How to run a weekly marketing attribution review in 30 minutes
A weekly marketing attribution review is a 30-minute meeting where you check spend against new meetings by channel, explain what moved, flag traffic with no clear source, and leave with one change to make. It works when it stays short and ends in a decision. It fails when it turns into a tour of every dashboard you own.
Most B2B teams already have the data for this. What they lack is a fixed agenda and a rule for which numbers to read each week and which to leave alone until the quarter closes. This guide gives you both.
What should a weekly attribution review answer?
Keep it to three questions. Are we spending what we planned, where we planned? Which channels created new meetings or opportunities this week? Is anything changing fast enough that it should change what we do next week?
Notice what is missing. A weekly review is not the place to judge which channel "works." B2B deals take months, so the deals that closed this week were mostly started by last quarter's spend. If you grade channels on this week's closed-won, you will cut programs right before they pay off. Our post on how long each B2B channel takes to create deals explains the lag in more detail.
What does the 30-minute agenda look like?
Use the same five blocks every week, in the same order, so the meeting runs itself.
- Spend and pacing (5 minutes). Pull spend by channel from Google Ads, LinkedIn Ads, and anything else you run. Compare it to plan. Overspend on one campaign is the most common thing to catch here.
- New meetings by channel (5 minutes). Count booked meetings and new opportunities, grouped by the channels on each account's path, not only the last click. This is your earliest sign of pipeline.
- What moved, and why (5 minutes). Look for any channel that rose or fell sharply against its recent average. Then ask for a reason: a new campaign, a paused ad set, a broken form, a holiday.
- Unclear sources (5 minutes). Check how much traffic and how many meetings landed in Direct or "unknown," and whether that share is growing.
- Pick one change (10 minutes). Agree on one adjustment, name an owner, and write down what you expect to see if it works.
Ten minutes for the decision sounds like a lot. It is the part teams skip, and it is the reason the meeting exists.
Which numbers belong in a weekly review, and which can wait?
Every number has a natural pace. Read it faster than that and you react to noise. Read it slower and you miss problems you could have fixed.
- Weekly: spend pacing, new meetings by channel, and new referring sources you have not seen before.
- Monthly: pipeline created by channel and the trend in customer acquisition cost (CAC).
- Quarterly: closed-won revenue by channel and any budget shifts that follow from it.
CAC is a good example. One week of spend divided by one week of new customers tells you almost nothing in B2B, because the customers who signed this week were sourced months ago. Our guide to reading CAC by channel when deals take months shows how to match spend to the cohort it actually started.
The same logic applies to the length of your lookback. If your attribution window is shorter than your sales cycle, early touches drop off the path before the deal closes. Check that your B2B attribution window covers your typical cycle before you trust the monthly and quarterly views.
How do you handle traffic with no clear source?
Every week some meetings will show up with a thin path or none at all. Two causes account for most of it.
The first is AI search. Buyers who find you through ChatGPT, Perplexity, or other assistants often arrive with no UTM, so the visit lands in Direct. Our explainer on LLM attribution and ChatGPT traffic covers how to recover those visits.
The second is dark social: podcasts, Slack groups, and word of mouth that never produce a click. A "How did you hear about us?" field on your demo form fills much of that gap. The self-reported attribution checklist walks through how to set it up and read it next to your tracked data.
In the weekly review, you do not need to solve either problem. You need to watch the share. If unclear sources jump from one week to the next, that is usually a tracking break, such as a form change or a stripped UTM, and it is cheaper to fix in week one than in month three.
Who should be in the room?
Keep it small: whoever owns the marketing budget, whoever runs paid channels, and someone from sales or RevOps who can speak to meeting quality. Sales matters here because a channel that books plenty of meetings that never turn into opportunities is not a good channel. Our post on how sales and marketing split deal credit covers how to keep that conversation from turning into a fight over numbers.
If your CRM is HubSpot, bring the meeting and opportunity data straight from it rather than a spreadsheet export someone built on Friday. Stale exports are the fastest way to lose the room.
What should come out of the meeting?
One written change, one owner, and one expected result. For example: "Shift a portion of the retargeting budget into the webinar campaign this week. Owner: paid lead. We expect more booked meetings from webinar attendees within two weeks." Next week, the first thing you read is whether that happened.
Keep a running log of these decisions. After a quarter, the log becomes the story you tell leadership, and it pairs well with a one-slide marketing attribution summary for the board.
Running the weekly review with Source
Source does the joining for you. Install one pixel or a Google Tag Manager tag, connect your ad accounts and CRM, and Source builds each account's path from first visit to closed deal, with no warehouse to set up. Multi-touch, view-through, offline, and LLM attribution all sit on that same path, along with real CAC, goals, and channel quality.
That covers most of the agenda before the meeting starts. Daily Slack reports show spend and new meetings by channel as they happen, AI recommendations flag what moved, and you can chat with your data to ask why. If your team uses an AI agent, it can pull the same answers through the Source MCP server, with the SQL behind each one. Source is SOC 2 Type II and ISO 27001 certified, and customer data is not used to train external models.
If you want your next weekly review to start from one shared set of numbers, start free at dash.source.app/signup and install the pixel.