Amazon KPI Guide 2026: 24 Metrics That Drive Profit

Key Takeaways

1. Track 24 KPIs in four groups: advertising, sales and conversion, inventory, and financial and account health.

2. ACoS equals CPC ÷ (conversion rate × order value), so fix conversion before you cut bids.

3. Set target ACoS from your pre-ad margin, and read it next to TACoS to see if ads build organic sales.

4. Keep ODR well under Amazon’s 1% limit, and count defects as well as percentages.

5. Review ad spend and weeks of cover together, so ads never push a nearly empty listing.

6. Atom11 joins ads, inventory, pricing and rank in one retail-aware dashboard, with root cause analysis for sales drops.

Amazon KPIs help sellers measure advertising efficiency, sales performance, conversion rates, and overall profitability. In 2026, tracking the right metrics is essential to understanding whether campaigns generate profitable sales or waste ad spend. 

Key Amazon KPIs include Advertising Cost of Sales (ACoS), Total Advertising Cost of Sales (TACoS), Return on Ad Spend (ROAS), conversion rate, click-through rate (CTR), and profit margin. Each reveals a different part of your business performance, from keyword effectiveness to organic sales growth. 

This Amazon KPI guide covers 24 essential metrics, their formulas, useful benchmarks, and how to use them to make smarter bidding, budgeting, and inventory decisions. 

What Is an Amazon KPI?

An Amazon KPI (key performance indicator) is a measurable number that shows whether your listings, ads, stock and account are moving toward profit. Revenue alone is not a KPI. A $100K-a-month seller at 8% net margin earns less than a $60K seller at 22%.

KPIs fall into two types. Leading KPIs such as click-through rate and Buy Box share move first and warn you early. Lagging KPIs such as net profit and refund rate confirm what already happened. A healthy dashboard has both.

Why Are Amazon KPIs Important?

They show where profit leaks. Fees, ads and returns can erase a sale’s margin. KPIs reveal which SKU or Amazon campaign is responsible.

They protect your account. Metrics such as Order Defect Rate carry thresholds. Miss them and Amazon can limit your selling privileges.

They guide spend. Ad budget follows the products with healthy conversion, stock and margin, not the ones that are simply loudest.

They make teams accountable. Agencies and in-house teams can agree on targets, not opinions.

They compare you fairly. Category benchmarks show whether a number is strong or weak for your niche.

How Do Amazon KPIs Work?

Every KPI follows the same chain, and breaking it is where most teams go wrong:

1. Data source. Metrics come from Seller Central or Vendor Central, the Amazon Ads console and your own cost data (COGS, shipping).

2. Calculation. A formula turns raw counts into a ratio, for example ad spend divided by sales.

3. Benchmark. The ratio is compared with a threshold, a category range or your own history.

4. Decision. The gap triggers an action: change a bid, reorder stock, fix a listing.

Two rules make the chain reliable. First, read KPIs in pairs: ACoS with TACoS, sessions with conversion, ad spend with inventory. Second, set targets from your own margin, because platform averages describe the market, not your business.

What Insights Can You Draw from Amazon KPIs?

Top 24 Amazon KPIs for 2026 at a Glance

The below section in detail explains 24 key Amazon KPIs, what they measure, and the advertising benchmarks to consider when evaluating campaign performance, sales, profitability, and growth.

Advertising KPIs (1-8)

1. ACoS (Advertising Cost of Sales)

ACoS is ad spend divided by ad-attributed sales. It measures how efficient your campaigns are at turning spend into sales. Lower is better, but only relative to margin.

Here is the relationship most guides skip. ACoS is built from three inputs:

ACoS = CPC ÷ (conversion rate × average order value)

Take an illustrative product with a $1.20 CPC, 10% ad conversion and a $30 order value. ACoS = 1.20 ÷ (0.10 × 30) = 40%. Lift conversion to 12% with better images or A+ content, and ACoS falls to 33% with no bid change. When ACoS is high, check conversion and price before cutting bids.

