How to Use Whatagraph to Blend Google Ads and Meta Spend

The short answer: to show one client's Google Ads and Meta spend in a single widget, a Whatagraph blend costs no extra credits, since blends sit outside the connected-sources count and the two accounts already use 2 of the 50 credits on Max, from €699 a month billed annually (source counting article, pricing page, checked on 26 October 2026). The catch is detail: Whatagraph's own help centre shows a Meta and Google blend tripling spend when Meta's placements stay in the join. The steps: connect both accounts in the right Space, choose a blend over a source group, build it on Date with a full outer join, match the dimensions on both sides, check the widget against the platforms, let an IQ Agent draft the commentary, and automate the report with review switched on.

Disclosure: I may earn a commission through some links on this site. It does not change what this page says.

Write these down before you blend:

1. Each platform's total for last month. For example, Google Ads €8,420 and Meta €6,135 give €8,420 + €6,135 = €14,555. The blended spend widget for that month has to show €14,555 and nothing else.
2. The dimension both platforms already share. Date is the safe join key. Campaign names only work if your naming already matches across Google and Meta.
3. The first three sends. Keep review on for the first three automated reports and compare the blended total with your sum each time.

What Whatagraph includes and costs is in the Whatagraph glossary entry.

What you need

• A Whatagraph account. New accounts start on a 14-day Max trial, which includes blends and source groups (upgrade article).
• An admin or manager role for step 6, since editors cannot use IQ Agents (IQ Agents article).
• Access to the client's Google Ads and Meta Ads accounts.

Step 1: Connect both accounts in the right Space

In a report, open the Sources drawer, click Manage all sources, then Connect new source, and add Google Ads and Meta Ads (adding sources article). If a connected source does not appear in the report, the Space is the usual cause: sources belong to Spaces, so move the report or reassign the source. Switching a report from sample data can reset widgets that used a metric the new source lacks, so look over each widget afterwards.

Step 2: Choose a blend, not a source group

A source group adds up the same metrics across several accounts, such as all of one client's Google Ads accounts, without formulas (source groups article). A blend joins different channels on a shared dimension (blends article). One spend line across Google and Meta is a blend. If the client also runs several Google Ads accounts, a source group handles that part.

Step 3: Build the blend on Date

Under Organize, Aggregations, click Create New, then Create a new blend (blends article). Select the two sources, pick the dimensions and metrics, choose Full outer join and set Date as the join key. A full outer join keeps every date from both sides and fills gaps with null, and Whatagraph recommends it for most cases; cross join is for special calculations only (join types article). Name the blend after the client and the metric.

Step 4: Give both sides the same level of detail

This step decides whether the number is right. If Meta brings spend by campaign, date and placement while Google brings campaign and date, every Google row repeats once per placement and the total inflates (inflated numbers article). Whatagraph calls this expected join behaviour, not a bug. Edit the blend so both sides carry the same dimensions, removing Placement from Meta or adding the matching dimension to Google. For a single widget, the Aggregate By setting does the same job without touching the blend.

Step 5: Build the widget and test it

Add a widget on the blend showing spend by month and compare it with the total from your list. Higher than your sum means a dimension still differs; go back to step 4. Export the widget to CSV if you want the comparison in a spreadsheet (export article).

Step 6: Let an IQ Agent draft the commentary

Open IQ Agents and give the job to Insight, the pre-made agent that answers data questions and drafts client-ready summaries into a text widget (pre-made agents article). For example: "Summarise last month's blended Google Ads and Meta spend for [client], with the change against the month before, in three sentences." Before an agent changes a report, widget or blend it shows an impact warning and asks you to confirm, and approvals gather in Activity, where only the person who started the run can approve (Activity article). The account manager reads the draft against the widget before approving it.

Step 7: Automate the report with review on

Pick a pre-made or rolling date range, click Automation, and set the frequency, delivery day, time and time zone (automation article). Choose "I'll review before sending", and under issue handling "Do not send reports until issues are reviewed". Add the recipients, attach a PDF or Excel file if the client wants one, and click Automate report.

Warnings from Whatagraph's help centre

• A report set to review is not sent until someone confirms it. Name the reviewer before the first cycle (automation article).
• YouTube blends are limited to the date dimension by its API (blends article).
• Amazon Ads data cannot be exported to Excel or CSV (export article).
• A reviewed report can go back. If you forgot a comment, Revert to Review in the report history reopens it (automation article).

Do the sum by hand, then let the trial decide

The free check needs no Whatagraph at all: last month's spend from each ad manager in two cells of a spreadsheet, added up. Every blended figure you see later has to match that cell, and keeping it during the trial costs nothing. Mind how the trial closes. New accounts run on Max for 14 days and then book a call to choose a plan, with everything built during the trial left in place (upgrade article), so the price conversation starts after two weeks of your blends and templates already living inside the product.

The summed spend is a fact. The revenue each platform reports beside it is that platform's claim, and blending two claims does not turn them into one answer. Later, if you want each channel's contribution measured against what would have happened without it, a causal attribution tool like Causality Engine does that, and I founded it.

FAQ

Which join type should I use?

Full outer join in most cases; left outer when one source is your main dataset; inner when only the overlap matters; cross join for a ratio with no shared field (join types article).

Can I fix an inflated widget without editing the blend?

Yes. Open the widget's settings, find Aggregate By and choose the dimension that matches the less detailed source (inflated numbers article).

Can Claude build the blend instead?

Through the Whatagraph MCP, Claude can combine channels with a blend or a source group, asking before it creates or changes anything; build tools depend on your plan (MCP article).