Puzzle & Activity Book Niche Radar — Amazon.fr & Amazon.de

Niche research for puzzle and activity books (word search / mots mêlés / Wortsuche, sudoku, crosswords / mots fléchés / Schwedenrätsel, logic grids / logigrammes / Logicals, coloring, mazes, brain training, nostalgia quizzes, kids' activity books) on Amazon.fr and Amazon.de, built for KDP publishers who sell in French and German.

For every keyword you get one niche report: how many results Amazon shows, the real Best Sellers Rank (BSR) of the top organic books, review counts, EUR prices, publication dates, and a transparent opportunity score whose formula is documented below — no black box.

Why this Actor (vs. generic KDP analyzers)

• FR/DE first. French and German result-count strings, BSR lines ("Nr. 4.469 in Bücher", "3 247 en Livres"), German/French publication dates and EUR prices are parsed natively.
• Puzzle-book specific signals that generic analyzers don't compute:
  • puzzleTypeMix — which puzzle formats dominate the top results (FR/DE/EN vocabulary).
  • shareLargePrint — share of large-print titles (gros caractères / Großdruck / XXL).
  • Launch-wave detection — shareLaunchesLastNDays and newestLaunches: spots seasonal gold rushes (e.g. a flood of new Christmas word-search books with zero reviews) before you join them.
  • shareNearDuplicateTitles — share of top titles that are near-copies of each other, a hint of mass-produced low-content saturation.
  • shareIndependentlyPublished — how much of the top is self-published vs. traditional publishers.
• Keyword batch mode with per-keyword store routing: fr: logigrammes, de: logicals in one run.
• One clean row per niche, with the top books nested as evidence.

Input

Field	Description
keywords	List of keywords. Prefix fr: / de: to force the store per keyword.
marketplace	amazon.fr, amazon.de or both (for keywords without prefix).
category	books (default), kindle, all.
maxResultsPerKeyword	Organic results opened per keyword (3–30, default 10). Sponsored results are excluded.
fetchProductDetails	Open product pages for BSR, publication date, publisher, pages (default true).
recentLaunchDays	Window for "recent launch" detection (default 90).
delaySeconds	Polite randomized delay between page loads (default 3 s).
proxyConfiguration	Residential proxy in the store's country recommended (also gives local delivery address + local-VAT prices).

Example:

{
  "keywords": ["fr: logigrammes", "fr: nés en 1956", "de: rätselbuch jahrgang 1956", "de: sudoku großdruck"],
  "category": "books",
  "maxResultsPerKeyword": 10
}

Output (one item per keyword × store)

Key fields: keyword, marketplace, searchUrl, resultCount, resultCountIsLowerBound (Amazon shows "over 1,000" style counts), medianBsr, top3MedianBsr, bestBsr, medianReviews, shareUnder20Reviews, shareZeroReviews, medianPriceEur, medianAgeDays, shareLaunchesLast90Days, newestLaunches, shareLargePrint, puzzleTypeMix, shareNearDuplicateTitles, demandScore, competitionEaseScore, floodPenalty, opportunityScore, verdict, scoreConfidence, and topProducts[] (title, ASIN, URL, price, currency, BSR, category and sub-category ranks, rating, review count, publication date, publisher, pages, language, puzzle types, large-print flag).

Opportunity score (0–100) — formula

Demand (0-100)   = 100 × clamp( (log10(500000) − log10(top3MedianBsr)) / (log10(500000) − log10(5000)) )
                   → median BSR of the 3 best-ranked top books: 5,000 or better = 100, 500,000 or worse = 0

ReviewEase       = 100 × clamp( 1 − log10(1 + medianReviews) / log10(501) )     (500+ median reviews = 0)
LowReviewShare   = 100 × share of top organic books with < 20 reviews
ResultCountEase  = 100 × clamp( 1 − (log10(resultCount) − 2) / (log10(50000) − 2) )   (100 results = 100, 50,000 = 0;
                   neutral 50 if Amazon's count is missing or smaller than the results shown → resultCountSuspect)
CompetitionEase  = 0.5 × ReviewEase + 0.3 × LowReviewShare + 0.2 × ResultCountEase

FloodPenalty     = if shareZeroReviews ≥ 0.30 : min(20, 20 × shareLaunchesLastNDays) else 0
                   (pre-orders with a future date are counted in `preorders`, not as launches)

Opportunity      = clamp( 0.4 × Demand + 0.6 × CompetitionEase − FloodPenalty , 0, 100 )
Verdict          = "low demand" if Demand < 25; "opportunity" if ≥ 60 and CompetitionEase ≥ 45; "possible - needs a strong angle" ≥ 45; else "saturated"
Confidence       = high (≥5 BSRs read), medium (3–4), low (<3)

BSR is a sales-velocity proxy, not a sales figure; this Actor does not invent sales estimates. Thresholds are constants at the top of src/metrics.py.

Pricing

Pay per event: $0.15 per niche report — one report = one keyword on one Amazon store, with up to maxResultsPerKeyword books analyzed (BSR, reviews, price, publication date). Reports that end as blocked or empty are not charged. Example: 4 keywords on one store = 4 reports = $0.60.

Tip: set a maximum cost per run in the run options if you want a hard budget cap.

Limits & good practice

• Data is read from public Amazon pages; Amazon may show CAPTCHAs. The Actor runs sequentially with delays and retries, and reports status: "blocked" instead of guessing.
• Without a proxy in the store's country, Amazon may show a non-EU delivery address; prices are still forced to EUR via the store currency setting but may differ slightly from local-VAT prices. Check priceCurrencies.
• Use the results as research evidence, not as a guarantee of sales. Respect Amazon's terms of use and KDP content guidelines.

En bref (FR)

Analyse de niches KDP pour livres de jeux et d'activités sur Amazon.fr et Amazon.de : nombre de résultats, BSR réel des meilleurs livres, avis, prix en euros, dates de parution, vagues de lancements récents, part de gros caractères et score d'opportunité transparent (formule ci-dessus). Préfixez un mot-clé par fr: ou de: pour choisir la boutique.

Kurz (DE)

KDP-Nischenanalyse für Rätsel- und Beschäftigungsbücher auf Amazon.de und Amazon.fr: Trefferanzahl, echter Bestseller-Rang der Top-Bücher, Bewertungen, Euro-Preise, Erscheinungsdaten, Erkennung von Veröffentlichungswellen, Großdruck-Anteil und ein transparenter Opportunity-Score (Formel oben). Mit de: oder fr: vor dem Suchbegriff wählen Sie den Marktplatz.