Incrementality
CTV campaign measurement
Plan CTV measurement around delivery, estimated reach, completed playback and business outcomes, with the limits of each measure made clear.
Measure a connected TV (CTV) campaign against the question it was meant to answer. Separate delivery, estimated audience reach, playback and business outcomes. An impression or completed play shows an advertising event; neither proves that someone paid attention or that the ad caused a sale.
Agree on the question before launch
For an awareness campaign, the review might ask whether the buy reached enough of the intended audience and whether awareness changed. For an enquiry campaign, it might ask what happened to recorded enquiries.
Define the outcome, Australian locations, campaign dates and reporting method in advance. Keep other measures for diagnosis rather than letting the easiest number become the result.
In a CTV report, distinguish delivery and reach, brand response, and recorded enquiries or sales. State how each was measured.
| Question | Evidence to request | Limit |
|---|---|---|
| Did the buy deliver? | Impressions, dates, eligible inventory, device category and creative | Delivery does not identify everyone present at the screen. |
| How widely did it reach? | Estimated unique reach and frequency, with the identity unit stated | Devices, households and modelled people are different units. |
| Did the ad play through? | Starts, completions and the rate’s denominator | Playback to the end does not demonstrate attention. |
| Did the intended outcome change? | The agreed brand or business measure and any suitable comparison | A later action alone does not establish causation. |
Key CTV Measurement Metrics and Their Limitations
- Did the buy deliver?Impressions, dates, eligible inventory, device category and creative
- How widely did it reach?Estimated unique reach and frequency, with identity unit stated (devices, accounts, households, modelled people)
- Did the ad play through?Starts, completions and rate’s denominator (e.g., starts or impressions)
- Did the intended outcome change?Agreed brand or business measure with suitable comparison; causal claim requires experiment or credible control
Pre-Launch CTV Campaign Measurement Checklist
- Define clear campaign objectiveAwareness, enquiries, sales, or other measurable outcome
- Specify Australian locations and datesEnsure geographic and temporal alignment with target audience
- Confirm reporting method in advanceAgree on KPIs, data sources, and measurement standards
- Avoid using easiest metric as primary resultUse delivery or playback only for diagnostics, not outcome claims
Label the reach unit
Ask whether each reach figure refers to devices, accounts, households or modelled people. A shared television may have several viewers, while one person may encounter the campaign on several devices. A household estimate cannot identify which resident watched; a modelled people estimate is not an observation of everyone in the room.
Request the dates and inventory covered, how co-viewing and overlap are handled, and any reporting thresholds or delays.
Ask how estimates are modelled, when they become available and which report cuts are covered; check the proposed buy’s actual reporting coverage. Do not add publishers’ unique-reach figures without a method that accounts for overlap.
Read completions as playback evidence
Ask how the supplier defines a complete event and what denominator it uses: starts, impressions or another count. Compare the same periods, placements and creative lengths before interpreting differences.
Playback may continue without a viewer’s interaction. Ask what the seller or measurement partner can detect and which impressions its checks cover. Keep completions beside reach and the chosen outcome, rather than using them as a substitute for either.
Decide how to assess the outcome
Define the business event precisely, such as a submitted enquiry rather than a page visit. For a brand goal, decide whether a suitable study is feasible.
Record other promotions and changes that could affect the result. Without a credible comparison, report what was observed without calling the change incremental.
Some systems can associate an eligible CTV impression with a later action on another device. The resulting count depends on matching coverage, the conversion event, the time window and the credit rule.
It describes attribution under those rules, not what would have happened without the ad. A suitable comparison or experiment is needed for a causal claim.
Before booking, obtain the supplier’s available reporting fields and definitions. In the final review, show the agreed objective, delivery against the booked scope, reach with its unit and coverage, playback diagnostics, and outcome evidence. State missing inventory or unsettled data beside the conclusion.
CTV Measurement Best Practices: Key Reporting Requirements
- Define business event precisely
- e.g., submitted enquiry (not just page visit)
- Record external influences
- Other promotions, product launches, seasonal changes
- Use attribution with caution
- Depends on matching coverage, time window, credit rule – not causation
- Verify reporting fields before booking
- Confirm available data and definitions from supplier
How to Assess CTV Campaign Outcomes Correctly
- Define the business event clearlye.g., form submission, app download, or purchase – not just a website visit
- Identify potential confounding factorsNote concurrent campaigns, market shifts, or product updates
- Use matched data across devices if possibleLink CTV impression to later action via cross-device tracking
- Apply attribution rules transparentlyState time window, matching threshold, and credit model used
- Conclude with evidence, not assumptionReport observed change – only claim incrementality with experiment or control group
In this guide
- Household reach versus individual viewersLearn what household and individual CTV reach estimates count, how co-viewing affects them and what to check before comparing reports.
- Comparing completed views with business outcomesSee what a CTV completion rate can establish, how to compare it with enquiries or sales, and which conclusions need stronger evidence.
- Understanding the limits of CTV conversion attributionLearn how CTV conversions are matched and credited, what coverage and windows change, and why attribution alone cannot prove sales impact.
