Test summary
This report presents the findings from an A/B test of Facebook advertisements for a candidate with immigrant background in a local election campaign. The experiment tested two messaging approaches: Version 1 (no mention of immigrant background) versus Version 2 (explicit mention of immigrant background). The test aimed to determine which messaging strategy would better resonate with all eligible voters in the candidate’s local area.
Based on the Meta advertising performance metrics, the version with no mention of immigrant background (Version 1) demonstrated superior performance across all key metrics and was declared the winner with 74% confidence. The conclusion is that for broad audiences in a local election the message that does not mention the candidate’s immigrant background works best.
Test Set-Up
Research question: Will a message that mentions the candidate’s immigrant background resonate more with voters, including immigrants eligible to vote, than a message that makes no mention?
Purpose: click throughs to register to vote, target all eligible voters.
Test Period: 4 days in April 2025 (prior to the May 1, 2025 local elections)
Budget: Approximately £2-3 daily spend (within the planned £25-40 budget)
Target Audience: All voting-age residents (ages 18+) in the target area
Primary Goal: Drive click-throughs to the voter registration website
Treatment
Version 1 (No-mention version):
- Made no explicit reference to the candidate’s immigrant background
- Focused on their experience as a county councillor and their achievements
- Emphasised the local impact of council decisions
Version 2 (Immigrant-mention version):
- Explicitly mentioned the candidate’s immigrant background in relation to their political career
- Otherwise maintained the same messaging about council responsibilities
- Aimed to establish connection with immigrant communities
Results
| Metric | Version 1 (No-mention) | Version 2 (Immigrant-mention) |
| Cost per click | £0.173 | £0.2049 |
| Reach | 3,269 | 2,888 |
| Confidence | 74% | 26% |
| Winner | ✓ | ✗ |
Key Findings
- Higher Efficiency: Version 1 delivered a 15.7% lower cost per click (£0.173 vs £0.2049), indicating better resource efficiency.
- Greater Reach: Version 1 reached 13.2% more people (3,269 vs 2,888), demonstrating better audience penetration within the same budget parameters.
- Statistical Confidence: Meta’s algorithm determined that Version 1 was the superior performer with 74% confidence.
Analysis and Implications
Contrary to initial assumptions that explicitly highlighting the candidate’s immigrant background would resonate more strongly with general voters, the data suggests that the more generalised messaging performed better. This could be attributed to several factors:
- Broader Appeal: The non-specific messaging may have appealed to a wider cross-section of voters, including but not limited to immigrant communities.
- Messaging Simplicity: The more straightforward messaging in Version 1 may have been more effective at driving the specific action (voter registration).
- Community Integration: The local immigrant population may identify more strongly with messaging that emphasises community-wide issues rather than segmenting them as a distinct demographic.
Recommendations
Based on these findings, we recommend the following actions for future campaign communications:
- Continue with Inclusive Messaging: Maintain the approach used in Version 1, which focuses on candidates’ experience and the local impact of council work rather than explicitly highlighting their identity and background.
- Test Additional Variables: Consider testing other elements such as video location settings, call-to-action phrasing, or visual elements.
Segmented Follow-up: Consider targeted follow-up messaging to specific community segments after initial registration, which might then incorporate more