How to Automatically Connect Amazon Ads to Claude in 60 Seconds

One of the most common problems with Amazon Ads is not that you don’t have enough data. It’s that you have too much of it.

Hidden deep within your Amazon Seller Central account, buried under layers of Campaign Manager tabs and Bulk Operations files, is the secret to your next 20% growth. But right now, that secret is trapped in a spreadsheet you don’t have time to open.

Most advertisers spend their weeks playing data janitor. They download, they filter, they pivot. Then they manually look for the leaks, such as the ASINs bleeding budget, the bidding wars they can’t win, and the high-ACOS campaigns that are quietly eating their profit.

What if you could get instant AI analysis across your entire Amazon Ads account, along with actionable insights and recommended next steps?

By automatically connecting Amazon Ads to Claude via the Windsor MCP, you stop being a data entry clerk and start being a strategist.

🚀 Start your 30-day free trial at Windsor.ai and build your first Amazon Ads to Claude integration right now.

You can audit your entire business through natural language and generate clear, visual reports like this one in just seconds.

With Amazon Ads to Claude integration, you can simply ask:

• “Which products have a high Click-Through Rate but a conversion rate below my break-even point?”
• “Which seasonal campaigns should I kill based on the last 14 days of SKU-level profit?”
• “Claude, find the ‘attribution gap’—which auto-campaigns are cannibalizing my organic sales?”

Claude doesn’t just read numbers; it understands your brand context. In moments, it surfaces optimization levers you’d normally spend hours hunting for manually.

Stop digging through spreadsheets and start making winning decisions. Here’s how Windsor.ai bridges the gap between raw data and impactful insights in under a minute.

The 60-second setup to connect Amazon Ads to Claude via Windsor MCP

Just three super simple steps. No code. No developer needed.

1. Go to onboard.windsor.ai and select Amazon Ads as your data source.

2. Connect your Amazon Advertising account via the standard OAuth flow.

3. In Claude, open the Windsor.ai connector and click Connect.

That’s it. Your Amazon Ads data is now live inside every Claude conversation and ready for automated reporting and analysis.

📖 Full guide: windsor.ai/documentation/windsor-mcp/how-to-integrate-data-into-claude/.

What to ask Claude about your data: Amazon Ads analytics prompt ideas

There’s a lot you can do with your Amazon Ads data in Claude—whether it’s generating a simple weekly performance report or tackling more advanced tasks, like identifying wasted spend, spotting campaigns that are cannibalizing each other, or uncovering the products that are quietly eating into your profit.

Feel free to copy and paste these prompts directly into your Claude chat to get started immediately.

If you’re running Sponsored Products campaigns

SP is where most Amazon ad spend lives. It’s also where most waste hides: in auto campaigns that found the wrong targets, in broad match keywords that drifted, in ASINs that get impressions but can’t convert because the listing isn’t ready.

The zero-conversion spend audit

Start here. Every account has it. This prompt finds it fast.

Prompt:

Pull all search terms from my Sponsored Products campaigns for the last 60 days. Filter to search_term records where impressions > 100 and orders = 0. Group by campaign_name and targeting_type (auto vs. manual). Sort by spend descending. Show total wasted spend at the top. For the top 20 terms by spend with zero orders, flag whether they also appear as a current keyword in a manual campaign — if they do, that's a negative keyword gap.

ACOS by ASIN: finding your profit drains

Blended ACOS across a campaign tells you very little. ACOS by ASIN tells you which products you’re subsidising with ad spend and which ones are generating margin. There’s almost always a handful of ASINs in every account where the unit economics simply don’t support the bids being used.

Prompt:

Break down spend, sales, orders, acos, and clicks by advertised_asin for all Sponsored Products campaigns in the last 30 days. Rank by spend descending. Flag any ASIN where acos exceeds [your target]% and spend is above $200. Then cross-check: for those flagged ASINs, what is the average cpc and average conversion_rate? High CPC + low conversion_rate usually means either a listing problem or wrong keyword targeting — note which pattern each flagged ASIN fits.

