How to Analyze Amazon Sales Data: A Practical Framework for Sellers

Amazon sales reporting is, at its core, combining Seller Central sales and traffic reports with fees, advertising, returns, inventory, and internal cost data to determine what drives revenue and contribution margin at the ASIN level. The objective is to have a process that is Deja Vu for each outlier, leading to an action in pricing, inventory, content, or advertising.

Amazon sales indicate that the revenue is insufficient to say whether an Amazon business is well established or not. Sales can be increasing while profits aren’t. Yes, ad performance might improve with overall demand staying flat. What's to stop this from being a conversion problem that is actually a traffic problem, stockout, or Buy Box loss?

Use this five-step process to keep your thinking on track:

Report names, menu paths, permissions, and data availability may vary by account type, marketplace, and Amazon interface rollout. Verify all fields in Seller Central, Vendor Central, or the advertising account before publishing screenshots, navigation instructions, or timing statements.

What Amazon Sales Data Analysis Should Help You Decide

An informative analysis begins with a decision, not a dashboard.

Many teams download every available report, add numerous metrics to a dashboard, and still cannot explain what changed. Start with one business question and collect only the data needed to answer it.

Your analysis should help you answer questions such as:

• Sales were down because traffic was down or because conversion was weaker?"
• Do we have revenue growth on a handful of SKUs and not enough contribution margin to cover overhead?
• Are the ad campaigns generating incremental demand, or is the account simply becoming more reliant on paid traffic?
• Are sales being constrained by stockouts, slow delivery, listing suppression, or lost Buy Box ownership?
• Which ASINs, SKUs, variations, or marketplaces should I focus on first?

What are the right metrics to look at? ACoS provides insight into how efficient an ad is, but it does not reveal whether a product is profitable. Sessions are a good way to measure traffic, but they won’t tell you if a stockout prevented demand from being met.

For broader definitions and formulas, see the Amazon seller metrics guide.

Step 1 — Define the Decision, Data Level, and Comparison Period

Do not start with a broad question such as:

How is the Amazon account performing?

That question is too open. It often leads to a crowded dashboard with no clear answer.

A better question is:

Why did contribution profit for Child ASIN A fall during the last complete four weeks compared with the previous four-week period?

This question defines the product, metric, time period, and comparison, giving the analyst a clear starting point.

Make a note before you export data.

• The scope of the product.
• The marketplace.
• The selling model.
• The reporting currency.
• Decision-maker: Identify the person responsible for approving or acting on the analysis.
• Is the review for a parent ASIN, child ASIN, SKU, campaign, or account?

This setup prevents team members from using the same dashboard to answer different questions.

Choose the Right Product Level

Select the appropriate level for your decision:

• Parent ASIN: See entire variation family.
• Child ASIN: Identify performance differences by size, color, flavor, or pack.
• SKU: Differentiate costs, fulfillment types, or inventory holders.
• Marketplace: Differentiate prices, fees, demand, and currency.
• Campaign or target: Review your ad results.

Do not compare or combine these levels without a clear reason. Parent-level totals can hide a poorly performing child ASIN, and grouped ASIN data can conceal a stock issue affecting one SKU.

Use Fair Comparison Periods

Compare full periods that have a comparable blend of selling days. Check:

• Promotions, coupons, and sales events.
• Holidays and seasonal demand.
• Price change.
• In-stock day.
• Listing suppression.
• Buy Box loss.
• Campaign launches or pauses.
• Variation family modifications.

Suppose an ASIN was available for only 14 days during one month. Comparing its total units with a full 30-day month would make the decline look larger than it really was.

Historical demand can help with planning. Deeper work on reorder amounts, lead times, and forecasting belongs in an Amazon inventory management process.

Choose the Review Cadence

• Daily: Follow alerts and short-term problems.
• Weekly: Do regular operating reviews.
• Monthly: Reconcile fees, returns, and earnings.

Step 2 — Gather the Right Amazon and Internal Data

Seller Central - where selling partners manage listings, sales, orders, inventory, payments, and related activities. Billing and Payment Options: Amazon also provides reports and dashboards that are useful for business reviews. See the official Seller Central overview for the latest information.

Rarely does one report answer all the questions. Good analysis is also a data synergy, where you integrate the Amazon reports with your brand’s own financial and operational data.

