Measuring Brand Lift for a Reach Campaign with Weak Signal
Verdant Outdoor sells hiking gear and runs a Video reach campaign for a spring brand campaign. It spent $60,000 over four weeks, delivered 3.1 million impressions, and reached 1.4 million unique users in a target audience of 12 million. View rate was 32%, and average frequency was 4.8. A Brand Lift study reports ad recall lift of 3.8 percentage points at 70% confidence and consideration lift of 1.2 points at 55% confidence. The campaign used a 45-second brand film and a 15-second cutdown, with no creator content. Site conversion tracking exists, but the campaign was optimized for reach, not conversions. The CMO wants to know whether to renew at $60,000, shift budget to Video action, or change creative. Sales are seasonal and the campaign ends in two weeks.
1. Interpret the Brand Lift results. What can and cannot be concluded, and what should Verdant do about the confidence levels?
Read the worked answer
The study does not give a reliable answer yet. A 3.8-point ad recall lift at 70% confidence is not statistically significant by the usual 90% or 95% standard, and the 1.2-point consideration lift at 55% confidence is indistinguishable from noise. Verdant cannot conclude that the campaign failed, but it also cannot claim that it worked. The reach and view-rate numbers show delivery, not persuasion; they do not substitute for a valid lift result.
The likely issue is insufficient statistical power. The campaign reached 1.4 million unique users, which sounds large, but Brand Lift depends on survey responses and the size of the lift being detected. A small true effect in a broad audience may need more responses, a longer study, or a more concentrated audience. Do not renew at $60,000 based on these confidence levels. First check whether the study had enough completed surveys and whether the target audience was defined broadly enough to dilute the effect. Then decide whether to extend the study or redesign it around a narrower audience and stronger creative.
2. Design the next measurement plan to determine whether to renew, including the role of a conversion-focused campaign and experiments.
Read the worked answer
Use a two-part plan. First, run a properly powered Brand Lift study with a clear primary metric, adequate survey volume, and a decision rule set in advance. Second, add a conversion-focused Video action campaign with a modest budget to see whether the brand film can produce lower-funnel behavior, not just recall. The two campaigns should have separate budgets and separate measurement so the reach campaign is not judged on direct-response CPA and the action campaign is not credited for brand impressions.
Where possible, use a geo holdout or Google Ads experiment to compare exposed and unexposed markets on branded search, site visits, and sales. Track view-through conversions separately and do not let them carry the entire case for renewal. Define success before launch: for example, significant ad recall lift at 90% confidence plus a measurable lift in branded search in exposed geos. If the action campaign produces efficient sales while the brand campaign remains inconclusive, the CMO has a clearer basis for reallocating budget than a single weak Brand Lift result.
3. Recommend budget and creative changes for the final two weeks and the renewal decision.
Read the worked answer
Do not make a wholesale shift to Video action with only two weeks left; that would contaminate the remaining brand measurement and leave neither question answered. Keep the reach campaign running but shift a small share, such as 15-20%, to a Video action test using the 15-second cutdown and a creator-style vertical asset. Use the final two weeks to gather creative learning and early conversion signal, not to declare a winner.
For the renewal decision, require a statistically credible Brand Lift read or a credible holdout result; otherwise renew only a smaller test rather than the full $60,000. If ad recall becomes significant but consideration does not, the creative may be memorable but not persuasive, so test a more product-led 15-second asset. If neither moves, reassess audience and message before adding budget. The judgment is to avoid treating a 70%-confidence result as proof and to avoid solving a measurement problem by simply buying more reach.