Set the target from the margin, not from a blog benchmark. Break-even ACoS equals your margin before ad spend. A 55%-margin product can run at 40% ACoS. A 12%-margin product cannot run at 15%.

Common misread: A falling ACoS is not always good. If you cut bids too far you lose impressions, and total sales fall with it.

Check first: Conversion rate and price, then CPC.

2. TACoS (Total Advertising Cost of Sales)

TACoS is ad spend divided by total sales, organic plus paid. It shows whether your brand leans on ads or is building organic momentum. A mature product with TACoS of 10-15% is generally considered healthy. Direction matters more than level: rising ACoS with falling TACoS usually means ads are lifting organic rank.

TACoS = (Total ad spend ÷ Total sales) × 100

Common misread: Cutting ad spend to lower TACoS can backfire when your organic rank depends on ad-driven sales.

Check first: Split total sales into ad and organic, then check the rank trend.

3. ROAS (Return on Ad Spend)

ROAS is ad sales divided by ad spend, the inverse of ACoS. A 25% ACoS equals a 4.0 ROAS. Use it when your team prefers a return multiple. Your target ROAS is 1 ÷ your break-even ACoS.

ROAS = Ad-attributed sales ÷ Ad spend

Common misread: A 4.0 ROAS is not automatically good. On a 20% margin product, break-even ROAS is 5.0, so 4.0 loses money.

Check first: Margin per SKU, before judging any ROAS.

4. CPC (Cost Per Click)

CPC is ad spend divided by clicks. It is the price of a visit and the first input to ACoS. Competition is rising: Amazon’s advertising revenue grew 26% to $19.8 billion in Q2 2026, nearly twice the growth of its online store revenue. Expect upward CPC pressure and track it by keyword type, not only in total.

CPC = Total ad spend ÷ Total clicks

Common misread: A rising CPC is not always a bidding problem. It can come from falling CTR or more competition on one keyword.

Check first: CPC by keyword and placement, not the account average.

5. CTR (Click-Through Rate)

CTR is clicks divided by impressions. Which ad appears depends on bid and expected click-through rate, so a weak CTR raises your effective cost. If CTR is low, test the main image, title and price before raising bids.

CTR = (Clicks ÷ Impressions) × 100

Common misread: A high CTR on a broad keyword can still lose money if those clickers do not buy.

Check first: Main image, title, price and star rating compared with competitors on the results page.

6. Ad conversion rate

This is orders divided by ad clicks, with a typical range of 8-15%. It tells you whether the traffic you pay for is the right traffic. A high CTR with a low conversion rate points to mismatched search terms or a weak listing.

Common misread: A low rate on a new campaign may simply reflect too few clicks. Judge it once you have enough data.

Check first: The search term report for irrelevant terms, then listing content.

7. Cost per acquisition (CPA)

CPA is ad spend divided by ad-attributed orders. It only means something next to profit per order. If CPA exceeds the profit you make on that order, the campaign loses money on the first sale and needs repeat purchases to recover.

CPA = Total ad spend ÷ Ad-attributed orders

Common misread: A CPA that looks small next to the product price can still be unprofitable. Profit per order is the test.

Check first: Profit per order, then negate search terms that never convert.

8. New-to-brand share

This is the share of ad sales or orders from shoppers who have not bought from your brand recently. It tells you whether ads are acquiring customers or recycling existing ones. A low share on branded campaigns is normal. A low share across non-branded campaigns means growth is stalling.

Common misread: A low share on branded campaigns is normal, because existing customers search your brand name.

Check first: The split between branded and non-branded campaigns, and any recent launches.

Sales and Conversion KPIs (9-13)

9. Unit Session Percentage

This is units ordered divided by sessions, Amazon’s version of conversion rate, and a strong driver of organic rank. Compare it within your category. A 10% rate is excellent in one and weak in the other.

Unit Session Percentage = (Units ordered ÷ Sessions) × 100

Common misread: A drop is not always a listing problem. It can be a lost Buy Box or a new competitor offer.