Match type cannibalization: when your campaigns compete against each other

You’ve built out exact match campaigns for your best keywords. But your auto and broad campaigns are still bidding on the same queries — and often at higher CPCs, because they haven’t been optimized. The result: you’re winning the auction against yourself and paying more than you should.

Prompt:

Show me search terms from the last 30 days that appear in both auto/broad campaigns AND exact match campaigns simultaneously. For each duplicate search term, compare: spend, orders, acos, and cpc between the two campaign types. Where auto/broad is outspending exact on the same query, flag it — those are cannibalisation cases where spend should be shifted to the exact match campaign and the search term negated from auto/broad.

Placement performance: is top-of-search worth the premium?

Amazon lets you apply placement bid multipliers to boost bids for top-of-search and product page placements. But most sellers set these blindly. The actual data often tells a different story.

Prompt:

Pull performance by placement_type (top_of_search, rest_of_search, product_page) across all Sponsored Products campaigns for the last 30 days. For each placement, show: spend, sales, orders, acos, cpc, and ctr. Calculate the ACOS premium of top_of_search vs. rest_of_search. For campaigns where top_of_search acos is more than 20 percentage points above rest_of_search, list the current placement_modifier and suggest whether reducing it would improve overall campaign ACOS without sacrificing volume.

If you’re running Sponsored Brands campaigns

SB campaigns are where Amazon’s funnel starts. They drive category awareness and brand search volume. But they’re also where measurement gets murky; attributed sales include a 14-day window and halo effects that can make SB look better or worse than it actually is.

SB vs. SP halo: understanding the true contribution of brand campaigns

Sponsored Brands attributed sales often include orders that would have happened anyway from organic search. The only way to understand true incrementality is to look at what SP performance does in the same period for the same ASINs.

Prompt:

Compare Sponsored Brands and Sponsored Products performance for the same advertised_asin values over the last 60 days. For each ASIN, show SB spend, SB attributed_sales, SB acos, SP spend, SP attributed_sales, SP acos, and total combined acos. Which ASINs have SB acos below SP acos — suggesting brand campaigns are pulling efficiently? Which have SB acos more than 2x SP acos — a signal that SB spend may be over-attributed or inefficient?

Keyword-level SB performance: which terms justify Sponsored Brands spend

Not every keyword needs both SP and SB coverage. Running both on the same keyword at similar bids means you’re paying for two positions in the same auction, one of which may be cannibalizing the other.

Prompt:

Pull SB keyword performance for the last 30 days: keyword_text, match_type, spend, clicks, orders, attributed_sales, acos, and ctr. Identify keywords where SB ctr is below 0.3% despite high impressions — these are likely wrong for brand format ads. Then find keywords where both SB and SP are active and SB acos > SP acos by more than 15 points — those are candidates for SB bid reduction or pausing.

If you’re running Sponsored Display campaigns

SD is the most opaque of Amazon’s three ad types. Remarketing audiences, category targeting, competitor ASIN targeting — all with a last-touch attribution model that makes ACOS look artificially good for some segments.

Audience vs. contextual: where SD spend actually performs

Prompt:

Break down Sponsored Display performance by targeting_type for the last 30 days: audience-based vs. contextual (product and category targeting). For each type, show spend, attributed_sales, orders, acos, ctr, and dpv (detail page views). Then within audience targeting, break by audience_name — which remarketing audiences (views remarketing, purchases remarketing) have the lowest acos? Flag any contextual targets where spend exceeds $100 and attributed_sales is zero over the full period.

Competitor ASIN targeting: is it converting or just burning budget?

Targeting competitor ASINs with SD placements is high risk, high cost. Conversion rates are typically lower because you’re interrupting a competitor’s customer. Most accounts are overspending here without knowing it.

Prompt:

Show all Sponsored Display campaigns using product ASIN targeting (not category). For each targeted_asin, show: spend, clicks, orders, acos, and detail_page_views. Sort by spend descending. Flag any targeted_asin where spend is above $50 and orders = 0. Group results into: competitor ASINs (brand name doesn't match your brand) vs. owned ASINs (cross-sell targeting). Calculate total wasted spend on zero-converting competitor targets.