Sales and Traffic Data

Use sales and traffic data to relate visits to orders, units, and revenue.

Possible fields include:

• Sessions.
• Page views.
• Units ordered.
• Ordered sales.
• Conversion metrics.
• Buy Box or Featured Offer details.
• Note the report name, field name, and definition before combining sources. Different reports may have different denominators for conversions.

Orders, Cancellations, Returns, and Refunds

Track:

• Ordered units.
• Canceled units.
• Refunded units.
• Refunded revenue.
• Return fees.
• Useful return reasons.
• Reimbursements.

Returns may appear well after the original order. This means the order period and the financial period may not match.

A fast-growing ASIN can look strong in an order report but still produce weak profit after refunds and fulfillment costs are included.

Settlement and Fee Data

Data from settlements can contain information including:

• Referral fees
• Fulfillment charges
• Storage fees.
• Refunds.
• Reimbursements.
• Additional corrections.

Do not treat an Amazon disbursement as sales. A payout cycle may include transactions associated with orders from earlier periods.

Advertising Data

Review ad reports beside total sales, not by themselves.

Collect:

• Impressions.
• Clicks.
• Spend.
• Attributed orders.
• Attributed sales.
• Campaign type.
• Targeting type.
• Search terms or targets when needed.

For deeper query analysis, use the Amazon Search Term Report guide.

Brand Analytics

Eligible brands can use Amazon Brand Analytics dashboards to add grouped search and shopping data to account analysis.

Amazon directs users to Brands → Brand Analytics in Seller Central. Access depends on the selling account, Brand Registry relationship, and user role. Check Amazon’s official Brand Analytics page for current access rules.

Brand Analytics can help explain search demand and shopping behavior. It does not replace sales, fee, return, inventory, or COGS data.

This process focuses on your own Amazon account data. Competitor and category estimates belong in a separate Amazon market intelligence analysis.

External Traffic Attribution

Brands can measure off-Amazon traffic with Amazon Attribution. It uses tagged links to show how outside marketing affects shopping activity on Amazon.

Possible channels include:

• Search.
• Social media.
• Email.
• Display ads.
• Video.
• Affiliates.
• Influencers.

See Amazon’s official Attribution page for current setup and reporting details.

Internal Business Data

Amazon does not provide every cost needed to calculate contribution profit. Add:

• Product COGS.
• Freight and landed cost.
• 3PL fees.
• Variable shipping and logistics.
• Packaging costs.
• Product-level commissions.
• Internal promotion costs.
• Currency rules.
• Approved overhead, when used.

These figures may come from an ERP, accounting platform, purchasing records, 3PL system, or finance file.

Maintain a source register next to the main table. For every file, note the following:

• Report name.
• Account.
• Marketplace.
• Period requested.
• Time of download.
• Time zone.
• Currency.
• Owner.
• Refresh status.

Classifying each source as final, provisional, or reconciled makes auditing easier and allows another team member to repeat the process without guessing how the figures were prepared.

What “Real-Time” Amazon Data Really Means

Use the term “real time” carefully.

Amazon Marketing Stream sends hourly ad metrics and campaign changes in near real time through the Ads API. It does not create a complete live profit-and-loss report. See Amazon’s official Marketing Stream page for current details.

Use same-day data for alerts. Use matched and final data for financial decisions.

A SalesDuo Business Intelligence Dashboard may reduce the time spent collecting data. Teams still need clear rules for timing, costs, attribution, and ownership.

Step 3 — Normalize and Join the Data

Many of the biggest mistakes happen when reports are combined.

A useful data model needs shared fields. At a minimum, include:

• Date.
• Marketplace.
• Currency.
• Parent ASIN.
• Child ASIN.
• SKU.
• Campaign or target, when needed.
• Source report.
• Extraction date.

Create a Product Mapping Table

Connect each SKU to its:

• Child ASIN.
• Parent ASIN.
• Marketplace.
• Product line.
• Fulfillment method.
• Cost record.

One item type can have multiple SKUs. A SKU doesn’t necessarily exist in every marketplace. Internal product identifiers may also be different from Amazon IDs.

Add an effective date when the mappings change. Substituting historical mappings with today’s structure biases past results.

Match Time Zones and Dates

Amazon reports and internal systems may use different date formats or time zones.

