How to Read a Marketing Agency Report: Six Numbers and What They Prove

A marketing agency report is the monthly document that says what your retainer bought. Reading one well means separating the numbers that prove booked work from the numbers that only prove activity, then checking whether each figure carries a denominator, a baseline and a date. Most reports fail on the baseline.

Every platform fact below comes from the platform's own documentation or from a dated study, read as of September 2026. The reason so many report columns mislead is rarely dishonesty. It is that the number was never capable of meaning what its label implies, and the vendor's own documentation often says so.

Ask what the denominator is before you read the number

Every number in a report is a fraction with the bottom half missing. Forty leads means nothing until you know forty out of how many calls, over how many days, from which channel, and against what the same month produced a year ago. Ask for the denominator and half the report answers itself.

The six numbers most reports lead with, and what each can prove

Six figures lead most monthly reports: impressions, keyword rank, an AI visibility score, clicks to call, leads, and a list of tasks completed. Each proves something real and none proves what its row label suggests. The table below pairs each with the question that turns it into a number about your business.

None of the six is a bad number. Impressions are the right diagnostic when calls fall and you need to know whether the cause is demand or visibility. The failure is using a diagnostic as a scoreboard, which is what happens when a report leads with the column that always rises.

A rank column is an average of answers that differ by where the searcher stands

A rank column reports where one tracking bot saw your listing from one fixed point. Google's Business Profile help page says local results are mainly based on relevance, distance and popularity, and distance is measured from the searcher. A customer in Ballantyne and a customer in NoDa get different answers on the same morning.

Google's wording on that page, as of September 2026: "Local results are mainly based on relevance, distance, and popularity." The same page states there is no way to request or pay for a better local ranking. A single rank cell summarizes many different answers, and a promised position promises something the seller does not control.

The ranking-factor lists reports cite deserve the same reading. Whitespark's 2026 Local Search Ranking Factors report, published November 2025, surveys 47 local search experts scoring 187 factors, and says so of itself: "None of the experts have special access to the internal workings of Google's local search algorithm." An honest document, and not a readout of the algorithm.

Search Console does not publish AI Overview traffic as its own line

Search Console folds AI Overviews and AI Mode into ordinary search traffic. Google's AI features documentation, last updated December 2025, states that sites appearing in AI features are included in the overall search traffic reported in the Performance report, within the Web search type. A row labeled AI Overview clicks needs to say where its figure came from.

Google is blunter still in its AI optimization guide, last updated July 2026: "No third-party tool has access to our internal ranking or AI systems." Read an AI visibility score the way you would read a poll. Useful for trend, honest about method, never a measurement of the thing itself.

When the baseline moved and nobody wrote it down

A flat number can mean the page it was measured on changed. Sterling Sky's own rank tracking, published June 2026, found local pack ads on about 1% of its mobile reports at the start of 2025 and almost 22% by December, with Local Services Ads rising from roughly 11% of tracked queries to 31%.

Those are Sterling Sky's figures from its own client set, not an industry census, and the share depends on the keyword mix tracked. The point holds anyway. Where paid units now sit above a map pack that carried none eighteen months earlier, an unchanged position produces fewer calls, and a report showing the position without that context has explained nothing.

Leads is the wrong last column

Leads is the wrong last column because a lead is a contact, not a job. ServiceTitan's platform data across more than 3,000 US and Canadian trade businesses, as of June 2022, put a typical shop's call booking rate at 42%: plumbing 43%, electrical 41%, HVAC 38%, and shops with fewer than five technicians at 24%.

Chris Hunter, Principal Industry Advisor at ServiceTitan, said of that finding: "the impact and the importance of capturing this thing on the front end can lead to huge results on the back end." The front end is the phone, the form and the booking link. A report that stops at leads measures the half of the funnel that was never the problem.

So the last column is booked work, and the honest version of it names its own limits. Not every booking can be attributed to a channel, some customers call after seeing four things, and a report that claims perfect attribution is claiming more than the tracking can carry. Ranges and stated assumptions beat false precision.

What the report looks like when the systems underneath it are built

A report reads in booked work when the booking runs on the same system the report reads from. Mirastart builds the booking, missed-call text-back, follow-up and review systems first, then runs the marketing on top of them, so calls, booked jobs and average ticket by channel come out of one record rather than four tools stitched together monthly.

The booking system behind this site's own discovery-call page calculates real availability and confirms automatically; the online booking and repair-status software Quick Auto NC runs its bays on is ours. Each writes a record, and a record is what a report is built from. So the test for any agency, this one included: ask where the numbers come from, and whether you can open the same screen yourself on any day of the month.