Build a MER and marginal return dashboard in Claude

Builds the Get MCP Ads MER and marginal return template as a Claude dashboard: MER (the site's GA4 purchase revenue over the ad spend of Meta, Google Ads, TikTok, Pinterest and Snapchat added up within one currency) by day and week against an equivalent comparison period, each platform's own claimed revenue and ROAS beside the site's measured revenue with the gap between them, first-time and returning purchasers, the incremental return of the latest spend change, and the marginal return fitted on 13 to 26 weeks of the account's own history with its interval, fit quality and shape, refused with its reason when the history is too short, too flat or the fit poor.

Recipe

Intent A live page that says whether the business gets more revenue for its total ad spend and what the next unit of spend returns, to decide whether to raise, hold or cut the total budget and where.

Steps: 3
Tools: 1
Rules: 6
Failure modes: 4

skills/mer-dashboard.md

When to use it

Use it when

Someone asks for their MER, blended ROAS, marketing efficiency ratio, whether more spend still pays, diminishing returns, or how much to raise or cut the total ad budget.
A monthly budget review that weighs total spend against the site's revenue rather than each platform's own claims.

Do not use it when

Incrementality or causal lift: the dashboard shows observed relations on the account's own weeks; a holdout or geo test answers causality.
Clients without GA4 connected: there is no site revenue to divide by; say so and offer the scorecard.
Hosts without Claude's Dashboard artifact type: answer in the conversation with the matching reports instead.

What it needs first

Optional Meta Ads, Google Ads, TikTok Ads, Pinterest Ads, Snapchat Ads, Google Analytics 4

No single source is required: this one reads across whatever you have connected.

Preconditions

The organisation is on Team or Enterprise: ready-made dashboards are not offered on Free or Pro.
The account is among the chosen accounts, confirmed via list_accounts.

Evidence every conclusion must carry

Every figure carries its account, metric and window.
Anything below a readable sample size is reported as unjudged, not as zero.

Inputs
Input	Required	What it is
`window`		Analysis window.Default: last 30 days.

Procedure

In order. Every tool named is one the gateway ships, and links to its reference.

Steps

Resolve the client's workspace, its GA4 property and its ad accounts (one per platform among Meta, Google Ads, TikTok, Pinterest and Snapchat) with list_accounts.
Fetch the mer template with every account (accounts, one per platform, each with its currency, the GA4 property included).get_dashboard_template
Create a Dashboard artifact, write every document exactly as given in one batch, read its check, hand over with the lines given.

Decision rules

The observable condition, what it lets you conclude, and what takes the conclusion back.

When The template is refused with PLAN_REQUIREDConclude Say the dashboards come with Team and Enterprise and that only the account owner changes the plan; answer the question in the conversation meanwhile.
When The ad accounts are in several currenciesConclude Say MER is given only when every ad account shares the site's currency: no exchange rate is applied, so the page shows each currency's spend apart and no blended MER.
When The marginal return is refusedConclude Read the reason aloud (too few weeks, spend too flat, no clear relation, a season or trend driving it) and offer what would make it readable: more weeks, deliberate spend changes, events noted in the client journal.
When The person asks whether to raise the budgetConclude Compare the marginal return at the current spend and its interval with the margin the business needs, say it is an observed relation, not proof of cause, and that 10% more spend past the levels seen is an extrapolation.
When A promotion, sale or tracking outage moved revenueConclude Note it in the client journal with its dates (note_event): those weeks are left out of the fit at the next open.
When A reader sees a source fail to loadConclude Their own Get MCP Ads access decides: check that their connection covers this client's workspace and that their plan includes dashboards before changing the dashboard.

Where the agent stops

The readout changes nothing. Any change it motivates goes through Safe Writes: preview, human confirmation, then apply.

How it goes wrong quietly

The cases where the analysis is wrong and still looks right. Read them before trusting a number.

Editing the template's page or queries by hand loses the provenance of each figure: ask for changes by name instead.
Adding the platforms' claimed revenue into the site's revenue: the platforms overlap and over-claim; their sum only shows the gap.
Reading the average MER as the return of the next unit of spend: it holds the revenue that comes without ads too.
Calling the marginal return incrementality: budgets follow demand, so a season raises spend and revenue together.