Every Monday: each Meta and Google Ads campaign's claimed purchases next to the orders Shopify or Stripe recorded with its UTM, and the campaign inflating most.
What lands on your board
Meta reports 41 purchases for Prospecting LAL last week; Shopify shows 23 orders with its UTM. Real ROAS 1.9, reported 3.4. 10 of the 41 were 1-day view-throughs.
Ads Manager says a campaign returned 3.4 times its spend. The store says something smaller, but only as one total, so nobody knows which campaign is padding its numbers. For ecommerce marketers and founders who move budget between campaigns every week, that missing per-campaign answer is the expensive part.
Every Monday the agent lines up each campaign's platform-reported purchases and value with the orders that actually carried that campaign's UTM in Shopify or Stripe, on the same conversion dates. You see reported ROAS and real ROAS side by side, the view-through share Meta folds in, and the one campaign where the gap is widest.
Call GET /act_<AD_ACCOUNT_ID>/insights with level=campaign, last Monday to Sunday as time_range, fields=campaign_id,campaign_name,spend,actions,action_values, action_attribution_windows=["7d_click","1d_view"] and action_report_time=conversion, so purchases sit on the day they happened. Read one action type only, omni_purchase (the Purchases column in Ads Manager), and keep its 7d_click and 1d_view values apart.
Run SELECT campaign.id, campaign.name, metrics.cost_micros, metrics.conversions_by_conversion_date, metrics.conversions_value_by_conversion_date FROM campaign WHERE segments.date BETWEEN '<monday>' AND '<sunday>'. The plain metrics.conversions sits on the click date, which would not line up with store orders.
From Shopify (whichever store you use), page through orders for the same dates (test:false, not cancelled) with currentTotalPriceSet, currentTotalTaxSet and customerJourneySummary { ready lastVisit { utmParameters { source medium campaign } } }. From Stripe instead, list Checkout Sessions with status=complete and that created range and read metadata.utm_campaign, or the customer's metadata for subscription first payments.
Normalize both sides: lowercase, trim, URL-decode (+ and %20 to spaces), then match utm_campaign to the campaign ID, or to the name when the person tags with names. Orders whose campaign value matches no campaign go in an "unknown campaign" row (renamed or paused campaigns); orders with no UTM stay unattributed and are never shared out; campaigns with spend and zero matched orders are flagged. Orders where ready is false are rechecked next run.
Per campaign: reported ROAS is platform value over spend, real ROAS is matched order revenue (after refunds, without tax) over spend, and the gap is the difference in revenue. Rank campaigns by gap in dollars, and say how many of Meta's purchases were 1d_view, since no UTM can ever confirm a view.
Title the card with the campaign that has the widest gap, both ROAS figures in the summary, the full table in the widget, and the unattributed and unknown-campaign totals in a callout. Use item_key insight-meta-ads-reported-vs-real-roas every week.
With BRIEFKIN_TOKEN set, the agent sends the result as one card that updates in place every run. Without it, you get the report in the chat.
Pick your agent, where the data comes from and how to schedule it; the commands below follow.
One command. It puts the SKILL.md in ~/.claude/skills and in ~/.agents/skills, which Codex, Cursor, Gemini CLI, Copilot CLI and OpenCode read too.
shellcurl -fsSL https://briefkin.dev/skills/meta-ads-reported-vs-real-roas/install | shAsk Claude Code:
promptRun the meta-ads-reported-vs-real-roas skillOr paste the prompt instead of installing:
promptFor last Monday to Sunday, get spend, purchases (omni_purchase) and purchase value per campaign from the Meta Marketing API insights (7d_click and 1d_view kept apart, action_report_time=conversion) and cost, conversions_by_conversion_date and conversions_value_by_conversion_date per campaign from Google Ads. Then take Shopify orders for the same dates with customerJourneySummary.lastVisit.utmParameters.campaign (or Stripe Checkout Sessions with metadata.utm_campaign). Match utm_campaign to campaign ID or name after lowercasing and URL-decoding. Give me a table per campaign: spend, reported purchases and value, matched orders and revenue, reported ROAS and real ROAS. List orders whose UTM matches no campaign and count orders with no UTM. Name the campaign with the widest gap and the one thing to change first.Mondays at 9:00, from cron on any machine where Claude Code is signed in:
crontab0 9 * * 1 claude -p "Run the meta-ads-reported-vs-real-roas skill"Add it with crontab -e. A laptop asleep at that time runs it when it wakes if you use launchd or Task Scheduler instead; GitHub Actions works for the API route.
