Blog

Amazon "bought in past month": what the number means, and how to collect it with Python

What Amazon's "bought in past month" label means, how it differs from best sellers rank, and a tested Python script that collects it for a whole search.

Published

"Bought in past month" is the label Amazon shows on many product listings, such as "100+ bought in past month" or "10K+ bought in past month". It says that at least that many units of the listing were bought recently, rounded down to a round number. It is the only sales figure Amazon shows publicly for a product, so sellers, brands and researchers use it to compare how much products sell.

This guide explains what the number means and what it does not, how it differs from the best sellers rank, and what the labels looked like across two real searches. It then gives a tested Python script of 108 lines that searches Amazon for a keyword, reads every organic result and saves each product's "bought in past month" figure with its price, rating, reviews and rank to a CSV file. Every number below comes from real runs on 22 November 2026.

What does "bought in past month" mean on Amazon?

It is a rounded count of recent purchases of one listing, written as a floor with a plus sign. "900+" means at least 900. "10K+" means at least 10,000.

Amazon started showing the label in March 2023 as a test. Marketplace Pulse reported at the time that the number "is rounded to hundreds, thousands, or tens of thousands and is capped at one hundred thousand", and that the time range is either a week or a month (Marketplace Pulse). Amazon has not published how it counts.

These are the labels we saw on 36 listings across our two searches, from the smallest to the largest:

Label on the page What it means Where we saw it
300+ At least 300 A small-brand electric toothbrush with 38 reviews
700+, 900+ At least 700, at least 900 A Sonicare toothbrush, a Ninja air fryer
1K+, 2K+ At least 1,000, at least 2,000 Oral-B, Philips Sonicare and Ninja listings
4K+ to 8K+ At least 4,000 to at least 8,000 Most of the air fryers and toothbrushes
10K+, 20K+, 30K+ At least 10,000, 20,000, 30,000 The best sellers in each search

Two things follow from the rounding. A listing with 10K+ may have sold 10,000 or 19,999, so two listings with the same label can differ a lot. And the step between labels grows with the number, so the label is precise for small sellers and rough for large ones.

What does it not tell you?

  • It is not exact sales. It is a floor, rounded down. Treat it as "at least".
  • It is not revenue. Multiplying the label by today's price gives a rough minimum at best, because the price may have been different during the month.
  • It belongs to one listing, not one product. In our air fryer search, three Ninja listings shared the same rating, the same 11,116 reviews and the same category rank, because they are variants of one product page. Their labels were different: 10K+, 10K+ and 1K+. Reviews and rank are shared across variants; the purchase label is not.
  • It is not a history. The label shows the recent past only. Last month's number is gone once the page changes, unless you saved it.
  • Not every listing has one. A listing without the label is not proof of zero sales; record it as "no badge" rather than as zero.

How is it different from best sellers rank?

The best sellers rank is Amazon's other public sales signal: a position, such as #5 in Air Fryers, based on how a product sells compared with the others in that category. The two answer different questions:

Bought in past month Best sellers rank
What it is A rounded count of purchases A position in a category
Unit Purchases of this listing Rank among products in the category
Shared across variants No, each variant has its own Often yes, in our runs
Compare across categories Yes, a count is a count No, #5 in one category is not #5 in another
Shown On the product page and in search results On the product page, in the product details

Sales estimators turn the rank into a guess at units with their own models. The label is Amazon's own number, rounded. The script below saves both, so you can compare them on your own data.

How do you collect "bought in past month" with Python?

The script searches amazon.com for a keyword, reads every organic result's product page, and adds one row per product to bought_in_past_month.csv. It prints a table and the sum of the labels, the least the listings sold together in the past month. It uses the standard library only and needs Python 3.9 or later.

Get access

Every account starts with 200 free credits, and sign-in is by email. The quickest way is to give your coding agent (Claude Code, Cursor or similar) this line:

set up https://mcp.datawebot.com/setup.md

The agent asks for your email and you approve the sign-in from the email datawebot sends. The key is saved to ~/.datawebot/credentials, readable only by you, and the script reads it from there. That key lasts 30 days. For a script that runs every day, create a key that lasts 90 or 365 days on the dashboard under API keys, and put it in an environment variable named DATAWEBOT_API_KEY from your secret store. Never commit it.

