Track Amazon Bestsellers by Category at Scale (2026)

Published April 27, 2026 · 1556 words · For researchers

Thirdwatch's Amazon Product Scraper feeds a structured bestseller-tracking pipeline at $0.002 per record — weekly snapshot category keyword searches across 19 marketplaces, compute review-count velocity, surface rising products before they peak. Built for e-commerce ops watching category trends, brand teams monitoring competitor launches, dropshippers researching opportunities, and market analysts studying Amazon marketplace dynamics.

Why track Amazon bestsellers by category

Amazon bestseller rankings move fast. According to Amazon's 2024 marketplace disclosures, the platform processed over $620B in gross merchandise volume across 19 marketplaces with hundreds of millions of unique SKUs — and category leaderboards turn over 30-40% of their top 100 within any given month. For e-commerce ops planning inventory, brand teams watching competitor launches, and dropshippers researching opportunities, programmatic bestseller tracking is the leading-indicator data layer.

The job-to-be-done is structured. An e-commerce ops team monitors 30 product categories weekly to time their own inventory decisions against rising competitor products. A brand team watches their category for new entrants gaining review velocity above 30% week-over-week. A dropshipper researches which categories are seeing fresh winners worth listing. A market analyst studies cross-market bestseller divergence (US Top-50 vs UK Top-50) to inform pricing and merchandising strategy. All reduce to category keyword + country + maxResults pulls returning structured product rows ready for velocity analysis.

How does this compare to the alternatives?

Three options for getting Amazon bestseller-tracking data into a pipeline:

Helium 10 and Jungle Scout offer Amazon-specific dashboards bundled with seller tools. The Amazon Scraper actor page gives you the raw weekly data layer at meaningfully lower unit cost, with full schema control for custom analytics.

How to track Amazon bestsellers in 4 steps

Step 1: How do I authenticate against Apify?

Sign in at apify.com (free tier, no credit card), open Settings → Integrations, and copy your personal API token. Every example below assumes the token is in APIFY_TOKEN:

export APIFY_TOKEN="apify_api_xxxxxxxxxxxxxxxx"

Step 2: How do I take a weekly snapshot of bestsellers per category?

Pass category keywords as queries; Amazon's search is relevance-ranked, which approximates bestseller order for broad category terms.

import os, requests, datetime, json, pathlib

ACTOR = "thirdwatch~amazon-product-scraper"
TOKEN = os.environ["APIFY_TOKEN"]

CATEGORIES = ["wireless headphones", "smartwatch", "robot vacuum",
              "air fryer", "yoga mat", "mechanical keyboard",
              "gaming mouse", "standing desk", "kindle ereader",
              "portable monitor"]

resp = requests.post(
    f"https://api.apify.com/v2/acts/{ACTOR}/run-sync-get-dataset-items",
    params={"token": TOKEN},
    json={"queries": CATEGORIES, "country": "us", "maxResults": 100},
    timeout=600,
)
records = resp.json()
week = datetime.date.today().isocalendar()
ts = f"{week.year}-W{week.week:02d}"
pathlib.Path(f"snapshots/amazon-bestsellers-{ts}.json").write_text(json.dumps(records))
print(f"{ts}: {len(records)} products across {len(CATEGORIES)} categories")

10 categories × 100 results = 1,000 records per weekly snapshot, costing $2.

Step 3: How do I detect rising products via review-count velocity?

Aggregate snapshots; track reviews_count change per ASIN week-over-week.

import pandas as pd, glob, re

def parse_count(s):
    if not s:
        return None
    digits = re.sub(r"[^\d]", "", str(s))
    return int(digits) if digits else None

frames = []
for f in sorted(glob.glob("snapshots/amazon-bestsellers-*.json")):
    week = pathlib.Path(f).stem.replace("amazon-bestsellers-", "")
    for j in json.loads(pathlib.Path(f).read_text()):
        frames.append({
            "week": week,
            "asin": j.get("asin"),
            "title": j.get("title"),
            "rating": j.get("rating"),
            "reviews_count": parse_count(j.get("reviews_count")),
            "is_prime": j.get("is_prime"),
            "price": j.get("price"),
        })

df = pd.DataFrame(frames).dropna(subset=["asin", "reviews_count"])

# Compute review velocity: rate of new reviews per week
pivot = df.pivot_table(index=["asin", "title"], columns="week",
                       values="reviews_count", aggfunc="first").fillna(method="ffill", axis=1)
weeks = sorted(pivot.columns)
if len(weeks) >= 2:
    pivot["wow_growth"] = pivot[weeks[-1]] - pivot[weeks[-2]]
    pivot["wow_pct"] = pivot.wow_growth / pivot[weeks[-2]].clip(lower=1)
    movers = pivot[
        (pivot[weeks[-1]] >= 100)
        & (pivot.wow_pct >= 0.30)
    ].sort_values("wow_pct", ascending=False)
    print(movers[[weeks[-2], weeks[-1], "wow_growth", "wow_pct"]].head(15))

A product gaining 30%+ in review count within a week with at least 100 baseline reviews is in active velocity ramp — the canonical "rising bestseller" signal.

Step 4: How do I cross-validate with rating and Prime availability?

