One number that includes all the spend and the actual orders.

The number your dashboards cannot produce

Ask Meta what it made you and it will tell you. Ask Google and it will tell you a different story about the same customer. Amazon reports its own slice. Each platform is being honest inside its own attribution model, and the models overlap, so the sum is not a real number.

Blended reporting steps around the argument instead of trying to win it. Total what you spent everywhere. Total what the store actually took. Divide. The result is deliberately unsophisticated, and that is exactly why it survives contact with reality: it cannot double count, because it never counted conversions in the first place.

It is the number a founder actually runs the business on, and until now it has been a spreadsheet someone rebuilds every Monday.

What the call returns

get_blended_performance takes whichever platform identifiers you already have, a Meta account_id, a Google customer_id, an amazon_profile_id, a Shopify shop, and all four are optional. It returns total ad spend against Shopify revenue with blended MER, ROAS and ACOS.

Two design decisions in that tool are worth calling out. First, a platform you have not connected contributes zero and is reported in a platforms_connected field rather than failing the whole call, so a partial setup still gets you a partial answer instead of an error. Second, it is a non-LLM read, described in its own documentation as cheap and safe to call often, unlike a full account analysis. You can put it at the top of every conversation without thinking about cost.

Your assistant will normally resolve the identifiers itself through get_my_accounts, list_amazon_profiles and list_shopify_stores rather than asking you to paste them.

The caveats, on the page rather than in a footnote

Three limits are baked into this data, and Muze prints them in the tool responses instead of leaving you to discover them.

Amazon spend on this connection is ad-attributed only. Muze holds the Amazon Ads campaign management scope and not SP-API, so Amazon marketplace sales are not included in the revenue figure and there is no total Amazon revenue number to be had here. Second, spend and revenue are summed without currency conversion, which means the blended figure is only meaningful when your connected accounts share a currency. Third, get_shopify_revenue returns null for revenue rather than a fabricated number when a store transacted in more than one currency in the window, and gives you a per-currency breakdown instead.

A tool that returns null when it does not know is more useful than one that returns a confident average. The response's own notes field lists which caveats applied to your specific call.

Around the blended number

The single call is the headline, but the supporting reads are what make the answer actionable.

get_shopify_revenue gives order revenue for a trailing period summed by currency, get_shopify_orders walks the orders themselves with pagination, and get_top_products ranks the catalogue by sales so you can see which products are carrying the blended number. On the Amazon side, set_amazon_revenue_estimate stores a monthly total revenue estimate used for TACOS, and it is worth being clear that this is a Muze-side setting you provide, not a figure pulled from Amazon.

If the numbers look wrong, the diagnostics are tools too. get_shopify_sync_status reports the state of the latest catalog and orders sync, get_shopify_diagnostics returns row counts and webhook state, and start_shopify_sync kicks off a background refresh that changes nothing in your store.

Blended is a decision tool, not an attribution model

We are not claiming to have solved attribution. Nobody has. Blended MER will not tell you which channel deserves the credit for a specific order, and if a page tells you it can, be suspicious.

What it does is answer the question you actually act on: at this total spend, is the business making money, and did that change this week. Use per-platform comparisons for diagnosis and the blended figure for the decision. The two together are a lot more useful than either alone, which is the argument for having all four platforms behind one connection in the first place.

What to type

These are prompts, not commands. Any MCP client will do: ChatGPT, Claude, Cursor or anything else that speaks the protocol.

• “What was my blended ROAS last month across Meta, Google, Amazon and Shopify?”
• “Total my ad spend this week and set it against actual Shopify orders.”
• “Is my blended MER better or worse than last month, and which platform moved it?”
• “Show me my top selling products and how much I spent advertising this month.”
• “Check my Shopify sync status, then give me the blended numbers for the last 30 days.”

The tools behind it

Thirteen of the fifteen Shopify tools are reads, and the two writes are a data refresh and applying an action on a Shopify agent card. Reporting does not need write access.

• get_blended_performance (read): Total ad spend against Shopify revenue with blended MER, ROAS and ACOS.
• get_shopify_revenue (read): Order revenue by currency. Returns null rather than guessing on mixed currencies.
• get_shopify_orders (read): The underlying orders, paginated.
• get_top_products (read): Sales-ranked products behind the revenue figure.
• set_amazon_revenue_estimate (write): Your monthly total revenue estimate for TACOS. A Muze-side setting.
• get_shopify_diagnostics (read): Sync health: catalog rows, order rows, webhook state.
• start_shopify_sync (write): Background catalog and orders refresh. Changes nothing in your store.

Platforms covered. Meta Ads, Google Ads and Amazon Ads for spend. Shopify for revenue.

Nothing spends without you

• The blended read is non-LLM and cheap, and is documented as safe to call often.
• Missing platforms contribute zero and are reported as not connected rather than failing the call.
• Revenue is returned as null with a per-currency breakdown when a store used more than one currency, rather than a fabricated single number.
• Amazon figures are ad-attributed only. Marketplace sales are not in this number and Muze says so in the response.

Frequently asked questions

What is blended ROAS?

Total revenue divided by total ad spend across every channel, rather than the platform-reported figure each ad network claims for itself. It cannot double count conversions because it does not count conversions at all, which is what makes it the number most operators actually steer by.

Can AI calculate my blended ROAS automatically?

Yes. get_blended_performance totals Meta, Google and Amazon ad spend against Shopify order revenue and returns blended MER, ROAS and ACOS in one call, so your assistant can produce the figure without you exporting anything.

Does it include Amazon marketplace sales?

No, and this is the most important caveat on the page. Muze holds the Amazon Ads campaign management scope, not SP-API, so Amazon figures are ad-attributed only and marketplace revenue is not included in the total.

What if I sell in multiple currencies?

The blended total sums without currency conversion, so it is only meaningful when your accounts share a currency. get_shopify_revenue is explicit about this: when a store transacted in more than one currency in the window it returns null for the single revenue figure and gives you a per-currency breakdown instead.

What if I have not connected all four platforms?

You still get an answer. Any platform you have not connected contributes zero and is reported as not connected in the response rather than failing the call.

Does this replace attribution?

No. Blended reporting deliberately avoids the attribution argument rather than resolving it. Use it to decide whether total spend is working. Use per-platform comparisons to work out where to look next.

Read next

Other features

Or see every Muze feature.