes without claiming false causality.
