Amazon Ads Marketing Analytics: Sponsored Products, DSP Performance, and Connecting Spend to Revenue

Amazon Ads Marketing Analytics: Sponsored Products, DSP Performance, and Connecting Spend to Revenue

Amazon Ads marketing analytics tracks performance across three core formats - Sponsored Products, Sponsored Brands, and DSP - using ACoS, TACoS, and New-to-Brand rate as your primary decision signals. Getting it right means separating what your ads paid for from what your store earned organically, then combining both into a single revenue view that flags when performance shifts.

Most DTC and ecommerce marketers running Amazon spend hit the same wall: Amazon Ads Campaign Manager shows clicks and spend, but doesn't tell you whether a revenue spike came from your Sponsored Products, your organic search rank, or a DSP campaign that ran two weeks ago. When the VP asks why ROAS dropped, you're pulling four separate reports with no common thread.

ACoS (Advertising Cost of Sales): ad spend divided by ad-attributed revenue. The primary efficiency metric for Sponsored Products and Sponsored Brands - lower is better, but too low can signal underinvestment in rank-building.

TACoS (Total Advertising Cost of Sales): ad spend divided by total store revenue - organic plus paid. The metric that tells you whether your ads are growing the business or merely subsidising it.

New-to-Brand (NTB) rate: the share of purchases from customers who have not bought from your brand on Amazon in the past 12 months. Your growth signal - rising NTB means you are acquiring, not just retaining.

DSP (Demand-Side Platform): Amazon's programmatic advertising layer. Runs display and video ads on and off Amazon. Measured differently from Sponsored ads - with eCPM, DPVR (Detail Page View Rate), and view-through conversions rather than ACoS.

Key takeaways

Amazon Ads Requires Three Primary Decision Metrics Not One

ACoS, TACoS, and New-to-Brand rate each measure a different dimension of Amazon Ads performance. No single metric captures the full relationship between paid spend and store performance - using only ACoS is as incomplete as using only ROAS for cross-channel paid media.

TACoS Reveals Whether Amazon Ads Are Growing the Business or Subsidising It

TACoS is ad spend divided by total store revenue including organic. A TACoS of 8% with a 12% ACoS means 33% of ad-attributed sales have an organic equivalent. When TACoS approaches ACoS, almost no organic sales are occurring independent of paid spend.

New-to-Brand Rate Above Sixty Percent Signals Genuine Customer Acquisition

A New-to-Brand rate above 60% means the campaign is acquiring net-new customers who haven't bought in the past 12 months. Below 40%, the campaign is primarily serving existing customers who would likely have purchased anyway - the correct primary signal for any advertiser with growth objectives.

A ROAS Drop in Sponsored Products Does Not Always Mean Underperformance

If organic search rank is improving simultaneously - visible in TACoS trending down while ACoS is stable - the brand may be transitioning from paid to organic traffic. This positive signal looks like deterioration in campaign-level reporting because the attribution model cannot distinguish the two effects.

Campaign Manager Hides the Source of Revenue Spikes Without TACoS Analysis

Campaign Manager reports clicks and spend without revealing whether a revenue spike came from Sponsored Products, organic rank changes, or a DSP campaign from two weeks prior. TACoS analysis over multiple time windows is required to separate these effects.

1. Separate your ad formats before reading any metrics

Amazon Ads analytics operates in two fundamentally different measurement contexts, and mixing them produces numbers that look plausible but mean nothing.

Self-service sponsored ads - Sponsored Products, Sponsored Brands, Sponsored Display - are direct-response: you pay for clicks, you measure ACoS against a 1-14 day conversion window. DSP is programmatic: you pay for impressions and measure awareness-to-purchase funnels over 30-day-plus horizons. A blended 28% ACoS might hide a 14% Sponsored Products ACoS (healthy) sitting next to a DSP campaign burning spend on irrelevant audiences.

