---
url: https://superduperai.co/library/weekly-growth-review-agent-prompt/
title: Weekly growth review: an agent reads the posts and the installs per campaign
description: Paste the week's posts and the installs per campaign link; the agent reports numbers first, then what worked by name, then 3-4 actions, judging by installs and by multiples, never by a 20 % swing.
category: growth
platforms: [tiktok, reels, shorts, youtube]
works_with: [claude-code, claude, chatgpt, any-llm]
tags: [weekly-review, analytics, installs, agent, creator-growth, measurement]
difficulty: intermediate
verified: false
created: 2026-11-01
skill: creator-growth
---

# Weekly growth review: an agent reads the posts and the installs per campaign

Paste the week's posts and the installs per campaign link; the agent reports numbers first, then what worked by name, then 3-4 actions, judging by installs and by multiples, never by a 20 % swing.

## Prompt

```text
You are reviewing one week of short-form posts for {app name}. Inputs below: the post log (date, account, platform, format, hook, views, link to the post) and installs per campaign link from the store consoles.

Rules:
- Views are attention, installs are the result. Rank formats and accounts by installs per campaign, with views as a secondary column.
- At small volumes, percentage swings are noise and multiples are findings. Report a change only when it is 2x or more; never recommend killing a format over a 20 % week.
- A format is judged after a week of posts, not after one post.
- Say 'probably' where the data does not show a cause.

Report in this order:
1. NUMBERS: posts, views, installs per campaign, installs total, cost (0 in Phase 1).
2. WHAT WORKED: named posts with their numbers; which format and which account converted; which got reach without installs.
3. WHAT DID NOT: named posts with low views or low installs, with the probable reason (hook, framing, format, account).
4. PHASE 2 TRIGGER: did any video pass {view threshold} views AND show installs on its link? Yes or no, with the numbers.
5. ACTIONS: 3-4 lines, each starting with a verb, for next week. Keep the cadence at 1-2 posts a day.

Post log:
{post log}

Installs per campaign:
{installs per campaign}
```

## Steps

1. Export the post log from your sheet and the installs from App Store Connect (App Analytics, Sources) and Play Console; paste both.
2. Run it the same weekday every week, so weeks are comparable.
3. Carry the actions into next week's shoot list; a review that changes nothing is a report.

## Tips

- YouTube Analytics lags about 2 days for day-level data; for a Short posted this week read the view count from the video itself instead.
- Multiples, not percentages: 3 installs to 9 is a finding; 30 to 36 is noise.
- In Claude Code, `/youtube-channel` pulls the YouTube side through the Analytics API; the TikTok and Reels numbers are pasted by hand.

## Works with

- Claude Code: Runs as a skill or a prompt in the terminal; can read your files and write the result.
- Claude: Paste into claude.ai or the Claude app.
- ChatGPT: Paste into ChatGPT; works with any GPT model.
- Any LLM: Model-agnostic: Gemini, Mistral, a local model.