email-performance-optimiser
Agent: Social Media Manager
L2 lifecycle and email marketing manager responsible for onboarding sequences, nurture campaigns, retention emails, re-engagement, and transactional email design.
Department ethos: ideal-marketing.md
Tool policy: allowed-tools.yaml
Skill Description
Runs A/B tests on subject lines, send times, and content across all email programmes, then reports open, click, and conversion rates to continuously improve performance.
When to Use
- Open rates or click-through rates drop below historical benchmarks for any email programme.
- A new email sequence launches and baseline performance data needs establishing through controlled experiments.
- Quarterly reviews reveal flat or declining conversion rates despite consistent send volume.
- The team debates a messaging or design change and needs data instead of opinions to decide.
Workflow
- Audit current email performance across all active sequences using the programme performance baseline table in
assets/email-optimisation-report-template.md: open rate, CTR, conversion rate, unsubscribe rate, and revenue per send vs. benchmarks.
- Identify the highest-impact optimisation opportunities by calculating the revenue gap between current and benchmark performance for each sequence. Prioritise tests by expected revenue lift.
- Formulate a test hypothesis: specify the variable (subject line, send time, CTA, layout, or copy), the expected lift, and the minimum sample size for statistical significance. Document in the test record format from
assets/email-optimisation-report-template.md.
- Build the A/B test variant in the email platform. Ensure only one variable differs between control and test to isolate causation.
- Launch the test to a randomised segment. Monitor deliverability and rendering during the first send window.
- Wait for statistical significance (≥ 95% confidence) before declaring a winner. Document the confidence level, sample size, and observed lift in the test record.
- Roll the winning variant into production and update the sequence template. Add the learning to the cumulative learnings library in
assets/email-optimisation-report-template.md.
- Publish the monthly optimisation report using
assets/email-optimisation-report-template.md. Score the programme quality using references/scoring-rubric.md before delivery.
Anti-Patterns
- Testing multiple variables simultaneously without a multivariate framework. Why: Changing subject line and CTA in the same test makes it impossible to attribute which change caused the result.
- Calling a winner before reaching statistical significance. Why: Premature decisions based on small samples lead to false positives that degrade performance when rolled out.
- Optimising for open rate while ignoring downstream conversion. Why: A clickbait subject line can lift opens but tank conversions and increase unsubscribes, net-negative for revenue.
- Running tests without a documented hypothesis. Why: Random testing produces random learnings; hypothesis-driven testing builds a compounding knowledge base about what the audience responds to.
Output
Success artifacts:
- A/B test briefs with hypothesis, variable, sample size, and success criteria
- Test result reports with statistical confidence and observed lift
- Monthly optimisation summary with cumulative performance improvement
- Updated email templates incorporating winning variants
Failure reporting:
- Flag tests with deliverability issues or rendering errors within 2 hours of launch
- Escalate sustained performance declines that persist after optimisation attempts to the marketing lead
Related Skills
No related skills defined yet.
1---2name: email-performance-optimiser3description: Runs systematic A/B tests across email programmes to maximise conversion rates and revenue per send. Use when asked to email performance optimiser. Suggest when relevant.4---56# email-performance-optimiser78## Agent: Social Media Manager910L2 lifecycle and email marketing manager responsible for onboarding sequences, nurture campaigns, retention emails, re-engagement, and transactional email design.1112Department ethos: [ideal-marketing.md](../../../../departments/marketing/ideal-marketing.md)13Tool policy: [allowed-tools.yaml](../../../../allowed-tools.yaml)1415## Skill Description1617Runs A/B tests on subject lines, send times, and content across all email programmes, then reports open, click, and conversion rates to continuously improve performance.1819## When to Use2021- Open rates or click-through rates drop below historical benchmarks for any email programme.22- A new email sequence launches and baseline performance data needs establishing through controlled experiments.23- Quarterly reviews reveal flat or declining conversion rates despite consistent send volume.24- The team debates a messaging or design change and needs data instead of opinions to decide.2526## Workflow27281. Audit current email performance across all active sequences using the programme performance baseline table in [`assets/email-optimisation-report-template.md`](assets/email-optimisation-report-template.md): open rate, CTR, conversion rate, unsubscribe rate, and revenue per send vs. benchmarks.292. Identify the highest-impact optimisation opportunities by calculating the revenue gap between current and benchmark performance for each sequence. Prioritise tests by expected revenue lift.303. Formulate a test hypothesis: specify the variable (subject line, send time, CTA, layout, or copy), the expected lift, and the minimum sample size for statistical significance. Document in the test record format from [`assets/email-optimisation-report-template.md`](assets/email-optimisation-report-template.md).314. Build the A/B test variant in the email platform. Ensure only one variable differs between control and test to isolate causation.325. Launch the test to a randomised segment. Monitor deliverability and rendering during the first send window.336. Wait for statistical significance (≥ 95% confidence) before declaring a winner. Document the confidence level, sample size, and observed lift in the test record.347. Roll the winning variant into production and update the sequence template. Add the learning to the cumulative learnings library in [`assets/email-optimisation-report-template.md`](assets/email-optimisation-report-template.md).358. Publish the monthly optimisation report using [`assets/email-optimisation-report-template.md`](assets/email-optimisation-report-template.md). Score the programme quality using [`references/scoring-rubric.md`](references/scoring-rubric.md) before delivery.3637## Anti-Patterns3839- **Testing multiple variables simultaneously without a multivariate framework.** *Why*: Changing subject line and CTA in the same test makes it impossible to attribute which change caused the result.40- **Calling a winner before reaching statistical significance.** *Why*: Premature decisions based on small samples lead to false positives that degrade performance when rolled out.41- **Optimising for open rate while ignoring downstream conversion.** *Why*: A clickbait subject line can lift opens but tank conversions and increase unsubscribes, net-negative for revenue.42- **Running tests without a documented hypothesis.** *Why*: Random testing produces random learnings; hypothesis-driven testing builds a compounding knowledge base about what the audience responds to.4344## Output4546**Success artifacts:**47- A/B test briefs with hypothesis, variable, sample size, and success criteria48- Test result reports with statistical confidence and observed lift49- Monthly optimisation summary with cumulative performance improvement50- Updated email templates incorporating winning variants5152**Failure reporting:**53- Flag tests with deliverability issues or rendering errors within 2 hours of launch54- Escalate sustained performance declines that persist after optimisation attempts to the marketing lead5556## Related Skills5758*No related skills defined yet.*59- [`nurture-campaign-builder`](../nurture-campaign-builder/SKILL.md) — sibling skill under the same agent — combine with nurture-campaign-builder for end-to-end coverage60- [`onboarding-sequence-designer`](../onboarding-sequence-designer/SKILL.md) — sibling skill under the same agent — combine with onboarding-sequence-designer for end-to-end coverage61- [`retention-email-designer`](../retention-email-designer/SKILL.md) — sibling skill under the same agent — combine with retention-email-designer for end-to-end coverage62- [`transactional-email-designer`](../transactional-email-designer/SKILL.md) — sibling skill under the same agent — combine with transactional-email-designer for end-to-end coverage