Filter velocity movers by rating ≥ 4.3 and Prime status to remove sponsored-listing noise and quality-issue false positives.

qualified_movers = df[df.asin.isin(movers.index.get_level_values(0))]
qualified_movers = qualified_movers[
    (qualified_movers.rating >= 4.3)
    & (qualified_movers.is_prime == True)
    & (qualified_movers.week == weeks[-1])
]

print(f"Validated rising bestsellers: {len(qualified_movers)}")
print(qualified_movers[["title", "rating", "reviews_count",
                         "is_prime", "price"]].head(15))

Forward newly-validated movers to a Slack channel for product-research review, and persist in a Postgres database for longitudinal trend analysis.

Sample output

A single record from the dataset for one trending product looks like this. Five rows of this shape weigh ~3 KB.

{
  "title": "Apple AirPods Pro (2nd Generation)",
  "price": "$189.99",
  "rating": 4.7,
  "reviews_count": 125000,
  "is_prime": true,
  "asin": "B0D1XD1ZV3",
  "image_url": "https://m.media-amazon.com/images/I/...",
  "url": "https://www.amazon.com/dp/B0D1XD1ZV3",
  "domain": "amazon.com"
}

asin is the canonical primary key for trend tracking; reviews_count is the velocity signal. rating filters quality-issue false positives — a product with rapidly-growing reviews but rating dropping below 4.0 typically indicates a manufacturing defect or unfulfilled-promise scam, not legitimate momentum. is_prime flag indicates whether the product is fulfilled by Amazon — rising bestsellers without Prime rarely sustain top positions.

Common pitfalls

Three things go wrong in production bestseller-tracking pipelines. Sponsored-listing noise — Amazon mixes sponsored placements into search results; sponsored products often show fewer reviews than organic top-rankers, so filtering on reviews_count >= 100 cuts most sponsored noise. New-product cold-start — products launched in the last 30 days have low review baselines that produce wild week-over-week percentage changes; either filter to products with reviews_count >= 500 baseline or use absolute review-count growth thresholds rather than percentages. Cross-marketplace ASIN reuse — the same ASIN can have different reviews_count across amazon.com and amazon.co.uk because each marketplace has its own review database; group by (asin, domain) rather than asin alone for cross-market analysis.

Thirdwatch's actor returns domain and asin on every record so cross-marketplace dedup and grouping are clean. The pure-HTTP architecture means a 1,000-product weekly snapshot completes in under five minutes and costs $2 — small enough to run across 30+ categories without budget consideration. Pair Amazon with our Flipkart Scraper for cross-marketplace India e-commerce trend analysis or AliExpress Scraper for sourcing-side bestseller research. A fourth subtle issue worth flagging: Amazon's review_count counter is sticky in the search-results layer (sometimes 6-24 hours behind the detail page), so for absolute-precision velocity tracking on top-rank-sensitive categories, fetch the specific ASIN URLs via the actor at the start of each tracking week to refresh review counts directly. A fifth pattern: seasonal categories (Halloween costumes, Christmas decor, summer gear) show massive seasonal swings that distort year-over-year baselines; segment your trend dashboards by season-aware categories rather than treating them as steady-state.

Related use cases

• Scrape Amazon products for price monitoring
• Build an Amazon product research tool
• Monitor Amazon competitor pricing and ratings
• The complete guide to scraping e-commerce data
• All Thirdwatch use-case guides

Frequently asked questions

Why track Amazon bestsellers programmatically?

Amazon bestseller rankings shift rapidly — sometimes hourly within a category. According to Amazon's 2024 marketplace data, the top 100 in any category turns over 30-40% in any given month. For e-commerce sellers, brand teams, dropshippers, and product researchers, programmatic tracking surfaces rising products before they peak, enabling timely inventory decisions, partnership outreach, or competitive responses.

How does this differ from price monitoring?

Price monitoring tracks the same SKUs over time. Bestseller tracking surfaces NEW SKUs entering the top performers — a different question entirely. Run both in parallel: price monitoring tells you what your watchlist is doing, bestseller tracking tells you what should be on your watchlist. The same Thirdwatch Amazon Scraper handles both with different query patterns.

What signals identify a rising bestseller?

Three: (1) review-count velocity (week-over-week growth in reviews_count exceeding 30% indicates active customer purchasing). (2) rating stability above 4.3 (rapid review-count growth with rating dropping below 4.0 signals quality issues, not legitimate momentum). (3) Prime availability (new entrants without is_prime: true rarely sustain bestseller positions for long). Combined, these three filter the genuine rising-bestseller signal from sponsored-listing noise.

How often should I refresh bestseller tracking?

Weekly cadence is the standard for trend detection. Daily catches noise (Amazon's algorithm rotates rankings hourly during prime shopping windows); monthly is too coarse to catch fast-moving categories like consumer electronics or beauty. Weekly Sunday-night snapshots align with how Amazon's category leaderboards stabilise over weekends.

Can I track bestsellers across multiple Amazon marketplaces?

Yes. The actor supports 19 country marketplaces. Run parallel queries with country=us, uk, in, de, fr, jp etc. and compare bestseller composition by market. Cross-market analysis reveals which products are launching globally vs locally, which categories are saturated in mature markets but rising in emerging ones, and where arbitrage opportunities exist for cross-border sellers.

How does this differ from Amazon's own bestseller pages?

Amazon publishes bestseller rankings at amazon.com/Best-Sellers/ but doesn't expose them as structured data. The Thirdwatch actor returns search-results data (sorted by relevance, which approximates bestseller-style ranking for category keyword queries) with full structured metadata — rating, reviews_count, is_prime, price, ASIN — ready for trend analysis. For the literal bestseller list, scrape the bestseller pages with our actor by passing those URLs as queries.

Run the Amazon Product Scraper on Apify Store — pay-per-product, free to try, no credit card to test.