Bonus: the questions Amazon Ads alone can’t answer

Here’s where it gets interesting. Amazon Ads tells you ad-attributed performance. It doesn’t tell you organic rank, total revenue, margin per unit, or what happened to BSR when you cut a campaign. But Windsor connects to your other data sources, too. You can connect all the required platforms and then ask Claude questions that span all of them.

TACOS: the metric Amazon doesn’t show you

Total Advertising Cost of Sale — ad spend as a percentage of total revenue, including organic — is the metric that actually tells you whether your ad investment is building or consuming your business. To calculate it, you need Amazon Ads spend alongside Seller Central or Vendor Central revenue. Connect both through Windsor and ask the following.

Prompt:

Using Amazon Ads spend data and Amazon Seller Central total sales data, calculate TACOS by ASIN for the last 30 days: tacos = total_ad_spend / total_revenue (organic + ad). Sort by TACOS descending. Flag any ASIN where TACOS exceeds [your target]% — these are products where advertising is consuming more than an acceptable share of total revenue. For the top 5 by TACOS, show the trend: is TACOS improving or worsening week-over-week?

Ad spend vs. BSR movement: are your campaigns actually moving rank?

Amazon’s flywheel runs on sales velocity. Advertising is supposed to generate that velocity, which improves organic rank, which generates more organic sales. But some campaigns spend heavily without moving BSR at all. Connect Amazon Seller Central data alongside Amazon Ads and try this prompt idea.

Prompt:

For my top 10 ASINs by ad spend last month, pull weekly Amazon Ads spend vs. weekly BSR movement from Seller Central. For each ASIN, is there a correlation between spend spikes and BSR improvement? Are there any ASINs where spend is consistently high but BSR has been flat or declining for more than 3 weeks? Those products may have listing issues that advertising can't fix — rank them by spend and flag them for listing audit.

Margin-adjusted ACOS: what your ad reports don’t account for

An ACOS of 25% looks fine until you factor in that the product has a 30% margin. At that point, advertising is consuming the entire profit. Every team should be running this analysis. Almost nobody does it weekly because it requires joining ad data with cost data manually.

Prompt:

Join Amazon Ads performance data with product margin data from [your other sources]. For each advertised_asin, calculate: reported acos, product_margin%, and adjusted_acos (spend / gross_profit, where gross_profit = sales * margin%). Flag any ASIN where adjusted_acos exceeds 100% — you are spending more on ads than you are making in gross profit on those products. Sort by total spend on those unprofitable ASINs. This is your most urgent optimisation list.

The Amazon Ads fields Windsor sends to Claude

While Windsor.ai supports hundreds of attributes, these core metrics and dimensions are the ones that drive 90% of high-level strategic audits. They allow Claude to see past the surface level of your account.

🔍  Explore the full list at windsor.ai/data-field/amazon_ads/.

Performance & efficiency metrics

These metrics allow Claude to calculate your actual profitability and identify budget leaks across different attribution windows:

• totalcost: The total advertising spend (also available as cost or spend).
• sponsored_products_campaign__attributedsales14d: The aggregate value of sales occurring within 14 days of an ad click (Windsor also supports 1d, 7d, and 30d windows).
• sponsored_products_campaign__acos: Pre-calculated Total Advertising Cost of Sales.
• sponsored_products_campaign__roas: Total return on advertising spend (Revenue divided by Cost).
• clicks &impressions: The raw traffic data used to audit reach and engagement levels.

Strategic dimensions

These dimensions allow Claude to “slice and dice” your account to find the root cause of performance shifts:

• ad_type: Differentiates performance between Sponsored Products, Sponsored Brands, and Sponsored Display.
• marketplace: The specific Amazon marketplace (e.g., USA, UK, Germany) for global account auditing.
• sponsored_products_campaign__campaign: The specific campaign name Claude uses to group data by your internal strategy.
• sponsored_products_campaign__campaign_status: Identifies whether a campaign is currently active or paused.

New-to-brand (NTB) metrics

For Sponsored Brands and Display, Windsor provides specialized fields to track growth beyond your existing customer base:

• sponsored_brands_campaign_non_video__attributedordersnewtobrand14d: The number of first-time orders for products within your brand.
• sponsored_brands_campaign_non_video__attributedorderratenewtobrand14d: The efficiency of your brand-building efforts relative to clicks.