Connect Briefkin once with curl -fsSL https://briefkin.dev/install | sh and every run lands on your board as a report card. How setup works
SKILL.md---
name: meta-ads-reported-vs-real-roas
description: "Every Monday: each Meta and Google Ads campaign's claimed purchases next to the orders Shopify or Stripe recorded with its UTM, and the campaign inflating most. Use when the user asks for this, or when it runs on a schedule."
---
# Reported vs real ROAS: Meta and Google Ads against Stripe or Shopify
Every Monday: each Meta and Google Ads campaign's claimed purchases next to the orders Shopify or Stripe recorded with its UTM, and the campaign inflating most.
Ads Manager says a campaign returned 3.4 times its spend. The store says something smaller, but only as one total, so nobody knows which campaign is padding its numbers. For ecommerce marketers and founders who move budget between campaigns every week, that missing per-campaign answer is the expensive part.
Every Monday the agent lines up each campaign's platform-reported purchases and value with the orders that actually carried that campaign's UTM in Shopify or Stripe, on the same conversion dates. You see reported ROAS and real ROAS side by side, the view-through share Meta folds in, and the one campaign where the gap is widest.
## Two ways to get the data
Use the API or MCP route when you have an MCP server, a CLI or an API token for the app. Otherwise, if you can drive a browser the user is signed in to (Claude in Chrome, a Playwright or Chrome DevTools MCP server), use the browser route. Say which one you used.
## Route A: API or MCP server
### What you need
- UTM tags on every ad: Meta URL parameters such as `utm_source=facebook&utm_medium=paid&utm_campaign={{campaign.id}}`, and a Google Ads final URL suffix such as `utm_source=google&utm_medium=cpc&utm_campaign={campaignid}`.
- Read access to both ad accounts (`ads_read` on the Meta ad account, a Read only Google Ads user or the Google Ads MCP server).
- Shopify Admin API `read_orders`, or a Stripe restricted key with Read on Checkout Sessions and Customers plus a checkout that copies the visitor's UTM into Checkout Session or customer metadata.
If something is missing, say what and stop; do not invent numbers.
### Steps
1. **Pull Meta's claim per campaign** (Meta Ads): Call `GET /act_<AD_ACCOUNT_ID>/insights` with `level=campaign`, last Monday to Sunday as `time_range`, `fields=campaign_id,campaign_name,spend,actions,action_values`, `action_attribution_windows=["7d_click","1d_view"]` and `action_report_time=conversion`, so purchases sit on the day they happened. Read one action type only, `omni_purchase` (the Purchases column in Ads Manager), and keep its `7d_click` and `1d_view` values apart.
2. **Pull Google's claim per campaign** (Google Ads): Run `SELECT campaign.id, campaign.name, metrics.cost_micros, metrics.conversions_by_conversion_date, metrics.conversions_value_by_conversion_date FROM campaign WHERE segments.date BETWEEN '<monday>' AND '<sunday>'`. The plain `metrics.conversions` sits on the click date, which would not line up with store orders.
3. **Collect real orders with their campaign** (Shopify): From Shopify (whichever store you use), page through `orders` for the same dates (`test:false`, not cancelled) with `currentTotalPriceSet`, `currentTotalTaxSet` and `customerJourneySummary { ready lastVisit { utmParameters { source medium campaign } } }`. From Stripe instead, list Checkout Sessions with `status=complete` and that `created` range and read `metadata.utm_campaign`, or the customer's metadata for subscription first payments.
