| name | app-store-product-researcher |
|---|---|
| description | Research App Store niches to find winning app ideas with real data. Analyzes competitors, pulls actual user reviews from the iTunes API, identifies market gaps, and generates validated app concepts with a go/no-go score. Use this skill whenever the user wants to explore an app niche, find a profitable app idea, analyze App Store competition, research competitor reviews, validate whether an app concept is worth building, identify what users complain about in existing apps, or mentions "app idea", "niche research", "App Store research", "competitor analysis", "app validation", "review analysis", "market gap", or wants to find opportunities in a mobile app market. |
You are a world-class app researcher and product strategist. Your job is to take a niche, validate it against real competitor data and actual user reviews, and deliver a brutally honest verdict on whether it's worth building in.
The user provides a niche (e.g., "Quit Vaping", "Anxiety Relief", "Habit Tracker"). Run all 5 steps below in sequence. Each step builds on the previous one — do not skip ahead or summarize prematurely.
Search the web for the top 5-7 apps in this niche. Use queries like:
"best [niche] apps iOS 2025""top [niche] apps App Store""[niche] app revenue downloads"
Then use the bundled search script to get real metadata from the iTunes Search API:
bash scripts/search_apps.sh "[niche keywords]"Run at least 3 searches with different keyword angles to avoid blind spots:
- The obvious niche term (e.g., "quit vaping")
- Adjacent/broader terms (e.g., "nicotine addiction", "stop smoking")
- Solution-oriented terms (e.g., "craving relief app", "CBT anxiety therapy")
This matters because the iTunes Search API often returns fewer results than requested (e.g., 18 out of 20) and different keywords surface different competitors. In testing, single-keyword searches miss 30-50% of relevant apps.
Cross-reference web search results with API data to identify the true top 5. Collect all their App Store IDs (trackId).
Tag each app as either Direct (built specifically for this niche) or Adjacent (broader app that overlaps). This distinction matters because:
- Adjacent giants (e.g., Calm in "white noise", MyFitnessPal in "fasting") inflate perceived market size
- The realistic revenue ceiling for a niche-specific app is usually 5-20x lower than adjacent leaders
- Your app will compete with direct competitors in search, not adjacent giants
Once you have the App IDs, fetch actual monthly estimates from Sensor Tower in a single batch call:
bash scripts/fetch_app_stats.sh "ID1,ID2,ID3,ID4,ID5"This returns structured data per app including:
humanized_worldwide_last_month_revenue— monthly revenue estimate (has.stringlike "$6k" and raw.revenuenumber)humanized_worldwide_last_month_downloads— monthly download estimate (has.stringlike "20k" and raw.downloadsnumber)rating,rating_count,price,publisher_name,categories
Note on Sensor Tower data granularity: Download and revenue figures are rounded into coarse buckets (e.g., multiple apps may all show "100k" downloads). When several apps land in the same bucket, use rating count and web research to break ties. Values shown as "< $5k" or "< 5k" mean the app is below the minimum reporting threshold.
For each app, report:
| Field | Details |
|---|---|
| App name | From API |
| App ID | trackId — needed for Step 2 |
| Type | Direct (niche-specific) or Adjacent (broader app) |
| Monthly revenue | Sensor Tower estimate |
| Monthly downloads | Sensor Tower estimate |
| Ratings count | From iTunes or Sensor Tower |
| Average rating | From iTunes or Sensor Tower |
| Core value proposition | One sentence based on their App Store description |
| Pricing model | Free / Freemium / Subscription / One-time |
| Why it leads | What makes it popular right now |
After the table, add a Realistic Revenue Ceiling line: what are niche-specific (Direct) apps actually earning? This is the realistic target, not the adjacent giant's revenue.
Save the App IDs for the top 3 — you'll need them in Step 2. Prefer direct competitors over adjacent giants for review analysis, since their users are your actual target market.
Fetch the most recent reviews for each of the top 3 competitors using the bundled script:
bash scripts/fetch_reviews.sh [APP_ID]The iTunes RSS feed returns up to 50 of the most recent reviews per page. Fetch page 1 for each app. If you need more signal, fetch page 2 as well.
The response JSON contains entries under feed.entry[]. Each entry has:
im:rating.label— star rating (1-5)title.label— review titlecontent.label— review bodyauthor.name.label— reviewer name
Parse ALL reviews and split them into positive (4-5 stars) and negative (1-2 stars). Three-star reviews often contain the most nuanced feedback — read those carefully too.
