Amazon Sales Estimates

How DataHawk estimates daily sales for tracked products and competitors; variant-level estimates by default, plus the V7 model methodology and table reference.

Sales Estimates give you a daily view of how many units any tracked product sells, including competitors you don't have access to.

⚠️

June 15, 2026; V7 model release. DataHawk now delivers sales estimates at the variant (child ASIN) level, with rollups to parent and brand. The full 2024+ history was reprocessed with the V7 algorithm on June 15, 2026.

All V5 and V6 tables are now deprecated aliases pointing to V7 data. Existing integrations continue to work, but you should migrate to the active table names below.

What business decisions can you make?

Combined with market share data, this powers decisions across three areas:

Product research

  • Identify high-demand, low-competition products before entering a category
  • Find underserved niches worth entering
  • Determine which of your variants actually drives most of the family's sales

Competitor benchmarking

  • Track whether a competitor is gaining or losing ground in your category
  • See which brands dominate the top 100 and how that's changing
  • Compare your estimated sales against category leaders
  • Measure your brand's share of total category volume

Inventory planning

  • Use sales trend data to forecast demand and avoid stockouts or overstock
  • Detect seasonality and how much it varies by category

Getting started

To receive sales estimates, you need to track at least one of the following:

  • Track a category → DataHawk collects Best Sellers Rank (BSR) data and calculates daily estimates for all top 100 products in that category. History is available immediately, and this is the primary way to get competitor estimates.
  • Track specific products (optional) → Estimates for individual ASINs you add manually. History starts from the day you begin tracking.
  • Create product segments (optional) → Group products with tags to compare your catalog against a specific competitor set.
  • Connect your Seller account (optional but recommended) → Your actual sales data helps calibrate the model for your own products, improving estimate accuracy.
💡

Once you track a category, DataHawk begins collecting and estimating daily. The more history accumulates, the more reliable trend analysis becomes.

DashboardWhat it showsBest for
Market Intelligence (Power BI)Market share trends, brand comparisons, top 100 analysisCategory-level competitive analysis
Product Dashboard (Power BI)Individual product performance and estimatesSKU-level and variant-level tracking
Analytics Essentials (Looker Studio)Cross-category overview with estimatesHigh-level monitoring

Capabilities overview

Product tracking vs category tracking

Category trackingProduct tracking
CoverageAll top 100 products in the category, including competitorsOnly the ASINs you explicitly track
HistoryUp to 2 years (from DataHawk collection start per marketplace), available immediately after category tracking startsFrom the day you start tracking
Best forMarket sizing, brand share, competitive landscapeFocused monitoring of a specific competitor set or your own catalog

Variant, parent, and brand levels

Amazon products often have multiple variations, such as sizes, colors, or pack counts, grouped under one parent ASIN. Parent-level estimates help with market share, brand comparison, and category sizing. Child-level estimates help with variant and SKU planning when you need to know which exact version of a product is driving sales.

Estimates are computed first at the family (parent ASIN) level, using the parent sales rank, which is the most reliable signal in the model. The family total is then allocated to each variant using public purchase signals (see How it works). The data is delivered at the variant (child ASIN) level, and rolls up to parent and brand for higher-level analysis.

LevelWhat it representsWhen to use
Variant (child)Estimated daily sales for each individual variantVariant-level decisions, specific SKU planning, understanding which size or color drives sales
ParentSum of all variants in the familyMarket share analysis, family-level comparisons, category sizing
BrandSum of all parents owned by a brandBrand-share and brand-vs-brand benchmarking
💡

Partially tracked families. When you track only one child ASIN from a product family, DataHawk uses the parent ASIN to discover the family structure, but it does not collect sales signals for the untracked children. Estimates for the tracked child rely on that child's own public purchase signals.

If the tracked child has no public purchase signal, DataHawk treats the absence of signals as a low monthly range of 0-50 units, represented as 25 monthly units. That can happen even when the parent family is a bestseller and the full family sells more than 30,000 units per month.

Data history by marketplace

Two full years of history are maintained to support year-over-year analysis.

Precision

Accuracy ratingMarketplaces
HighUS, SA, IE, AE, BR, GB, FR, DE, IN, IT, EG, MX, TR, ES, CA, JP
ModerateAU, SE, BE, NL, PL, SG

Estimates are approximations, not exact figures. Revenue estimates depend on the accuracy of captured price data. Sudden rank changes (flash sales, stockouts) may cause temporary spikes that the model smooths over subsequent days.

How it works

Collect Best Sellers Rank data

DataHawk monitors BSR for all products in tracked categories daily.

Apply the model per category

A rank of #50 in one category can represent a very different sales velocity than #50 in another. DataHawk calibrates estimates by category so rank position is translated into category-specific daily unit estimates.

Estimate at the family (parent) level

The model first computes the estimate at the family (parent ASIN) level, based on the parent sales rank. This is the most reliable signal available and produces the most accurate value in the model.

Distribute to child variants

The family total is then allocated across the children of the family, guided by the strongest evidence available for that family on that day:

  • Direct purchase signal. When public purchase evidence exists for a variant, its observed share is used.
  • Conservative fallback. When some children are covered by the catalog but lack their own signal, a conservative fallback share is applied.
  • Equal split. When no useful child-level signal is available, the family estimate is split evenly across all known variants.

Child estimates are not computed independently: they are allocated from the family total, so the sum of UNITS_CHILD across a complete family reconciles with UNITS_PARENT. Brand-level views aggregate parent estimates across a brand.

Adjust for sales events

Daily adjustments to estimates during high-velocity periods (Prime Day, Black Friday, etc.).

Fill gaps

When BSR data is temporarily unavailable, DataHawk applies linear extrapolation. Extrapolated rows are flagged (IS_EXTRAPOLATED = true).

Accuracy & reliability

Each row in the dataset carries a TRUST_INDEX score indicating model confidence for that product and date. Rows where extrapolation was applied are flagged with IS_EXTRAPOLATED. Estimates are most reliable in categories with stable bestseller patterns and strong public purchase-signal coverage.


Technical reference

Need active table names, deprecated aliases, grain, key columns, or migration notes?

On this page