Agent Skills: Gaia Product Portfolio Retrospective

Provides the Stage-10 closure and portfolio-compounding playbook for consumer one-off / in-app-purchase products - end-of-life decision triggers and review cadence, a three-way continue/maintenance/sunset evaluation plus sequel/DLC, a humane sunset that honors recent one-off purchases, a portfolio retrospective that harvests capabilities and calibrates the unit-economics model, documenting realized priors, and transferring the audience/CRM/payer-base forward into a new Discovery cycle. Use to close a title and compound learnings back to Stage 1. Money-gated and net-of-fee aware (decisions run on NET-LTV vs operating cost and realized-vs-predicted calibration).

UncategorizedID: frostaura/ai.toolkit.gaia/fa-product-portfolio-retrospective

Install this agent skill to your local

pnpm dlx add-skill https://github.com/frostaura/ai.toolkit.gaia/tree/HEAD/plugins/product/skills/fa-product-portfolio-retrospective

Skill Files

Browse the full folder contents for fa-product-portfolio-retrospective.

Download Skill

Loading file tree…

plugins/product/skills/fa-product-portfolio-retrospective/SKILL.md

Skill Metadata

Name
fa-product-portfolio-retrospective
Description
Provides the Stage-10 closure and portfolio-compounding playbook for consumer one-off / in-app-purchase products - end-of-life decision triggers and review cadence, a three-way continue/maintenance/sunset evaluation plus sequel/DLC, a humane sunset that honors recent one-off purchases, a portfolio retrospective that harvests capabilities and calibrates the unit-economics model, documenting realized priors, and transferring the audience/CRM/payer-base forward into a new Discovery cycle. Use to close a title and compound learnings back to Stage 1. Money-gated and net-of-fee aware (decisions run on NET-LTV vs operating cost and realized-vs-predicted calibration).

Gaia Product Portfolio Retrospective

Scope and when to use

Stage 10 (closure) of the money-gated lifecycle: decide a live title's end-state and compound its learnings back to Stage 1. Scope is one-off + IAP/IAG; auto-renew passes are red-lined unless an explicit in-scope decision is recorded. Decisions run on NET-LTV vs operating cost and realized-vs-predicted model calibration.

Use when:

  • a live title trips an end-of-life trigger or its review cadence comes due
  • a continue / maintenance / sunset (or sequel/DLC) decision must be made
  • a sunset must be run humanely and the payer base/tooling harvested forward

Do not use when:

  • the title is healthy and only needs day-to-day operations (use fa-product-liveops)
  • a new bet is being framed from scratch with no prior priors (use fa-product-discovery, S1)

Required inputs

  • the title's NET-LTV-vs-operating-cost trend and content-runway economics
  • the realized cohort data (retention, ARPPU, whale, churn-cause) and model version
  • platform-obsolescence signals and the opportunity-cost picture across the portfolio

Owned outputs

  • the continue / maintenance / sunset (or sequel/DLC) decision with rationale
  • the humane sunset plan (stop-selling, honor/refund, archive) when chosen
  • harvested capabilities + documented priors + a go/no-go on a follow-on cycle

Core workflow

  1. Watch end-of-life triggers and review cadence (standing NET-LTV-vs-operating-cost trend, content-runway economics, platform obsolescence, opportunity cost).
  2. Run the three-way evaluation: continue/invest, maintenance-mode, or sunset, plus a sequel/DLC option.
  3. On the sequel/DLC path, scope the follow-on, design migration/loyalty rewards, and build a pre-launch CRM bridge that converts the base into day-one buyers.
  4. On maintenance-mode, templatize evergreen events and harvest the long tail at minimal cost.
  5. On sunset, run it humanely: stop selling IAP ahead of shutdown; honor/refund recent one-off purchases per platform + consumer law; communicate timelines; archive IP/assets.
  6. Run the portfolio retrospective + capability harvest (event templates, live-ops tooling, economy models, CRM playbooks, ASO/creative database, the versioned unit-economics model + realized-vs-predicted calibration) and document realized LTV/retention/whale/churn-cause PRIORS.
  7. Transfer audience/CRM/payer-base forward, run the go/no-go on a follow-on, and if GO hand priors + payer base + tooling into a NEW Discovery cycle (Stage 1) — the portfolio compounding loop.

Stage focus — close the loop, never just close the title

The point of closure is compounding: a sunset still produces calibrated priors, a harvested toolchain, and a transferable payer base. Decisions are made on NET-LTV vs operating cost, and the realized-vs-predicted calibration is fed back so the next cycle's unit-economics model starts smarter than this one did.

Anti-patterns

  • do not keep a title live past the point where NET-LTV covers operating cost
  • do not sunset while still selling IAP or without honoring/refunding recent one-off purchases
  • do not discard event templates, CRM playbooks, or the calibrated model on shutdown
  • do not start a sequel without bridging the existing payer base into day-one buyers
  • do not skip documenting realized priors before handing off to Stage 1

Handoff and downstream impact

  • return the end-state decision, sunset plan, and harvested priors to the coordinator
  • on a GO follow-on, hand priors + payer base + tooling to fa-product-market-researcher (S1)
  • feed the realized-vs-predicted calibration into the next unit-economics model version

Completion checklist

  • the continue/maintenance/sunset (or sequel/DLC) decision is explicit and NET-justified
  • any sunset honored/refunded recent one-off purchases and archived IP/assets
  • realized priors documented and payer base/tooling transferred into the next cycle

References