Check first: Buy Box and price against competitors, then reviews and ratings.

10. Buy Box percentage

This is the share of time your offer is the featured one on the product page. 85% or more for brand owners. Below that, ad clicks may land on a competitor’s offer, so your ad spend builds their sales. Check price, stock and ODR first.

Buy Box % = (Page views or time with your offer in the Buy Box ÷ Total page views or time the product was available) × 100

Common misread: A healthy account-wide figure can hide one top ASIN that keeps losing the Buy Box.

Check first: Price, stock and ODR, ASIN by ASIN.

11. Sessions and glance views

Sessions count unique visits to your listings and page views count every load. Vendors see a similar measure called glance views. Falling traffic with stable conversion is a visibility problem. Stable traffic with falling conversion is a listing, price or review problem.

Common misread: More traffic is not growth. If the extra sessions come from poor search terms, they only dilute conversion.

Check first: Traffic source (ads or organic) and the search terms behind the rise

12. Average order value (AOV)

AOV is revenue divided by orders. It feeds directly into ACoS, because a higher AOV lowers ACoS at the same CPC and conversion rate. Bundles, multipacks and add-on items are the usual levers.

AOV = Total revenue ÷ Number of orders

Common misread: Raising AOV through deep discounts can lift revenue while shrinking margin.

Check first: Margin on the bundle or multipack, not just its price.

13. Best Seller Rank

This is Amazon’s sales ranking within a category. Use it as a trend line, not a daily target, and plot it against ad spend to see whether your spend moves rank. Short-term swings are normal.

Common misread: Rank is relative and updates often, so a one-day spike is not a trend.

Check first: The 7-day and 30-day trend plotted against ad spend.

Inventory KPIs (14-18)

14. Weeks of cover

This is units on hand divided by average weekly sales. With an 8-week supplier lead time, cover below 10 weeks puts you at stock-out risk. A stock-out hurts twice: sales stop and the rank you built with ad spend can erode.

Weeks of cover = Units available (on hand + in transit) ÷ Average weekly units sold

Common misread: Cover calculated on normal sales looks safe until a promotion lifts demand. Prime Day can use it up quickly.

Check first: Lead time, units in transit and upcoming sales events.

15. Inventory turnover

This is cost of goods sold divided by average inventory value. A higher number means cash is not stuck in stock. A very high number can also mean you are running too lean.

Inventory turnover = Cost of goods sold (COGS) ÷ Average inventory value

Common misread: A very high turnover can mean you are running out of stock, not running efficiently.

Check first: Stock-out days per SKU.

16. Aged and excess inventory

This is stock sitting in fulfillment centers for a long time and collecting storage fees. Review it monthly, then discount, bundle or remove units before fees exceed their value.

Common misread: Aged stock is not always a product problem. Sometimes you over-ordered after a short spike.

Check first: Sales velocity trend and how reorder quantities are set.

17. Inventory Performance Index (IPI)

IPI is Amazon’s own score of stock health. A score below 400 can shrink storage capacity. Confirm the current threshold in Seller Central, as Amazon changes it.

Common misread: Dumping stock at a loss to lift the score can cost more than the storage limits it avoids.

Check first: Excess units, stranded units and in-stock rate.

18. Stranded inventory

This is unsellable stock at fulfillment centers, often caused by listing errors. It should be kept under 2% of inventory. It is a quick win, because fixing the listing usually frees the stock.

The overlooked link is between inventory and ads. Ads running on a nearly empty listing waste budget. Review spend and weeks of cover together.

Common misread: Stranded stock is usually a listing problem, not a warehouse problem.

Check first: Listing errors in Seller Central’s stranded inventory view.

Financial and Account Health KPIs (19-24)

19. Contribution margin per SKU

This is what remains after COGS, referral fees, FBA fees, storage, returns and ad spend. Check it at SKU level, because a healthy average can hide loss-making products.

Common misread: A margin that looks strong before returns can turn thin once refunds are counted.