4. **Join orders to campaigns on utm_campaign**: Normalize both sides: lowercase, trim, URL-decode (`+` and `%20` to spaces), then match `utm_campaign` to the campaign ID, or to the name when the person tags with names. Orders whose campaign value matches no campaign go in an "unknown campaign" row (renamed or paused campaigns); orders with no UTM stay unattributed and are never shared out; campaigns with spend and zero matched orders are flagged. Orders where `ready` is false are rechecked next run.
5. **Work out reported and real ROAS**: Per campaign: reported ROAS is platform value over spend, real ROAS is matched order revenue (after refunds, without tax) over spend, and the gap is the difference in revenue. Rank campaigns by gap in dollars, and say how many of Meta's purchases were `1d_view`, since no UTM can ever confirm a view.
6. **Post the card**: Title the card with the campaign that has the widest gap, both ROAS figures in the summary, the full table in the widget, and the unattributed and unknown-campaign totals in a callout. Use item_key `insight-meta-ads-reported-vs-real-roas` every week.
## Route B: the browser
### What you need
- Signed in to Meta with at least Analyst access on the ad account, to Google Ads with Read only access, and to Shopify with Reports permission or to Stripe with a role that can view payments.
If something is missing, say what and stop; do not invent numbers.
### Steps
Start at https://adsmanager.facebook.com/. The user is signed in already; never type passwords or codes, and ask if a sign-in page appears.
1. **Read Meta per campaign**: Pick the ad account, open the **Campaigns** tab and set the date to last Monday to Sunday. Under **Columns > Customize columns**, add **Purchases** and **Purchases conversion value**, tick **Compare attribution settings** and choose 7-day click and 1-day view, then note spend, purchases and value per campaign with the click and view split.
2. **Read Google Ads per campaign**: Open `https://ads.google.com/`, go to **Campaigns** for the same dates, and add the **Cost**, **Conversions (by conv. time)** and **Conv. value (by conv. time)** columns. Note each campaign's ID too (add the **Campaign ID** column) if your UTMs use IDs.
3. **Read orders by UTM campaign**: Open `https://admin.shopify.com/`, pick the store, then **Analytics > Reports** and the **Sales attributed to marketing** report for the same dates, grouped by UTM campaign. For Stripe, open **Payments**, filter to the week and read the `utm_campaign` metadata on each succeeded payment (your checkout must copy it onto the payment, or open the linked Checkout Session), or export the list with metadata columns.
4. **Match and compare**: Match each UTM campaign value to a campaign ID or name, lowercased. For every campaign, divide the matched revenue by spend for real ROAS, set it beside the platform figure, and list UTM values that match nothing and the orders with no UTM at all.
Note: Shopify's marketing report applies its own attribution model, so its campaign totals can differ a little from a per-order count through the API. Ads Manager shows conversions on the impression date unless you change the reporting setting, so a few purchases can shift between weeks. Over a dozen campaigns, the API route is the better one.
## The report
- **The worst gap**: the campaign whose reported revenue the store backs up least, with both ROAS figures.
- **Per campaign**: spend, platform purchases and value, matched orders and revenue, reported and real ROAS.
- **View-through share** of Meta's purchases.
- **Unmatched**: orders with an unknown campaign value, and orders with no UTM.
- One thing to do first, for example "Cut Prospecting LAL's budget by a third until real ROAS clears 2.5".
## Post it to Briefkin
When `BRIEFKIN_TOKEN` is set, or `~/.config/briefkin/env` exists (the installer from https://briefkin.dev/docs writes it), also send the result to the user's Briefkin board as one card. Keep the `item_key` the same on every run so the card updates in place, use a new `event_id` per run, and put today's real numbers in place of the example ones. If the request fails, say so once and carry on.