Make one rule for the following:

• Daily sales reviews.
• Monthly financial reviews.
• Recent ad attribution.
• Return and settlement timing.

Write up the rule and apply it every time. Don't combine reports just because the date columns look alike.

Keep Sales Terms Separate

Ordered sales, shipped revenue, net sales, ad-attributed sales, and settlement cash are different figures.

Keep each measure in a separate field. This prevents teams from comparing ad-attributed sales with Amazon disbursements or treating gross orders as net revenue.

Handle Currencies Carefully

For the accounts you sell on multiple marketplaces:

1. Retain the original currency.
2. Record the exchange rate applied.
3. Apply only one rule for the conversion date.
4. Keep both the local and the reporting-currency amount.
5. Apply the same procedure to the periods compared.

Changing the currency method can create a false gain or loss.

Match Cancellations, Returns, and Repeated Rows

Units may be pulled before shipment due to cancellations. Returns can take weeks to come in. Settlement reports might have multiple rows for one order because fees and adjustments are shown on separate rows.

Do not remove the duplicate order IDs without verifying them. They could correspond to different financial transactions,n ot duplicates.

Recommended Fields for the Main Data Table

Keep raw reports in separate tabs and preserve their original values. Add formulas only in the normalized data and KPI tabs.

Recommended tabs include:

• Read Me.
• Source inventory.
• Raw sales and traffic.
• Raw advertising.
• Fees and returns.
• Inventory.
• COGS.
• Normalized data.
• KPI dashboard.
• Period comparison.
• Action log.
• Data checks.

Step 4 — Calculate the Metrics That Explain the Decision

A useful scorecard does not need every available metric. It needs the metrics that answer the business question.

Start with sales, units, sessions, and conversion.

If traffic falls while conversion stays steady, check visibility and availability. If conversion falls, review price, content, reviews, delivery, and offer quality.

ACoS and ROAS show ad results inside the attribution setup. TACoS compares ad spend with total sales. None of these metrics proves that ads created new demand.

Contribution profit is often more useful than revenue when deciding whether to scale—clearly state which costs are included.

Adjust sales velocity and inventory cover for:

• Stockouts.
• Inbound units.
• Lead time.
• Seasonal demand.
• Suppressed listings.

For more formulas, see the Amazon seller metrics guide.

Step 5 — Diagnose Common Amazon Performance Patterns

Do not act until you connect the change to a likely cause and check data quality, timing, inventory, and contribution profit.

Pattern 1 — Sales Increased but Contribution Profit Fell

Revenue growth can mask lower contribution profit.

Verify if:

• Ad spend increased at a higher rate than sales.
• Returns were up.
• Average price dropped.
• More coupons or discounts were offered.
• Fees for Amazon or fulfillment increased.
• Sales shifted toward lower-margin products.

Do not pause advertising immediately. First align attribution windows, separate branded and non-branded campaigns, and determine whether the additional spend supported a planned launch or promotion.

Pattern 2 — Traffic Fell but Conversion Stayed Stable

Stable conversion means the remaining visitors still buy at a similar rate.

Check:

• Organic ranking.
• Ad impressions and clicks.
• Campaign budgets.
• Search demand.
• Listing suppression.
• Buy Box ownership.
• Stock availability.
• Delivery times.
• Outside traffic.
• Variation visibility.

Do not rewrite the listing unless search relevance or click data supports that action. Lower traffic may come from weaker demand, lower rankings, fewer ads, or poor availability.

Pattern 3 — Traffic Stayed Stable but Conversion Fell

Review:

• Price.
• Promotions.
• Main image.
• Title.
• Reviews and ratings.
• Delivery promise.
• Buy Box ownership.
• Variation setup.
• Competitor offers.
• Product availability.
• Customer complaints.
• Ad-to-page fit.

Check child ASINs separately when possible. Parent-level conversion can hide a weak variation. Broader campaigns may also bring in lower-intent traffic.

Pattern 4 — Ad Sales Increased, but Organic Sales Stayed Flat

Review:

• TACoS.
• Branded and non-branded campaign mix.
• Search-term concentration.
• Organic ranking.
• Total units.
• New-to-brand signals, when available.
• Price and promotions.
• Attribution windows.
• Retargeting activity.

