A quarterly review should answer five questions, each with one number attached: what did we spend, what did it produce in pipeline, what closed from it, what are we killing, and what changes next quarter. Anything else is a status update. Ask the agency to send those five numbers 48 hours before the call, not during it.
By Avi Vatsa, CEO, Exchange Four Agency. Avi leads the team that installs and runs AI-leveraged revenue systems for owner-led companies. His background spans law, technology, and marketing; he also co-founded Dialora, an AI voice-agent platform for automated lead capture and booking. (Marketer of the Day #1411, Jeremy Ryan Slate Show)
Why does accountability slip after the first 90 days?
Because onboarding has a deadline and quarter two doesn't. The first 90 days come with a plan, a kickoff, and a visible finish line, so everyone shows up prepared. Then the cadence softens into a monthly deck, the deck becomes a dashboard link, and the dashboard fills with impressions.
An industry survey distributed via Businesswire found 71% of brands cite frustration demonstrating marketing ROI effectiveness. That is not 71% of brands with no data. Most of them have more data than they can read. The gap is between the data they have and a number the owner can take to the bank, and quarterly reviews are where that gap gets either closed or permanently papered over.
The structural problem: the agency sets the agenda. When the agency sets the agenda, the agenda covers what went well. You set the agenda, and the shape of the meeting changes before anyone speaks.
What are the five questions to ask in a quarterly review with a marketing agency?
Send these in advance. Ask for one number per question, pre-filled, before the call starts.
- What did we spend this quarter, all-in? Retainer plus ad spend plus tools plus anything billed as a pass-through. One figure.
- How many qualified opportunities did that produce, and by what definition of qualified? The definition matters more than the count. Lock it in writing this quarter so it can't float next quarter.
- What closed revenue is attributable to those opportunities, and what's still open in pipeline? Two numbers, both traced to named deals, not modeled percentages.
- What are we stopping? Every channel that didn't produce should have a kill date or a stated reason it gets another quarter.
- What changes next quarter, and what number does that change move? A forecast you can hold them to in 90 days.
Five questions. Twenty minutes if the agency is prepared, two hours if they aren't — and the length of the answer to question three tells you almost everything.
How do you run the meeting so it doesn't turn into a presentation?
Ask for the numbers 48 hours early and read them alone. That single move changes the meeting more than any question on the list, because it removes the agency's ability to narrate the data while you see it for the first time. You arrive with your three real questions already formed.
Then set the order: start with question three — closed revenue — and work backwards. Most reviews start at the top of the funnel because that is where the numbers look best, and by the time anyone reaches revenue there are nine minutes left and someone has a hard stop. Start at the bottom. Impressions can wait.
Keep a running document, not a fresh deck each quarter. The document carries forward last quarter's forecast next to this quarter's actual, which is the only view that makes a pattern visible. One missed forecast is noise. Three is a diagnosis.
What should be on the agenda, and what should be left off?
| On the agenda | Off the agenda |
|---|---|
| Closed revenue traced to named deals | Impressions, reach, follower growth |
| Cost per qualified opportunity, trended across four quarters | Cost per click with no downstream number |
| Pipeline created vs. pipeline forecast last quarter | "Engagement is up" with no dollar attached |
| What we're killing, and the date | A list of everything produced this quarter |
| Who actually worked the account, by name and hours | A slide of logos and awards |
| Next quarter's forecast, written down | A new strategy with no reference to the last one |
The right column isn't worthless. It's just not quarterly-review material — it's weekly operating detail, and letting it occupy the quarterly slot is how a year passes without anyone asking whether the engagement paid for itself. If you aren't sure which metrics belong in the left column for your business, we worked through the full list in what metrics a marketing agency should report every month.
What does a good answer to "what closed?" actually look like?
It looks like a list of deal names, each with a source, a close date, and a value. It looks like someone saying "these four came from the paid search campaign, this one we can't trace, and these two were referrals the campaign didn't touch — so I'm not claiming them."
A weak answer sounds like a methodology. "We use a multi-touch model that attributes a weighted share across the customer journey." That might be perfectly legitimate. Ask the follow-up anyway: show me one deal and walk me through where the credit went. If nobody can do that for a single named customer, the model isn't measuring your business, it's decorating it.
The honest version of this, including why modeled attribution and traced attribution diverge, is laid out in how to calculate marketing ROI.
What if the numbers are bad?
Then the review is working. A quarterly review that only ever produces good news isn't a review, it's a retention mechanism.
The question that separates a partner from a vendor is what the agency does in the ten seconds after you say the number is down. Watch for one of three responses. The first explains the number away — seasonality, the algorithm, a competitor's budget. The second accepts the number and proposes activity: more content, more tests, more posts. The third accepts the number, names a specific cause, and proposes a change with a forecast attached.
Only the third is accountability. At Exchange Four we hold a plain internal rule on this: run on the numbers, never explain away a falling number. It is easy to say and uncomfortable to practice, which is exactly why it's worth watching for in someone you pay.
As Riggs Eckleberry, Chairman of OriginClear, put it about working with us: "They know exactly how to connect marketing execution to real business outcomes." That connection is the entire job. Everything else is how.
How does this fit with the rest of the operating rhythm?
The quarterly review is the accountability layer, not the management layer. If the only structured contact you have is four meetings a year, the quarterly becomes a tribunal and nobody performs well under one. A weekly or biweekly working session handles execution; the quarterly handles whether the execution is paying.
That split is the core of the operating rhythm most owners never get told to set up. It also connects backwards: the baselines you agreed in the first 90 days are what quarter two is measured against. If no baseline was set at onboarding, set one at this review, and accept that the first quarterly is a calibration rather than a verdict.
One practical note on the fragmented setups. Owners running an SEO shop, a PPC shop and a design shop separately end up running three quarterly reviews where no one owns the revenue number — each vendor reports their own slice, and the gaps between slices belong to nobody. Vendor-management research puts the coordination overhead alone at 8–15% of annual vendor spend in hidden labor that never appears on an invoice, with businesses reporting roughly 30% higher total spend versus a single integrated partner. The quarterly review is usually where that tax becomes visible, because it's the one meeting where someone has to add it all up. We broke down the full arithmetic in what multiple marketing vendors actually cost.
When does a bad quarterly review mean it's time to leave?
One bad quarter is data. The pattern worth acting on is three consecutive quarters where the forecast missed, the cause was never named, and the proposed fix was more activity. Add a fourth signal: the person answering your questions has changed twice and neither one could reconstruct what was decided last quarter.
At that point you aren't managing an underperforming engagement, you're subsidising one. The signals, and what to check before you act on them, are covered in when to fire your marketing agency — and if you decide to move, switching without losing a quarter covers the handover sequence.
The one-page agenda to send before your next review
Copy this into the calendar invite:
- Pre-read, due 48 hours before: all-in spend, qualified opportunities with the written definition, closed revenue by named deal, open pipeline, last quarter's forecast vs. actual.
- Minute 0–15: closed revenue and pipeline. Walk one deal end to end.
- Minute 15–25: cost per qualified opportunity, trended across every quarter to date.
- Minute 25–35: what we're stopping, with dates.
- Minute 35–50: next quarter's forecast, written into the running document.
- Minute 50–60: staffing — who works this account, how many hours, any changes coming.
Sixty minutes, four times a year. That is the entire accountability structure most owner-led companies never build, and the reason an agency relationship drifts for eighteen months before anyone checks whether it paid.
If you want the version where the system reporting the numbers is the same system producing them, that's what we install and run — the offer, the funnel, the follow-up, the CRM, and the tracking that reports back in revenue rather than impressions.