The script

"""Amazon "bought in past month" for a keyword: search amazon.com, read every
organic result, save one dated row per product to a CSV file and print a table.
Standard library only. Usage: python3 bought_in_past_month.py "air fryer" """
import csv, json, os, sys, time, urllib.request
from concurrent.futures import ThreadPoolExecutor

API = "https://api.datawebot.com/v1/requests"
OUT = "bought_in_past_month.csv"
FIELDS = ["checked_at", "keyword", "position", "asin", "title", "price", "rating",
          "reviews", "bought_past_month", "rank_category", "rank", "seller", "url"]


def api_key():
    key = os.environ.get("DATAWEBOT_API_KEY")
    if key:
        return key
    # Written by the agent setup (https://mcp.datawebot.com/setup.md).
    with open(os.path.expanduser("~/.datawebot/credentials")) as f:
        return json.load(f)["profiles"]["https://mcp.datawebot.com"]["api_key"]


def call(method, url, body=None):
    req = urllib.request.Request(url, method=method,
        data=json.dumps(body).encode() if body else None,
        headers={"Authorization": "Bearer " + api_key(),
                 "Content-Type": "application/json",
                 "User-Agent": "amazon-bought-in-past-month-example/1.0"})
    with urllib.request.urlopen(req, timeout=180) as r:
        return json.loads(r.read())


def run(body):
    """Send one request and wait for the answer; None when it failed."""
    try:
        res = call("POST", API, body)
        for _ in range(60):  # when the answer is "working", poll until it is done
            if res.get("status") != "working":
                break
            time.sleep(res.get("check_after_seconds") or 4)
            res = call("GET", f"{API}/{res['request_id']}")
        return res
    except (OSError, ValueError):  # a network or server error counts as a failure
        return None


def read(item):
    """Read one product page; None when the read failed or was incomplete."""
    position, url = item
    res = run({"capability": "product", "country": "US", "url": url})
    p = (res or {}).get("product")
    d = (p or {}).get("details") or {}
    if not d.get("delivery_location"):  # an incomplete read: try it again later
        return None
    rank = (d.get("best_sellers_rank") or [{}])[-1]  # the narrowest category
    return {"checked_at": res["meta"]["as_of"], "position": position,
            "asin": url.rsplit("/dp/", 1)[-1].split("/")[0],
            "title": p["title"], "price": (p.get("price") or {}).get("amount"),
            "rating": p.get("rating"), "reviews": p.get("review_count"),
            "bought_past_month": d.get("sold_count") if d.get("sold_count_basis") == "monthly" else None,
            "rank_category": rank.get("category"), "rank": rank.get("rank"),
            "seller": (d.get("seller") or {}).get("name"), "url": url}


def badge(n):
    """Amazon's wording: 900 -> 900+, 6000 -> 6K+."""
    if n is None:
        return "no badge"
    return f"{n // 1000}K+" if n >= 1000 else f"{n}+"


def main():
    keyword = " ".join(sys.argv[1:]) or "air fryer"
    found = run({"capability": "search", "query": keyword, "site": "amazon.com",
                 "country": "US", "limit": 20})
    items = list(enumerate((p["url"] for p in (found or {}).get("products") or []), 1))
    if not items:
        sys.exit(f"The search for {keyword!r} returned no organic results.")
    with ThreadPoolExecutor(max_workers=2) as pool:
        rows = dict(zip(items, pool.map(read, items)))
    missed = [i for i, row in rows.items() if not row]
    if missed:  # failed reads are free, so each one is tried once more
        time.sleep(130)
        rows.update((i, read(i)) for i in missed)
    done = sorted((row for row in rows.values() if row), key=lambda r: r["position"])
    for row in done:
        row["keyword"] = keyword
    new_file = not os.path.exists(OUT)
    with open(OUT, "a", newline="", encoding="utf-8") as f:
        writer = csv.DictWriter(f, fieldnames=FIELDS)
        if new_file:
            writer.writeheader()
        writer.writerows(done)
    print(f'"{keyword}" on amazon.com, {len(done)} of {len(items)} organic results read')
    print(" #  bought      price  rating  reviews  rank       title")
    for r in done:
        rank = f"#{r['rank']:,}" if r["rank"] else "-"
        price = f"${r['price']:,.2f}" if r["price"] else "-"
        print(f"{r['position']:>2}  {badge(r['bought_past_month']):9} {price:>8}  {r['rating'] or '-':>6}"
              f"  {r['reviews'] or 0:>7,}  {rank:9}  {r['title'][:38]}")
    floor = sum(r["bought_past_month"] or 0 for r in done)
    print(f"At least {floor:,} bought in the past month across these {len(done)} listings")
    failed = [str(i[0]) for i, row in rows.items() if not row]
    if failed:
        print("Not read: result " + ", ".join(failed))