A higher attributed sales level may involve customers who would have made a purchase regardless. A launch, however, might still require some paid traffic before organic visibility builds.

Evaluate performance relative to the campaign goal. Amazon advertising management should be used only after the analysis indicates a genuine execution gap.

Pattern 5 — Demand Looks Healthy, but Sales Are Limited

Check:

• Sellable inventory.
• Inbound inventory.
• In-stock days.
• Listing suppression.
• Buy Box percentage.
• Delivery estimates.
• Variation availability.
• Vendor availability.
• Inventory placement.
• Fulfillment method.
• Price.

Use Buy Box analytics for offer problems. Use Amazon inventory management when stock or placement limits demand.

A Simple Order for Finding the Cause

When sales fall, ask:

1. Is the period complete and reliable?
2. Did traffic fall?
3. Did conversion fall?
4. Did price, product mix, or ads change?
5. Was the product fully available?
6. Did Buy Box ownership change?
7. Did cancellations, returns, fees, COGS, or logistics change?
8. Which cause has enough evidence for action?

Do not recommend a change until the data points to a likely cause.

Worked Example — From Revenue Growth to an Action Plan

The figures below are examples, not benchmarks.

Revenue rose by $10,000, while contribution profit fell by $4,500.

Advertising, fees, fulfillment, and returns need review. The data does not prove that any one of these caused the decline.

Before acting:

• Match ad and retail periods.
• Compare branded and non-branded campaigns.
• Review product mix.
• Match returns to the correct period.
• Check fees by SKU.
• Use the same cost rules in both periods.

If ad spend had stayed at $10,000 while all other current values stayed the same, contribution profit would have been $30,000.

This does not prove that the extra spend was wasted. It shows where to investigate first.

A sensitivity table changes one assumption at a time. The $5,000 difference shows the size of the advertising question. It does not prove that the added spend was unnecessary or unprofitable.

Run a Weekly Amazon Sales Review

A dashboard does not create accountability. A regular review process does.

Suggested Weekly Meeting Plan

1. Check for missing or incomplete data.
2. Find the products driving the biggest changes.
3. Review traffic, conversion, price, ads, inventory, returns, and costs.
4. Confirm the evidence.
5. Assign an owner, due date, and review metric.
6. Review past actions and results.

End the meeting with a decision log.

Rank issues by:

• Financial impact.
• Confidence in the evidence.
• Effort.
• Urgency.

A large decline with weak evidence may need more research before action. A smaller issue with clear evidence and an easy fix may deserve immediate ownership.

Record rejected actions so the team does not repeat the same discussion the following week.

Choose the Right Analysis Setup for Your Business

The right setup depends on catalog size, marketplaces, data sources, update needs, and team structure.

Before choosing a tool, test the process manually with one marketplace and a small group of ASINs.

Confirm that:

• Reports join correctly.
• Formulas follow finance rules.
• The action log leads to useful decisions.
• Owners understand their tasks.
• Review metrics are clear.

Once the method works, automate repeated collection and refresh steps. This prevents technology from hiding weak definitions, missing costs, or unclear ownership.

A spreadsheet may become inefficient as SKU count, marketplace count, currencies, connected systems, refresh frequency, or team size increases.

For comparisons, see the Amazon analytics tools guide. For automated reporting, consider the SalesDuo Business Intelligence Dashboard after the team agrees on definitions and rules.

Common Data and Analysis Mistakes

Check these points before approving a finding:

• Incomplete periods: A partial week is not a full week.
• Mixed time zones: One order may fall on different dates in different systems.
• Recent ad attribution: Attributed sales may still change.
• Late cancellations and returns: Gross orders can overstate results.
• Mixed ASIN levels: Child problems may disappear in parent totals.
• Repeated rows: One order may appear on several fee lines.
• Sales treated as cash: Orders, net revenue, and payments are different.
• Profit without COGS: Amazon cannot provide costs that only the brand knows.
• Ignored in-stock days: Low unit sales may come from poor availability.
• Assumed causation: Two metrics moving together do not prove that one caused the other.
• Universal targets: Goals vary by margin, category, and product age.
• Changed currency rules: Different methods can create false changes.
• Mixed fulfillment models: FBA and merchant-fulfilled costs differ.
• Changed mappings: Current parent-child links may not match past ones.

Every conclusion should connect to a source, calculation, and clear set of assumptions.

