How We Estimate Sales.
Every sales estimate across all 60+ tools is produced by a single, audited engine. One formula, one source of truth — so the number you see in Search is the same number that feeds Competitor Analysis, Stock Forecast, and Wholesale Scanner.
The Sales Estimation Engine
Every monthly sales number you see is produced by a three-step priority chain. AI is never part of this chain.
Amazon itself publishes a '1,000+ bought last month' badge on high-velocity listings. Keepa captures this figure and stores it as monthlySold. When available, this is the only number we use — no formula, no estimation.
When monthlySold is absent, we apply our piecewise-linear BSR→sales formula (estimateMonthlySales). This single function is defined once in keepa.ts and called by every tool. The formula is continuous at all boundaries — higher BSR always means lower sales with no jumps.
Claude, Grok, DeepSeek, and any other AI model are never used to generate a monthly sales number. AI is used for content (listing copy, reviews, keywords) — never for market size or demand figures.
Confidence Levels by BSR Range
Accuracy degrades as BSR increases. Here is what each confidence level means for interpreting the sales figure.
Dense Keepa data; many products have native monthlySold from Amazon's '1,000+ bought last month' badge.
BSR oscillates less frequently; formula coefficients validated against Keepa monthlySold samples.
BSR at this range can represent very different category-level demand. Category multipliers not yet applied.
Thin Keepa coverage; few products have monthlySold badges. Formula extrapolation risk is high.
Products this deep in the catalog may have near-zero real sales. Treat as directional signal only.
Tool-by-Tool Data Sources
Which source feeds which metric in each tool — no ambiguity.
Data Sources
Keepa maintains a complete historical record of every Amazon product's BSR, price, and review count. The monthlySold field (when present) comes directly from Amazon's backend badge ('1,000+ bought last month') and is always preferred over our BSR formula.
DataForSEO is the SINGLE source of truth for all keyword search volumes. Every tool that displays a volume number — Search Volume, KeySpy, Reverse ASIN, PPC Keywords — pulls from the same DataForSEO bulk_search_volume endpoint with identical normalization, so the same keyword always returns the same number across all tools.
Used to verify organic keyword rankings by checking live Amazon search results, and to power Google Trends seasonality signals in Search Volume analysis.
Claude AI is used ONLY as a fallback for keyword suggestions when DataForSEO returns no results. It is NEVER used to estimate monthly sales, BSR, or revenue. AI-generated data is labeled with the 'AI Est.' badge.
Grok is used in Reverse ASIN to discover which keywords an ASIN might rank for, using live web access. After Grok generates keyword candidates, DataForSEO bulk_search_volume overrides all volumes — so the final numbers come from DataForSEO, not AI.
Known Limitations
Every estimation model has limits. Here are ours — explained honestly.
BSR 5,000 in Books means ~4× more sales than BSR 5,000 in Sports & Outdoors. Our current formula does not apply per-category coefficients. We are validating category-specific curves and will ship them when we reach ±15% accuracy across 5+ categories.
A product with a BSR spike due to a lightning deal or coupon will show inflated sales estimates for that period. Keepa's monthlySold badge (when available) mitigates this as it's a rolling 30-day figure from Amazon.
Amazon shows both a top-level category BSR (e.g., Sports & Outdoors #5,000) and sub-category BSRs (e.g., Yoga Mats #12). Our formula uses the top-level BSR. Sub-category BSR would give a very different — and often inflated — sales estimate.
DataForSEO keyword volumes are updated monthly, not in real time. A keyword experiencing a sudden trend spike (e.g., a viral TikTok product) may show last month's volume for up to 30 days. We indicate the data date where available.
Our formula was calibrated against Amazon US data. For Amazon DE, UK, or TR, the BSR-to-sales relationship differs due to smaller catalog sizes and different shopper volumes. Treat non-US estimates as ±40% or worse until marketplace-specific calibration is complete.
Products with fewer than 30 days of BSR history produce unreliable estimates because their BSR fluctuates wildly during the launch phase. If Keepa's monthlySold is absent for a new product, treat the sales estimate as approximate.
Methodology Changelog
Unified engine: replaced bsrToSales() (wholesale/scan) and bsrToDaily() (stock-forecast) with the canonical estimateMonthlySales() from keepa.ts. All 60+ tools now use a single, identical formula.
Market-share endpoint now checks Keepa monthlySold before applying BSR formula — consistent with all other tools.
KeySpy and Reverse ASIN: normalized keyword map keys to ensure DataForSEO bulk_search_volume overrides Grok/ranked-keyword volumes correctly. Same keyword now returns same volume across all tools.
resolveKeywordVolumes() helper introduced as the single DataForSEO volume resolver for all keyword tools.
canonicalMonthlySales() introduced — priority wrapper that prefers Keepa monthlySold over BSR formula.
Questions about our methodology?
We publish our formula and coefficients openly. If you find an inaccuracy, we want to know.