```bash
[ -n "${BRIEFKIN_TOKEN:-}" ] || . ~/.config/briefkin/env
curl -sS -X POST "${BRIEFKIN_URL:-https://briefkin.dev}/v1/events" \
-H "authorization: Bearer $BRIEFKIN_TOKEN" \
-H "content-type: application/json" \
-d @- <<'JSON'
{
"v": 1,
"event_id": "meta-ads-reported-vs-real-roas/<date of this run, e.g. 2026-10-16>",
"item_key": "meta-ads-reported-vs-real-roas",
"project": "Acme Store",
"source": "Reported vs real ROAS: Meta and Google Ads against Stripe or Shopify skill",
"kind": "update",
"title": "Prospecting LAL: real ROAS 1.9, Meta says 3.4",
"summary": "Meta reports **41 purchases** for Prospecting LAL last week; Shopify shows **23 orders** with its UTM. Real ROAS **1.9**, reported **3.4**. 10 of the 41 were 1-day view-throughs.",
"widget": [
{
"type": "Table",
"props": {
"columns": [
"Campaign",
"Spend",
"Platform says",
"Shopify shows",
"Reported",
"Real"
],
"align": [
"left",
"right",
"right",
"right",
"right",
"right"
],
"rows": [
[
"Meta: Prospecting LAL",
"$2,400",
"41 / $8,160",
"23 / $4,560",
"3.4x",
"1.9x"
],
[
"Google: PMax all products",
"$1,800",
"38 / $6,840",
"26 / $4,680",
"3.8x",
"2.6x"
],
[
"Meta: Retargeting 30d",
"$900",
"28 / $5,040",
"24 / $4,320",
"5.6x",
"4.8x"
],
[
"Google: Search brand",
"$600",
"52 / $9,360",
"49 / $8,820",
"15.6x",
"14.7x"
],
[
"All four",
"$5,700",
"159 / $29,400",
"122 / $22,380",
"5.2x",
"3.9x"
]
]
}
},
{
"type": "Callout",
"props": {
"text": "Prospecting LAL claims $3,600 more than the store can trace to it, the largest gap of the four. Not counted: 9 orders ($1,530) with a utm_campaign that matches no live campaign, and 64 orders with no UTM.",
"tone": "warn"
}
}
]
}
JSON
```
Widgets come from the Briefkin catalog (https://briefkin.dev/v1/catalog). Without a token, just give the user the report.
## Run it on a schedule
Suggested: Mondays at 9:00 (`0 9 * * 1`). From cron, with the agent signed in on that machine:
```bash
0 9 * * 1 claude -p "Run the meta-ads-reported-vs-real-roas skill"
# Codex: codex exec, Gemini CLI: gemini -p, Copilot CLI: copilot -p, OpenCode: opencode run
```
Or use the agent's own scheduler (Claude Code routines, Codex automations, Claude or ChatGPT scheduled tasks): https://briefkin.dev/skills/meta-ads-reported-vs-real-roas#schedule
Source: https://briefkin.dev/skills/meta-ads-reported-vs-real-roas
UTM tags on every ad: Meta URL parameters such as utm_source=facebook&utm_medium=paid&utm_campaign={{campaign.id}}, and a Google Ads final URL suffix such as utm_source=google&utm_medium=cpc&utm_campaign={campaignid}.
Read access to both ad accounts (ads_read on the Meta ad account, a Read only Google Ads user or the Google Ads MCP server).
Shopify Admin API read_orders, or a Stripe restricted key with Read on Checkout Sessions and Customers plus a checkout that copies the visitor's UTM into Checkout Session or customer metadata.
https://mcp.facebook.com/ads, with tools for reporting, ad and catalog management, datasets and troubleshooting. You sign in through Facebook Login for Business. For Claude Code, Meta documents claude mcp add --transport http --client-id <META_APP_ID> meta-ads https://mcp.facebook.com/ads, using a Meta developer app with the ads MCP use case added; clients that send headers can pass a user access token as a Bearer token instead (get started). Codex and Cursor take the same URL.pip install meta-ads, reads the token from ACCESS_TOKEN, and returns insights with meta ads insights get --fields spend,impressions,ctr,cpc, which suits scheduled runs from Claude Code or Codex.ads_read permission. The skill reads META_ACCESS_TOKEN and META_AD_ACCOUNT_ID (act_1234567890).The MCP server can create and edit ads when you grant management permissions. For unattended reports, a system user with only ads_read and View performance can't touch a bid, a budget or an ad.