How to verify: in Campaign Manager, use the "Ad type" breakdown filter. If Sponsored Products, Sponsored Brands, and DSP results are collapsed into a single summary row, you are reading averaged data - not actionable data.

Common failure mode: reporting connectors that pull the summary endpoint instead of campaign-level data will aggregate all formats by default. If your ACoS appears without a campaign-type label, check what your reporting tool is actually summing before drawing conclusions.

2. Calculate your ACoS baseline before optimising

ACoS is only meaningful relative to your gross margin and category - not as an absolute number. The industry benchmark for Sponsored Products across ecommerce categories is a 32-35% ACoS platform average, with top-performing accounts in mature categories consistently reaching 23-26%.

Your target depends on your unit economics. If your gross margin is 45%, an ACoS above 40% means you are paying to lose money on Amazon. If your margin is 65%, you have more room. The practical formula is straightforward:

Maximum ACoS = gross margin − target profit margin

TACoS is the signal that matters more at scale. A brand with 30% ACoS but 12% TACoS is generating strong organic lift - the ad spend is building search rank and review velocity that drives unattributed organic sales. A brand with 22% ACoS and 22% TACoS has no organic halo at all: every dollar of revenue still requires a paid click.

How to verify: pull Sponsored Products attributed revenue and total store revenue for the same 30-day window. Calculate TACoS as (ad spend ÷ total revenue) × 100.

Common failure mode: using Amazon-attributed revenue for your TACoS denominator instead of total store revenue. Campaign Manager only counts orders tagged to your ads - organic sales are invisible there. TACoS calculated this way will always equal ACoS, which is useless.

3. Use New-to-Brand rate as your growth signal

NTB rate separates customer acquisition from retention in your Amazon ad spend. A high ACoS with 70%+ NTB is defensible for a growth-stage brand scaling into new audiences. A low ACoS with 12% NTB means your ads are profitable but targeting people who already buy from you.

To access NTB data, you need Amazon Brand Registry. The metric appears in Sponsored Brands and Sponsored Display campaign reports - not Sponsored Products. Brands running only Sponsored Products have no direct acquisition signal from their Amazon campaigns.

NTB benchmarks vary by category, but a working rule: if NTB falls below 25% for a brand under three years old, ad spend is largely retargeting existing buyers. That is a defensible lifecycle-marketing choice, but it should be a conscious one rather than a default.

How to verify: download the campaign-level report with "New-to-Brand purchases" as a column. If the column is absent, you are either not in Brand Registry or your campaign type does not support it.

Common failure mode: NTB has a 60-day look-back window. Pulling NTB for a 7-day period produces unreliable numbers because Amazon is still attributing purchases from customers who last bought within 60 days. Use 30-90 day windows for any NTB analysis.

Prooflytics

Stop stitching platform exports together

Every channel in one brief — plus the memory of what each one actually drove.

Start free trialSee pricing

14 days free · no credit card

4. Track DSP metrics on a separate scorecard

DSP performance lives in Amazon's DSP Console - separate from Campaign Manager - and requires different KPIs:

• eCPM - cost per thousand impressions. The efficiency benchmark for your bid strategy.
• DPVR (Detail Page View Rate) - the share of impressions that resulted in someone viewing your product detail page. A proxy for creative relevance.
• NTB from DSP - the same metric as above, but sourced from display and video impressions rather than search clicks.
• Total sales impact - DSP uses view-through attribution, the methodology for which changed substantially in January 2026 (covered in the next section).

The most common DSP analytics mistake is evaluating it against an ACoS target. DSP operates on different time horizons: a Sponsored Products click typically converts within 3-7 days; a DSP display impression influences a purchase 14-30 days later. Evaluating both on a 7-day ROAS window will always make DSP look inefficient - because it is being measured against the wrong clock.

For teams managing Amazon alongside other paid channels, the marketing analytics for DTC brands guide covers how to build a unified ROAS view that accounts for these different attribution lags across channels.