Check first: Any SKU at or below zero, then its return rate.

20. Order Defect Rate (ODR)

ODR counts orders with negative feedback, A-to-z claims or chargebacks. Amazon’s target is under 1%. Count defects, not just percentages: at 2,000 orders a month, 1% is 20 defects, and a few bad weeks in Q4 can use your buffer. Many agencies alert at 0.5%.

Common misread: At low volume the percentage misleads: one defect on 50 orders is 2%.

Check first: The reasons behind negative feedback and A-to-z claims.

21. Late shipment rate

This is late shipments divided by shipments for seller-fulfilled orders. Amazon’s target is under 4%. Fix handling time settings or carrier cutoffs before peak season.

LSR = (Orders shipped late ÷ Total orders shipped) × 100

Common misread: Being under the limit is not safe in peak season, when carrier delays arrive late in the window.

Check first: Handling time settings, daily cutoff and carrier scans.

22. Pre-fulfillment cancel rate

This is orders you cancel before shipping, divided by total orders. Amazon’s target is under 2.5%. Most cancels come from stock errors, so keep inventory counts accurate.

Common misread: Most cancels trace to stock sync errors, not shopper demand.

Check first: Whether your inventory count matches your actual stock.

23. Refund rate

This is units returned divided by units sold. Apparel and electronics tend to run highest. Track it by SKU and treat returns as a cost in your margin, not a footnote.

Refund rate = (Units returned or refunded ÷ Units sold) × 100

Common misread: A flat overall rate can hide one SKU returning at 30%.

Check first: Return reasons by SKU, such as size, defect or not as described.

24. Valid tracking rate (VTR)

This is the share of shipments with valid tracking. GrowithAmazon lists the target as above 95%. Poor tracking can hurt account health and shopper trust.

VTR = (Shipments with valid tracking ÷ Total shipments) × 100

Common misread: Failures often come from the carrier or label process, but they still count against you.

Check first: Carrier integrations and when tracking is uploaded.

How to Diagnose a Sales Drop in Five Steps

1. Account and Buy Box. Did you lose the Buy Box, or is a listing suppressed?

2. Inventory. Is weeks of cover near zero, or did ads pause for low stock?

3. Traffic. Compare ad clicks with organic page views to see which side dropped.

4. Conversion. Check price changes, new competitor offers and recent reviews.

5. Ad efficiency. Only now look at ACoS, CPC and budget caps. A campaign that ran out of budget at noon looks like a demand problem but is not.

Done by hand across several reports, this takes hours. That is the gap a unified dashboard closes.

How Atom11 Turns Amazon KPIs into One Control Dashboard

The hardest part of KPI work is not the math. The data lives in four places: the ads console, Seller Central, a pricing tool and a spreadsheet. By the time you join them, the week is over.

Atom11 is an Amazon PPC and analytics platform built by former Amazon employees for sellers, vendors and agencies. Its approach is “retail-aware”: ad data is read alongside inventory, pricing, Buy Box and organic rank. Here is how that maps to the KPIs above.

Agencies can see every client account on one page and export it. Atom11 also supports Amazon Marketing Cloud reports and audience building.

A clarification for brand owners: Amazon Brand Analytics is Amazon’s own report on shopper search behavior. Atom11 does not replace it. It connects your ad and retail KPIs so you can act on what you learn there.

It is also not for everyone. If you run a few campaigns on a small ad budget, Amazon’s console and a spreadsheet may be enough. Atom11 pays off when ad spend, SKU count or client count makes manual joining a part-time job.

Want to see your own KPIs in one view? Contact Us, and our team will walk through your account.

Conclusion

The best Amazon KPI strategy in 2026 is ordered and connected. Start with account health, then conversion and Buy Box, then ACoS read beside TACoS, then inventory cover and margin per SKU. Set targets from your own margin and review weekly. The sellers who win are not tracking more metrics. They link ads, stock and price in one view and act before a small drop becomes a lost month.