if __name__ == "__main__":
    main()

How it works

  • The search asks amazon.com for the keyword and keeps the organic results in their order. Sponsored results come back in a separate list and are left out, so ads never take a place in the table.
  • read sends one product read for each result, two at a time. The label comes back as details.sold_count, the number on the label (900 for "900+", 10000 for "10K+"), with details.sold_count_basis set to monthly. The script keeps it only when the basis is monthly.
  • The rank is the narrowest category rank on the page, such as Air Fryers rather than Kitchen & Dining, because that is the one shoppers compare.
  • An incomplete read is retried. A full read always says which delivery ZIP code it was read for. A read without one did not finish, so it is tried again after about two minutes, together with any read that failed outright. Failed reads are free. A result that still fails is named in the last line and gets no row.
  • badge writes the number back the way Amazon shows it, so the table reads like the page.
  • The file only grows. Each run appends rows with the time they were read, so running it every week builds the history the label itself does not keep.

What it printed for "air fryer"

The first run started at 02:20 UTC on 22 November 2026 and took 5 minutes 36 seconds, including the wait before the retries. All 16 organic results were read:

"air fryer" on amazon.com, 16 of 16 organic results read
 #  bought      price  rating  reviews  rank       title
 1  10K+        $49.99     4.6    7,617  -          Chefman Air Fryer – 4 QT Compact Airfr
 2  10K+        $79.99     4.7   11,116  #5         Ninja Air Fryer Pro, 5 Qt, 4-in-1, Gre
 3  10K+        $89.98     4.8   22,892  #4         Cosori TurboBlaze Air Fryer, 9-in-1, 6
 4  6K+         $59.99     4.6    7,617  #2         Chefman Air Fryer – 6 QT Compact Airfr
 5  10K+       $149.99     4.3   18,401  #7         Emeril Lagasse 26QT French Door Multi-
 6  6K+         $69.99     4.3   52,382  #8         CHEFMAN Multifunctional Digital Air Fr
 7  6K+         $83.99     4.7      639  #6         Cuisinart Air Fryer, Glass Basket & Cl
 8  6K+         $69.99     4.5    1,140  #9         Chefman Air Fryer 9 Qt | TurboFry 7-in
 9  900+        $99.99     4.6      362  #10        Ninja Air Fryer Pro XL, 6.5 Qt, 7-in-1
10  5K+         $47.99     4.6    1,190  #3         bella Fits-Anywhere™ 6qt Slim Air Frye
11  6K+        $109.95     4.6      139  #1         Cuisinart Air Fryer Toaster Oven Combo
12  10K+        $99.99     4.7   11,116  #5         Ninja Air Fryer XL with MaxCrisp, 6.5 
13  7K+        $199.95     4.6      841  #13        Ninja French Door Premier Air Fry Oven
14  5K+        $149.99     4.6   26,286  #1         Ninja Digital Air Fry Countertop Oven,
15  6K+        $159.95     4.8   25,225  #17        Ninja Foodi Air Fryer, 8 Qt DualZone, 
16  1K+        $109.99     4.7   11,116  #5         Ninja Air Fryer Pro, 5 Qt, 5-in-1, Gre
At least 104,900 bought in the past month across these 16 listings

What it printed for "electric toothbrush"

The second run started at 02:26 UTC and took 7 minutes 45 seconds. All 20 organic results were read:

"electric toothbrush" on amazon.com, 20 of 20 organic results read
 #  bought      price  rating  reviews  rank       title
 1  10K+        $59.99     4.3   19,603  #1         Oral-B iO Series 3 Limited Deep Clean 
 2  4K+         $59.96     4.4    2,115  #2         Philips Sonicare 5950 Rechargeable Ele
 3  6K+         $59.99     4.3   19,603  #1         Oral-B Rechargeable Electric Toothbrus
 4  5K+         $59.99     4.3   19,603  #1         Oral-B Rechargeable Electric Toothbrus
 5  2K+        $199.99     4.6      615  #11        Oral-B Rechargeable Electric Toothbrus
 6  5K+         $99.99     4.4    4,562  #6         Oral-B Rechargeable Electric Toothbrus
 7  2K+        $119.99     4.5    3,627  #3         Oral-B iO Deep Clean + Protect Recharg
 8  30K+        $49.94     4.5  166,019  #2         Oral-B Pro 1000 Rechargeable Electric 
 9  30K+        $29.99     4.2   21,421  #1         Philips Sonicare 4100 Rechargeable Ele
10  10K+        $29.99     4.4   14,821  -          Philips Sonicare 4100 Rechargeable Ele
11  20K+        $49.94     4.5  166,019  #2         Oral-B Pro 1000 Rechargeable Electric 
12  20K+        $29.99     4.3   23,486  #1         Philips Sonicare 4100 Rechargeable Ele
13  10K+        $49.99     4.5  166,019  #2         Oral-B Pro 1000 Rechargeable Electric 
14  4K+         $99.96     4.5    1,098  #3         Philips Sonicare 7300 Rechargeable Ele
15  2K+         $59.96     4.4    2,115  #2         Philips Sonicare 5950 Rechargeable Ele
16  300+        $23.99       5       38  #37        OasiSmile X610 Electric Toothbrush for
17  8K+          $8.49     4.3   14,658  #6         Sonic Electric Toothbrush for Adults, 
18  4K+         $33.95     4.5   35,887  #2         Aquasonic Vibe Series Ultra-Whitening 
19  700+        $39.96     4.5      298  #10        Philips Sonicare 4200 Rechargeable Ele
20  6K+         $49.99     4.4    2,679  #5         Oral-B iO Series 2 Limited Simple Clea
At least 179,000 bought in the past month across these 20 listings

To check a row against the page, we opened the first air fryer, the Chefman 4-quart, in a browser right after a read of the same listing. The page showed "10K+ bought in past month" and $49.99, the same as the row.

What do the two runs show?

  • The search order is not the sales order. The air fryer in first place and the one in twelfth place both had 10K+. In the toothbrush search, the two listings with 30K+ were in eighth and ninth place, below several listings with 2K+ to 6K+.
  • Cheap does not always win, but often does. The Philips Sonicare 4100 at $29.99 had 30K+, and a sonic toothbrush at $8.49 from a little-known brand had 8K+. The Oral-B listing at $199.99 had 2K+.
  • Variants split the sales. Three Oral-B Pro 1000 listings with the same 166,019 reviews had 30K+, 20K+ and 10K+. Adding up the variants of one product gives a better picture of that product than any one listing.
  • The sum is a floor. The 16 air fryer listings had at least 104,900 purchases between them in the past month, and the 20 toothbrush listings at least 179,000. The real totals are higher, because every label is rounded down.

That is the kind of answer the label is good for: which products in a search really sell, and roughly how much more one sells than another. Our guide to ecommerce market research with marketplace data shows how to turn rows like these into category shares, and the Amazon search scraper and keyword rank tracker records where each product appears for a keyword over time.

How do you track it over time?

Run the script on a schedule with the same keyword. Every run appends rows with their time, so after a few weeks the file shows how each listing's label moved: a product going from 2K+ to 6K+ is taking off, and one going from 10K+ to 4K+ is losing ground. Once a week is enough for most categories. Check more often around sale events: Black Friday is on 27 November 2026, and labels can move within days.

To watch specific listings rather than a search, put their ASINs in a list and read only those. Our ASIN lookup guide shows how to get the ASIN from a link, a barcode or a keyword, and our Amazon buy box tracker has a script built around a watchlist of ASINs that you can extend with the same field.

