optimize-nurture — Claude Code command: plan and setup
optimize-nurture is a Claude Code command that turns nurture metrics and test results into an optimization plan. It flags weak stages and assigns follow-up work with owners and due dates.
Apache-2.0 ★ 450 updated 2026-04-03
.claude/commands/optimize-nurture.mdWhat optimize-nurture does
optimize-nurture combines MAP statistics with CRM pipeline, product usage and SDR assist metrics. It spots drop-offs, audience fatigue and high unsubscribe rates, then reviews experiment results and suggests changes such as new tests, refreshed content or a different message sequence.
The command produces an optimization brief, an experiment tracker and an alert tracker with SLAs. It also writes a plan to .claude/plans/plan-<timestamp>.json, including data sources, metric owners, experiment IDs, risk guardrails and success targets. Inputs can include a reporting window, KPIs, audience segments, active tests and alert thresholds. The default retry limit for automation fixes is 2.
If activation or pipeline KPIs miss their guardrail for two consecutive windows, the workflow triggers an escalation and notifies Marketing and Sales leadership. It can invoke nurture-analytics-partner and the nurture-testing and lifecycle-cadence skills; recommendations include owners, deadlines and dependencies.
What it helps with
- spot nurture stages with drop-offs or audience fatigue
- compare experiment outcomes against activation and pipeline KPIs
- plan content or sequence changes from test findings
- assign owners and deadlines for segment-level follow-up
Example prompts for the agent
- Build a nurture optimization plan for the last 30 days using activation and pipeline KPIs for the enterprise segment.
- Review the active nurture tests and suggest what to test next.
- Find segments with high unsubscribe rates and assign follow-up tasks.
How to install optimize-nurture
Manually: put the files in place
.claude/commands/optimize-nurture.mdCall it with /optimize-nurture. For all projects use ~/.claude/commands/
Manually: put the files in place
~/.codex/prompts/optimize-nurture.mdCall it with /prompts:optimize-nurture. Codex marks this format deprecated in favour of skills
Manually: put the files in place
.github/prompts/optimize-nurture.prompt.mdManually: put the files in place
.gemini/commands/optimize-nurture.tomlGemini stores commands as TOML: the prompt text goes into the prompt field
Questions about optimize-nurture
How do I install optimize-nurture in Claude Code?
Add the command file at .claude/commands/optimize-nurture.md.
What does optimize-nurture do?
It analyzes nurture performance and experiment results, then creates an action plan plus experiment and alert trackers.
What inputs can optimize-nurture use?
Provide a reporting window, KPIs and audience segments. You can also include running experiments and alert thresholds.
Source text
optimize-nurture.md · Apache-2.0
--- name: optimize-nurture description: Builds a nurture optimization plan with insights, experiments, and remediation tasks. usage: /lead-nurture-orchestration:optimize-nurture --window 30d --kpis "activation,pipeline" --segments "enterprise" --- # Command: optimize-nurture ## Inputs - **window** – reporting period (7d, 14d, 30d, quarter). - **kpis** – comma-separated metrics (opens, clicks, conversion, pipeline, velocity, unsub). - **segments** – optional audience filters. - **experiments** – optional list of running tests to review. - **alerts** – optional thresholds for escalations. ### GTM Agents Pattern & Plan Checklist > Based on GTM Agents orchestrator blueprint @puerto/plugins/orchestrator/README.md#112-325. - **Pattern selection**: Optimization usually runs **pipeline** (data → insights → experiments → action plan → governance). If insights and experiment analysis can proceed in parallel, log a **diamond** segment + merge gate. - **Plan schema**: Update `.claude/plans/plan-<timestamp>.json` with data sources, query/metric owners, experiment IDs, remediation tasks, risk guardrails, error handling, and success targets (activation %, pipeline lift, unsub ceiling). - **Tool hooks**: Reference `docs/gtm-essentials.md` (Serena for pipeline diffs + data pulls, Context7 for platform docs, Sequential Thinking for retrospection, Playwright for QA of revised assets). - **Guardrails**: Define retry limit (default 2) for automation fixes; escalation ladder = Nurture Analytics Partner → Marketing Ops Partner → Revenue Operations if alerts breach thresholds. - **Review**: Run `docs/usage-guide.md#orchestration-best-practices-puerto-parity` before sharing to ensure agent coverage + deliverables are complete. ## Workflow 1. **Data Consolidation** – MAP stats, CRM pipeline, product usage, SDR assist metrics. 2. **Insight Detection** – highlight stages with drop-offs, fatigue, or high unsub rates. 3. **Experiment Analysis** – summarize test results, significance, and rollout recommendations. 4. **Action Plan** – propose new tests, content refresh, audience refinements, or sequencing changes. 5. **Governance** – log owners, deadlines, and dependencies for each recommendation. ## Outputs - Optimization brief (insights, impact, recommended action, owner, due date). - Updated experiment tracker with verdicts and next tests. - Alert + SLA tracker for segments needing fast intervention. - Plan JSON entry stored/updated in `.claude/plans` for auditability. ## Agent/Skill Invocations - `nurture-analytics-partner` – leads analysis. - `nurture-testing` skill – validates experiment rigor. - `lifecycle-cadence` skill – checks pacing vs fatigue thresholds. ## GTM Agents Safeguards - **Fallback agents**: record substitutes (e.g., Lifecycle Cadence covering analytics) when primary owners unavailable. - **Escalation triggers**: if activation/pipeline KPIs fall below guardrail for two consecutive windows, invoke lifecycle rip-cord and notify Marketing + Sales leadership. - **Plan maintenance**: every optimization cycle must append results + changes to the plan JSON and status packets, matching GTM Agents audit expectations. ---
© gtmagents · Apache-2.0 · gtmagents/gtm-agents