Hire an AI new-user retention test planner

Everything it can hand back

Your task sets the actual outputs. The specialist picks the methods that cover them.

• A funnel with the share kept at each step and the biggest leak named
• The few actions that mark an active user, and one first-visit change with a target and a date
• A one-morning test plan with tasks, a read-aloud script and a short fix list
• A scorecard of what moved after the fix, against your target, with what to do next

How it does the job well

• Start from the user problem and the evidence for it, then the solution.
• Write decisions down with the reason and the options you rejected.
• Keep scope small enough to ship and learn. Name what is out of scope.
• Define how success will be measured before building.

Mistakes it avoids

• Treating opinions or one loud request as proven demand.
• Specs that skip edge cases, error states or dependencies.
• Roadmaps with dates but no owners or confidence levels.

What a good result looks like

An example, not your task: Sign-ups look fine but few people are still around a month later.

• Step 1: New user drop-off finder. Your funnel with the share kept at each step, the biggest leak, the actions of an active user and one first-visit change.
• Step 2: Small usability test plan. Tasks for the leak, the people to test, a read-aloud script, watcher rules, a note sheet and a short fix list.
• Pause. You run the test with three people and one outsider, ship the fix with the owner's approval and wait for a full window.
• Step 3: Product metrics review. Your new numbers set against the old ones and your target, with what moved, what did not and what to do next.
• HAND-BACK SUMMARY. Each step's file, any step not run and why, and every MISSING line in one list.

How the work is checked

• Each step passed its own checks, or the run stopped at the step that failed and says why.
• The steps ran in order, and each one used the result of the one before.
• Nothing was invented to fill a gap. Every gap is a MISSING: line.
• Nothing was sent, published or paid for without a clear yes.
• States which checks passed, failed or could not run, and names missing inputs.

These checks have not been run against a model yet. Your AI uses them to check its own work.

A second opinion

Before it hands back, your AI reviews the draft the way Andrew Chen and Melanie Perkins would. This is a way of reviewing the work. It does not mean they wrote or endorse it.

Capabilities

• Find the leak. Lays out visits, sign-ups and 30-day users with the share kept at each step, names the biggest drop and plans one first-visit change with a measure and a date.
• Plan a one-morning test. Picks three to five tasks at the leak and plans three testers plus one outsider. Writes a script that tests the product and not the person. Builds a short fix list.
• Pause, test, ship and wait. You run the test, ship the first-visit fix and wait until every new user since the fix has had 30 days. Then you export the same numbers again for the same length of time.
• Check the result. Sets the new numbers against the old ones and your target. Says what moved, why it likely moved and what to do next.

How the role works

1. Checks your task, context and available tools.
2. Picks the methods needed and prepares a work plan.
3. Carries out supported tasks, with each method's limits and review gates.
4. Checks the real deliverables and corrects supported faults.
5. Hands back files, evidence and unresolved gaps.

Your specialist's working setup

Download the base specialist without a task form or account. README.md and SETUP.md explain how to use the folder. SOUL.md defines its scope. INTAKE.md asks missing questions inside your AI. PLAYBOOK.md, SKILLS.md, BEST_PRACTICES.md, CHECKS.md and HANDOFF.md guide the work and its checked deliverables.

Complete associated skills, supporting files and licences are included. Choose only the methods needed for the actual task. Keep context, working files and accepted feedback in your AI project.

Use this specialist through MCP or the API

Call prepare_specialist with role_id find-where-new-users-leave-and-prove-the-fix-works only. No task is required to load it. Your AI asks what job you need done, what you want back and which relevant context it can use.

For task-led hiring, call find_specialist, explain the suggested specialists and let the user accept or change the team.

Complete ZIP: Download base specialist. API: GET /api/v1/specialists/find-where-new-users-leave-and-prove-the-fix-works/pack. The host AI supplies its own model, tools and execution; a pack does not connect accounts or run a task.

Inputs and helpful context

The specialist starts from the task brief. Only active methods require their inputs.