How do you get it from an AI agent?

If you work in Claude, ChatGPT, Cursor or another agent that supports MCP, you can ask without the script. After the setup line above, ask in plain language:

Use datawebot to search amazon.com for "electric toothbrush", read the top 10
organic results, and list each one's "bought in past month", price, rating and
best sellers rank, sorted by bought in past month.

The agent calls search_products and then get_product for each result, and answers from the same JSON the script reads. For a check that has to run every week without you, use the script and a schedule. The Amazon product data API page lists every field and the Amazon countries, and the MCP server page shows how to connect each client.

What should you check in "bought in past month" data?

  1. Read it as "at least". Every label is rounded down. Never report it as exact sales.
  2. Keep variants apart, then add them up on purpose. Reviews and rank are often shared across variants; the label is not. Group by product only when you mean to.
  3. Store "no badge" as its own value. A listing without a label is not a listing with zero sales.
  4. Keep the time of every read. The label describes the recent past at the moment the page was read. A row without a time cannot be compared with anything.
  5. Do not compare a week label with a month label. The script keeps only labels marked monthly, so a row never mixes the two.

What does it cost?

A search costs 5 credits and a product read 1 credit; failed reads are free. A run therefore costs about 5 credits plus 1 for each result: about 21 credits for the 16 air fryers and about 25 for the 20 toothbrushes. The 200 free credits every account starts with cover about eight runs of a 20-result search. More credits are one-time packs with no subscription, listed on the pricing page.

To read other fields from the same product pages, such as the seller, every variant and the GTIN, see our Amazon scraper API guide.

Need this at volume?

If you need the label for thousands of ASINs or many keywords every day, send a volume request and we will size it in credits. Our published samples show the result format.

Frequently asked questions

What does "100+ bought in past month" mean on Amazon?

That at least 100 units of that listing were bought in the past month. The number is rounded down, so the listing may have sold anywhere from 100 to just under the next label. It is a recent count for that one listing, not its sales over all time.

What does "1K+ bought in past month" mean?

At least 1,000 purchases of the listing in the past month. "K" means thousand, so 10K+ is at least 10,000. Above a thousand, the label moves in steps of a thousand, then ten thousand, so the range behind each label gets wider.

Is "bought in past month" accurate?

It is Amazon's own figure, but rounded down and with no published method, so treat it as a floor rather than exact sales. It is useful for comparing listings with each other, especially listings in different categories, where the best sellers rank cannot be compared.

Is it a week or a month?

Both exist. Marketplace Pulse reported in 2023 that the time range is either a week or a month. Read the label's wording; the script in this guide keeps only labels the read marks as monthly.

Why do some Amazon products not show "bought in past month"?

Amazon does not show the label on every listing, and it has not published the rule. Treat a listing without a label as "no badge", not as zero sales.

How can I see how many sales an Amazon product has?

The label is the only sales figure Amazon shows publicly for a product. For more precision, compare it with the best sellers rank and save both over time, or use a sales estimator, which models units from the rank. Your own exact sales are in your seller account; other sellers' exact sales are not public.

Can I collect "bought in past month" for many products at once?

Yes. The script in this guide collects it for every organic result of a search, one credit per product plus five for the search. For a fixed list of products, read their ASINs directly.

Does "bought in past month" include every seller's sales?

The label sits on the listing, and Amazon has not said how it counts sales by different sellers of the same listing. Treat it as the listing's figure, not one seller's.

Sources

  1. https://www.marketplacepulse.com/articles/amazon-made-sales-data-public
  2. https://www.datawebot.com/llms.txt
  3. https://mcp.datawebot.com/setup.md
  4. https://www.datawebot.com/solutions/amazon-product-data-api
  5. https://www.datawebot.com/blog/amazon-scraper-api
  6. https://docs.python.org/3/library/csv.html
  7. https://docs.python.org/3/library/concurrent.futures.html

Read Amazon search results and product pages as JSON.

Every account starts with 200 free credits, with no subscription: a search costs 5 credits and a product read 1. Failed requests are free.

← All articles