What the data shows about Amazon ads marketing analytics benchmarks in 2026

The operational problem for anyone managing Amazon ad budgets right now: historical DSP performance data from before January 2026 was measured under a fixed 14-day view-through attribution window. Any year-on-year comparison you run is not apples-to-apples.

On January 1, 2026, Amazon retired its fixed 14-day view-through window for DSP campaigns and replaced it with ML-driven attribution. According to Code3's analysis of the change, the model now weights touchpoints dynamically based on observed conversion patterns rather than applying flat 14-day credit. Some DSP campaigns will show lower attributed revenue - where the ML model determines the view-through was not causal. Others will show higher attributed revenue - where the model gives credit to longer-lag conversions the old flat window missed entirely.

What this means for your reporting:

• Your 2025 DSP ROAS and TACoS baselines used the old measurement model. 2026 numbers will differ for structural reasons before accounting for any real performance change.
• Benchmarks published before Q1 2026 for DSP metrics are based on the old model and are not reliable reference points.
• Before optimising DSP based on year-on-year comparisons, establish a new Q1 2026 baseline with the ML-driven attribution in place.

Prooflytics amazon ads marketing analytics surfacing works through the daily briefing, which flags metric anomalies with date-aware context - so if your DSP ROAS shifts at a specific date, the briefing identifies whether the shift reflects a real performance change or a measurement methodology change. Separating those two causes is exactly the kind of question a static dashboard cannot answer.

5. Connect Amazon spend to your full channel revenue view

Amazon spend is one input in a multi-channel marketing mix. A performance marketer running Meta Ads, Google Shopping, and Amazon Sponsored Products simultaneously needs a single revenue view that contextualises sales across all three - not three dashboards with incompatible attribution logic.

The practical approach for most in-house teams is two-step: connect Amazon Selling Partner order data to your analytics layer for revenue and ASIN-level performance, then connect your ad spend from Meta, Google, and Amazon to the same view so you can calculate cross-channel TACoS.

For brands running Shopify alongside Amazon, connecting both platforms to the same intelligence layer shows whether an Amazon price promotion is cannibalising Shopify revenue or growing total demand - a question neither Amazon Seller Central nor Shopify Analytics can answer independently.

Teams evaluating multi-touch attribution tools specifically for Amazon - Triple Whale, Northbeam - often find the comparison between Prooflytics and Triple Whale useful for clarifying where each tool fits. Triple Whale specialises in Shopify/Amazon ROAS reconciliation at the order level; Prooflytics sits above the attribution layer and explains what changed and why across all channels, including revenue shifts that no attribution model surfaces.

For the complete ecommerce analytics framework - covering inventory velocity, CLTV, and creative performance alongside paid channel data - the marketing analytics for ecommerce guide covers the full decision stack.

You can see the full list of data source connections, including Amazon Selling Partner, in the Prooflytics integrations hub.

The bottom line

Amazon Ads marketing analytics works when you apply the right measurement logic at each layer:

• Separate ad formats first. Sponsored Products and DSP use different metrics, different attribution windows, and different evaluation horizons. Never blend them into a single ROAS.
• ACoS tells you ad efficiency; TACoS tells you business health. Build your break-even ACoS before optimising. Track TACoS divergence from ACoS as your primary signal of organic programme health.
• NTB is your growth indicator. Use 30-90 day windows. If NTB falls below 25% on a growth-stage brand, re-evaluate your audience targeting.
• 2026 DSP baselines need resetting. Amazon's ML-driven attribution replaces the fixed 14-day window. Year-on-year DSP comparisons are structurally different - establish a new Q1 2026 baseline before drawing performance conclusions.
• Cross-channel view requires connecting Amazon to your other channels. Selling Partner order data combined with Meta and Google Ads spend shows you total demand signals that no single platform dashboard can provide.

You can read independent reviews of Prooflytics on G2 and compare it to alternatives in the marketing analytics category.

Book a walkthrough to see how Prooflytics connects Amazon and your other paid channels into a single daily briefing.