For the primary method, when used

• Visits, sign-ups and users still active at 30 days, for one named period
• What the goal of a new user is, in a sentence
• The page or the first five clicks of a new visit (optional)
• A description of your typical new user (optional)
• A second export of the same numbers, once every user since the fix has had 30 days

Tools and access

• Any good chat AI: Funnel shares and plain-language findings are text and arithmetic work.
• Your analytics tool: Gives the exact counts the skill needs, by period.
• Any good chat AI: Planning tasks and writing the script is text work any strong chat AI does with this skill.
• A screen and voice recorder: Record each session from the first screen so you can replay the pauses.

Functions and deliverables

• New-user retention test planner: Find where new users leave. Plan a small user test and review comparable results after the approved change.
  • A funnel with the share kept at each step and the biggest leak named
  • The few actions that mark an active user, and one first-visit change with a target and a date
  • A one-morning test plan with tasks, a read-aloud script and a short fix list
  • A scorecard of what moved after the fix, against your target, with what to do next
• New user drop-off finder: Find where new users drop off between visit, sign-up and day 30. Get the biggest leak, the missing first-visit step and one fix to measure.
  • Your funnel, with the count and share at each step
  • The biggest drop, ranked, written before any feature idea
  • An estimate of how many sign-ups the feature would reach
  • The step or message missing from the first visit
  • One change for new users, with a measure, a target and a date
• Small usability test planner: Plan a small usability test in one morning. Name what people should do. You get tasks, a read-aloud script and a fix list of the worst problems.
  • The three to five tasks to test, in order
  • Who to recruit and how, including one outsider
  • A read-aloud script and a note sheet
  • A fix list template, limited to the worst problems
• Product metrics review writer: Turn your product numbers into a metrics review: a scorecard against targets, what moved, likely causes and next steps. Paste your figures to start.
  • A scorecard against targets
  • Trends and notable changes
  • Likely causes for spikes and drops
  • Recommended actions
  Source: Anthropic.
• Product context file writer: Share product facts, audience and positioning. Get a reviewed context file that other marketing skills can reference.
  • A product marketing context file
  • A summary of what the file already holds
  • A change log of updates
  • Gaps you still need to fill
  Source: Corey Haines.
• Andrew Chen: Review product growth, returning users or a two-sided market using Andrew Chen’s public work. Start with the funnel and the smallest useful network.
  • Your problem worked on the way Andrew Chen would: their questions, their method, their push-back
  • The answer or the work, in Andrew Chen's voice, using their ideas by name
  • Anything that goes beyond their public record, marked
  • One suggested next step
• Melanie Perkins: Review a first-use flow or pitch using Melanie Perkins’s recorded approach. Watch beginners, clarify the problem and plan a focused early test.
  • Your problem worked on the way Melanie Perkins would: their questions, their method, their push-back
  • The answer or the work, in Melanie Perkins's voice, using their ideas by name
  • Anything that goes beyond their public record, marked
  • One suggested next step

Tasks you can use it for

• Sign-ups look fine but few people are still around a month later.
• Your team wants a new feature to bring users back and you are not sure it will reach them.
• You shipped a change and cannot tell whether it helped.

Limits to check before starting

• Your product has no 30-day history yet: run a short interview or a small usability test with a few people first.
• You want to know why ad sales fell: use the Ad sales drop analyst.
• You want to check if an A/B test has enough visitors: use the A/B test traffic checker.

Best way to use this role

Give it one clear outcome, the actual files and a definition of done. Keep the work in one AI project so it can reuse checked context and earlier files.

• Download the handoff pack. It's a zip with HANDOFF.md and every skill file this workflow uses.
• Give the folder to Claude (in Claude Code, or as files in a Claude Project) and say "Follow HANDOFF.md".
• Paste your funnel numbers for one named period. Do not mix periods.
• Stop after step 2, run the test, ship the fix and wait until every user who signed up after the fix has had 30 days.
• Come back with the new export and run step 3 to see what moved.

Read it as data

Read the complete role setup as JSON. Connect your AI to the AI Blocks MCP library and read aiblocks://roles/find-where-new-users-leave-and-prove-the-fix-works for its role instructions and pinned sources.

This is a prepared AI role. No task has run. Your AI carries out authorized work using its available tools and reports the actual deliverables and checks.