Conclusion

Amazon Ads rewards precision. Small accounts can get away with loose management. At scale, when you’re running SP, SB, and SD across hundreds of ASINs, the difference between a 25% ACOS and a 35% ACOS is not a bidding strategy. It’s the quality of your analysis.

Most teams are doing that analysis in spreadsheets, once a week, with data that’s already 48 hours stale by the time someone looks at it.

Windsor MCP gives Claude a live window into your Amazon Ads account. Not a snapshot. Not an export. The actual data, right now. Ask a question, get an answer. Ask a follow-up, get another. Keep asking until you know exactly where to move the budget and why.

🚀 Connect Amazon Ads to Claude — start your 30-day free trial: https://onboard.windsor.ai/app/amazon_ads.

FAQs

What’s the fastest way to get Amazon Ads data into Claude?

Windsor MCP. 60-second setup at onboard.windsor.ai, and your data is live permanently. The alternative is downloading bulk reports from Seller Central, cleaning them, and uploading to Claude, which gives you a flat file that’s stale before you’ve finished the analysis, with no ability to cross-reference campaign data against search term data without manual joins. There’s also no comparison.

What are the main ways to connect Amazon Ads to Claude?

• Windsor MCP (recommended): No-code, live API connection. SP, SB, and SD all included. Supports multi-account setups. Cross-source blending with Seller Central, Shopify, and 325+ other sources. 60-second setup.
• Manual bulk report downloads: Download individual campaign, search term, or placement reports from Seller Central and upload to Claude. Works for isolated questions. Falls apart for anything that requires joining multiple reports or seeing current data.
• Amazon Advertising API (direct): Full programmatic access. Requires developer resources, API credentials, Amazon’s specific report request-and-poll flow, and ongoing maintenance. The right choice if you’re building a custom analytics product — not if you just want to ask Claude why your ACOS went up this week.

Which Amazon Ads campaign types does Windsor support?

All three: Sponsored Products (SP), Sponsored Brands (SB), and Sponsored Display (SD). You can query them individually or compare across types in a single Claude prompt. For example, comparing ACOS across SP and SB for the same ASIN to understand which campaign type is more efficient for each product.

Can Claude see my search term report data, not just keyword data?

Yes, and this is one of the most valuable things Windsor unlocks. The search_term field gives Claude access to the actual customer queries triggering your ads, with spend, orders, ACOS, and match type for each. This makes zero-conversion spend audits, cannibalisation analysis, and negative keyword identification all possible from a single conversational prompt.

Does Windsor support Amazon Seller Central data alongside Amazon Ads?

Yes. Windsor connects to both Amazon Ads and Amazon Seller Central (Amazon SP). Having both blended in Claude is what enables TACOS calculation, BSR-to-spend correlation, and margin-adjusted ACOS analysis — the three analyses most teams want but can’t do easily because the data lives in two separate places.

Can I manage multiple Amazon Ads accounts via Windsor MCP (e.g. for an agency)?

Yes. Windsor supports multi-account setups. For agencies managing multiple brand accounts, you can connect all accounts in Windsor and ask Claude to compare performance across them, flag anomalies in any account, or roll up spend and ACOS across your full portfolio.

Does Windsor pull attribution window data separately?

Yes. Amazon Ads offers 1-day, 7-day, 14-day, and 30-day attribution windows. Windsor can pull attributed sales by window, which matters significantly for fast-reorder categories (where 1-day attribution is more accurate) vs. considered purchases (where 14 or 30-day attribution is more representative). Knowing which window your ACOS figures are based on is critical for meaningful analysis.

Can Claude write back bid changes or negative keywords to my Amazon account?

No, Windsor MCP is strictly read-only. Claude can identify exactly which search terms to negate, which bids to reduce, and which ASINs to pause, and give you a prioritised list with the reasoning behind each recommendation. The changes are made by you in Seller Central or via your bid management tool. This keeps human judgement in the loop before anything touches live spend.