google-ads-mcp, source). Run it locally with pipx run --spec git+https://github.com/googleads/google-ads-mcp.git google-ads-mcp, or deploy it to Cloud Run for remote clients. Its tools are list_accessible_customers, search for GAQL and get_resource_metadata (docs).gcloud auth application-default login with the adwords scope) or an OAuth client, and set GOOGLE_PROJECT_ID. Add GOOGLE_ADS_LOGIN_CUSTOMER_ID when you reach client accounts through a manager account, and write customer IDs without dashes.The adwords scope can't be narrowed, so sign in as a Google Ads user with Read only access; the MCP server has no write tools either way.
https://mcp.stripe.com. Add Stripe from Claude's connector directory, or run claude mcp add --transport http stripe https://mcp.stripe.com/, and sign in with OAuth; clients without OAuth can send a restricted key as a Bearer token.stripe login once, then stripe subscriptions list --status active --limit 100 or stripe invoices list --status open. The CLI defaults to test mode, so add --live for real data.rk_live_...) in the Dashboard under Developers > API keys, store it in STRIPE_API_KEY, and give it Read on the resources the report touches: Subscriptions, Invoices, Charges, Customers, PaymentIntents, Balance and Disputes.Never hand an agent the secret key (sk_live_...); a restricted key with Read where needed and None everywhere else can't refund, cancel or charge anything.
npx -y @shopify/dev-mcp@latest) covers docs and API schemas, not your store's orders; for store data, community MCP servers exist.read_orders, read_products and read_inventory, and store SHOPIFY_STORE_DOMAIN (acme.myshopify.com) and SHOPIFY_ACCESS_TOKEN. Requests go to https://acme.myshopify.com/admin/api/2026-10/graphql.json with the X-Shopify-Access-Token header. Orders older than 60 days also need read_all_orders.Request only read_ scopes when you install the app; Shopify enforces them per token, so the agent can't edit products, refund or cancel orders.
Meta's default attribution counts a purchase up to 7 days after a click and 1 day after someone merely saw the ad, and it models some conversions it cannot observe. A UTM only survives a click in the same browser, so view-throughs and cross-device purchases never show up in the store. The skill shows the 1d_view share separately so you can tell a tracking gap from view-through credit.
Orders tagged with the old name land in the "unknown campaign" row instead of being lost. Tagging with the campaign ID ({{campaign.id}} on Meta, {campaignid} on Google) avoids this, because IDs never change. If you tag with names, tell the agent the old and new names once and it maps them.
No. The skill runs with your own agent and gives you the report in the chat. With a Briefkin token set, each run also posts it to your board as a report card, so you can glance at it instead of asking again.
Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot CLI, OpenCode, Windsurf and Amp all read SKILL.md files. The install command puts it in ~/.claude/skills and ~/.agents/skills, where they look. ChatGPT, the Claude app and Grok take the prompt instead. Pick your agent under "Install and run it" and the commands change to match.
Yes. Switch the data source to "Your browser" and the steps change to the Meta Ads, Google Ads, Stripe and Shopify pages themselves: the agent opens adsmanager.facebook.com in a browser you are signed in to and reads the numbers off the page. Claude Code, Codex, Cursor, Claude app and ChatGPT can drive a browser out of the box; Gemini CLI, Copilot CLI and OpenCode need a Playwright or Chrome DevTools MCP server first. It is slower than the API and needs you signed in, so it suits a report you run by hand or from a scheduler on your own computer.
Skills are free to use with any agent. Briefkin is where their results land: one board for every report, number and question your agents send, with a ping when something needs you.
Free during the open beta. No card.