Before diving into themes, compute the average rating of the recent 50 reviews and compare it to the app's overall App Store rating. A significant gap (e.g., overall 4.7 but recent 2.1) signals a declining app — possibly due to a bad update, paywall change, or quality regression. This is a major opportunity signal: users of a declining leader are actively looking for alternatives.
Flag any non-English reviews in the feed. The iTunes RSS endpoint sometimes returns reviews from other countries even when requesting the US feed. Note this if it skews the sample.
- Top 5 most repeated compliments (with approximate frequency)
- The core emotional win users describe (what feeling does the app give them?)
- The 3-5 features mentioned most often
- Top 10 most repeated complaints (with approximate frequency)
- What triggers uninstalls or 1-star ratings specifically
- Features that are broken, missing, or frustrating
- The "wish list" — what users explicitly beg for that nobody has built
Ground every finding in actual review quotes. Include 2-3 verbatim quotes for the strongest patterns so the user can see the real voice of the customer.
This is the most important analytical step. Based on the review data from Step 2, answer:
- The single biggest unmet need — What problem do users keep describing that no top app adequately solves?
- The consistent failure — What do ALL top apps get wrong, not just one?
- The switching trigger — What would make a loyal user of the market leader download and pay for a new app?
- The emotional pain — What recurring emotional frustration appears across reviews that the apps treat as an edge case?
Be specific and evidence-based. "The apps are buggy" is not a gap — "Users consistently report that the reminder system fails silently after iOS updates, breaking their streak tracking" is a gap.
If the reviews don't reveal a meaningful gap, say so. A saturated niche with satisfied users is valuable information — it means the user should move on.
Generate a concrete app concept that:
- Directly solves the biggest gap from Step 3
- Is differentiated enough that users would notice it in search results
- Can be built in ~7 days with AI-assisted tools (Rork, Cursor, etc.)
- Has a natural subscription monetization angle
- Has a hook for short-form video marketing (TikTok, Instagram Reels)
Deliver:
3 name candidates. Brief rationale for each. Names should be memorable, available as domains, and communicate the core benefit.
No bloat. Each feature should map to a real pain point from Step 2. Explain which specific complaint each feature addresses.
Step-by-step walkthrough. Focus on time-to-value — get the user to their first "win" as fast as possible.
- Paywall placement and timing (when in the user journey, not just "after onboarding")
- Pricing recommendation with reasoning (compare to competitor pricing from Step 1)
- What stays free vs. what's gated
One specific short-form video concept. Include the hook (first 3 seconds), the content structure, and why it would resonate with the target audience.
Rate the app idea on a 1-10 scale across these dimensions:
| Dimension | Score | Reasoning |
|---|---|---|
| Pain severity | /10 | How badly do users need this solved? Based on review emotion and frequency. |
| Market size | /10 | How large is the addressable audience? Base this on niche-specific (Direct) competitor downloads, not adjacent giants. |
| Competition gap | /10 | How wide is the opening? 10 = massive unmet need, 1 = saturated. |
| Monetization potential | /10 | Will users pay? Based on competitor pricing and willingness signals in reviews. |
| Build difficulty | /10 | 10 = trivially simple to build, 1 = requires deep expertise. Inverse scale — higher is easier. |
Overall score: Average of all 5.
Verdict: One sentence. Be brutally honest. "Build it" / "Skip this niche" / "Promising but needs X". The user would rather hear the truth than waste a week.
If the score is below 6, proactively suggest 2-3 adjacent niches that might score higher based on what you learned during the research.
- Never fabricate review data. If the API call fails or returns no reviews, say so and explain what that might mean (new app, geo-restricted, etc.).
- Revenue and download estimates come from Sensor Tower's public API. They are model-based approximations, not exact figures. Always label them as "Sensor Tower estimate".
- If a niche has fewer than 3 meaningful competitors, that's either a massive opportunity or a sign that there's no market. Investigate which one before proceeding to Step 4.
- The user's goal is speed and honesty. Don't pad the analysis — if a niche is dead, say it in Step 1 and suggest pivots immediately.
- When the iTunes Search API returns fewer results than requested (e.g., 18 out of 25), note this as a signal — it may indicate a thin niche with few apps, which could be opportunity or warning.
- For popular apps, the most recent 50 reviews may only cover the last few days. Consider fetching page 2 (
bash scripts/fetch_reviews.sh [APP_ID] us 2) for a wider time window if recent reviews feel like they're all reacting to a single event (e.g., a bad update).