Turn Amazon Sales Analysis Into Action

A reliable Amazon sales review follows five steps:

1. Define the decision and comparison period.

2. Collect the required Amazon and internal data.

3. Normalize the reports.

4. Diagnose the variance with a focused metric set.

5. Assign the action, owner, due date, and recheck metric.

Use the downloadable Amazon Sales Analysis Workbook to organize source files, mappings, formulas, changes, and tasks.

A spreadsheet can work for a limited catalog and a few data sources. When matching reports takes more time than using them, the SalesDuo Business Intelligence Dashboard can provide a more organized reporting system.

For help turning findings into action, book a 1:1 growth call with SalesDuo.

Want to 5X Your Revenue?

Let's Connect!

Book Your 1:1 Growth Call

Want to 5X Your Revenue?

Let's Connect!

Book Your 1:1 Growth Call

FAQs About Amazon Seller Account Protection

1. What Data Do I Need to Analyze Amazon Sales?

Collect sales, units, traffic, orders, ads, inventory, returns, and Amazon fee data.

To calculate contribution profit, add COGS and variable logistics from internal systems.

The exact source set depends on whether you are reviewing traffic, advertising, inventory, or profit.

2. Where Can I Get Amazon Sales Data?

Sellers can download reports through Seller Central. Vendors may use Vendor Central when their role allows it.

Approved teams may also use SP-API and Amazon Ads API connections. Internal costs usually come from accounting, ERP, purchasing, or 3PL systems.

3. Can I Analyze Amazon Sales Data in Real Time?

Some ad signals are available in near real time through Amazon Marketing Stream.

Full sales and profit analysis takes longer because orders, attribution, returns, and fees may update later. Use fast data for alerts and matched data for financial decisions.

4. How Do I Analyze Historical Sales Trends on Amazon?

Compare similar periods and adjust for:

• Seasonality.
• Promotions.
• Price.
• In-stock days.
• Delivery speed.
• Advertising.
• Variation changes.
• Partial periods.
• Marketplace differences.

Review both totals and rates so stockouts do not look like weak demand.

5. What Is the Difference Between Amazon Sales Data and Amazon Brand Analytics?

Amazon sales data shows your account’s orders, units, traffic, fees, ads, and inventory.

Brand Analytics gives eligible brands grouped search and shopping data. Use it to add demand context, not as a full profit report.

6. Which Metrics Matter Most for Amazon Sales Analysis?

The decision determines the metrics.

For a sales drop, start with traffic, conversion, price, availability, and units.

For ads, review spend, attributed sales, ACoS, ROAS, TACoS, and total sales.

For profit, use contribution profit and contribution margin.

7. Why Do Amazon Sales and Ad Reports Disagree?

They track different events and use different timing rules.

Differences may come from:

• Attribution windows.
• Click-based or view-based attribution.
• Time zones.
• Canceled orders.
• Returns.
• Report updates.
• The difference between attributed and total sales.

8. How Do I Calculate Profit From Amazon Sales Data?

Start with net sales and subtract:

• COGS.
• Amazon fees.
• Fulfillment costs.
• Ad spend.
• Refunds.
• Returns.
• Variable logistics.
• Other product-level costs included in your company’s policy.

State what is included and use the same cost rules across products and periods.

9. Do I Need an Amazon Analytics Tool?

Not always.

A spreadsheet may be enough for a small catalog. A BI platform becomes more useful with several marketplaces, many SKUs, frequent updates, or several data sources.

Define the process before choosing software.

10. How Often Should I Review Amazon Sales Performance?

Use daily checks for urgent issues, weekly reviews for diagnosis and action, and monthly reviews for financial matching.

High-volume accounts may need more alerts. Low-volume products may need longer periods before the data is reliable.

About the Author

Giridhara Prasad is an Associate Director at SalesDuo and a startup enthusiast. With extensive expertise in e-commerce, this ex-Amazonian has been instrumental in driving success for businesses worldwide. Apart from his passion for creating innovative sales strategies and optimizing online retail experiences, Giri finds interest in watching and playing sports, including starting to play pickleball, traveling, and exploring political science, and philosophy.

Struggling with

Amazon Growth?

Book Your 1:1 Growth Call

Struggling with

Amazon Growth?

Book Your 1:1 Growth Call