Marketing Reports Explained: Metrics, Data Sources and What They Measure

What this covers

• 1. Impressions Leading the Report

• 2. Branded Search Counted as a Win

• 3. a Comparison Window Chosen After the Fact

• 4. One Ranking Number for a Local Business

• 5. Activity Reported as Outcome

• 6. Traffic Reported Without Source

• 7. the AI Metric That Does Not Exist

• What a Usable Report Looks Like

• The Numbers That Look Like Progress and Are Not

• What a Store’s Report Should Contain Instead

• The Test for Any Report

A monthly marketing report is one of the few documents a business owner receives that is written by the person being evaluated by it. That is not an accusation, it is a structural fact, and it produces predictable habits.

Here are seven, with the question that exposes each one.

1. Impressions Leading the Report

Impressions count how many times you were eligible to be shown. Not seen, not clicked. Eligible.

It is the easiest number in digital marketing to grow, because a site can start appearing at position 40 for hundreds of new searches and generate an impressive-looking climb while producing no visits at all. A report that opens on impressions has chosen the metric that moves most and matters least.

The question: how many of those impressions turned into clicks, and how has that ratio changed over three months?

2. Branded Search Counted as a Win

When someone searches for your business by name, they already know you exist. That search is a result of your other marketing, not a product of search optimization.

Branded terms tend to rank first and convert well, so including them inflates every number in the report. A company running radio ads will see branded search rise and can be sold that rise as an SEO result.

The question: show me the same numbers with every search containing my business name removed.

The gap is often dramatic. Worth noting: brand-name misspellings count as branded too, and a simple filter on the exact spelling misses them.

3. a Comparison Window Chosen After the Fact

Any month can be made to look good by choosing what it is compared against. Against last month. Against the same month last year. Against the three-month average. Against the launch date.

When the comparison window changes between reports, somebody is picking the flattering one.

The question: use the same comparison basis every month and tell me in advance which one it is.

4. One Ranking Number for a Local Business

Local search results vary with where the searcher is standing. Distance is one of the three inputs Google names for local results, so a position checked from one location describes one location.

A rank checked from the agency’s office, or from your own address, is flattering by construction. You are standing on top of the business.

The question: show me rankings measured across a grid of points in my service area, with the same grid each month.

A business is routinely first at its own address and outside the top ten five miles away, and only the grid shows it.

5. Activity Reported as Outcome

“Published 4 blog posts. Built 12 citations. Optimized 6 pages.”

All of that may have happened and none of it is a result. Activity reporting is the most common pattern and the most understandable one, because in the early months of an engagement activity is genuinely most of what there is. The problem is when it never graduates.

The question: for each item on that list, what changed as a result, and if nothing has yet, when should I expect it to?

A good agency answers this straightforwardly, including “too early to tell, check in month four.”

6. Traffic Reported Without Source

Total sessions is a number with many parents. An email campaign, a social post, a press mention, an ad, or genuine organic growth all move it.

The question: break that into organic, paid, direct, referral and social, and show me only the one you are responsible for.

Reports that do not segment traffic are frequently taking credit for other channels, sometimes without meaning to.

7. the AI Metric That Does Not Exist

This is the newest one and it is spreading quickly.

A growing share of searches get answered by an AI summary at the top of the results page, or by an assistant that returns a reply with no list at all. Businesses want to know whether they are being named inside those answers, and agencies have started reporting on it.

The problem is that Search Console does not report it. There is no impressions figure for a mention inside an AI-generated answer and frequently no click at all. No standard analytics platform reports it either.

So when a report contains a number labeled “AI visibility” or “AI impressions,” that number came from somewhere else and got relabeled, or it was estimated.

The question: where does this number come from, and can you show me the raw source?

The honest method available right now is manual: run the questions your customers actually ask against the live systems on a schedule, record which businesses were named and which sources were cited, date it, and compare month to month. That produces a real answer and a slow one. It also has to come with a caveat, which is that these results are genuinely volatile. We have run identical queries an hour apart and had an AI answer appear once and not the second time, so a single run is a snapshot rather than a verdict.

An agency reporting one good bounce as a win it caused is either fooling you or fooling itself.

What a Usable Report Looks Like

Short. Four parts.

What was done this month, in plain nouns. What changed, measured the same way as last month. What that means, in a sentence a person can act on. What happens next month.

If a month was poor, it says so and says why. A report where every month is a win is not a report, it is marketing about marketing.

Reporting on a store is the hardest version of this, and e-commerce SEO in Kansas City sets out which numbers are worth showing and which only look like progress. The business profile shows the market it is written from, and there is more on the same work for a local store. Ten years of doing this locally has mostly taught us that the reports clients actually read are the short ones.

The Numbers That Look Like Progress and Are Not

Reported metric	Why it looks good	What it hides
Total impressions	Large and always rising	Impressions are not visits
Keywords ranking	Grows automatically over time	Most are irrelevant or position 80
Average position	Single clean number	Averages hide the pages that matter
Traffic, site-wide	Easy to grow with any content	Which pages actually earn
Social reach	Impressive in isolation	Rarely connects to revenue

Average position is the most misleading of the five, because it improves when a page falls out of the index entirely. A number that gets better when something breaks is not a measurement.

What a Store’s Report Should Contain Instead

Metric	Why it is honest
Revenue by landing page	Ties the work to money
Category page performance, separately	Categories do the ranking, so track them alone
Conversion rate by traffic source	Separates a traffic problem from a site problem
Products with impressions and no clicks	A specific, fixable list
Indexed product count against total	Catches quiet exclusion at scale

The last row is the one that catches serious problems in stores. Pages get excluded in bulk through a setting nobody revisits, and a site-wide traffic number can look stable while a third of the catalog has quietly left the index.

The Test for Any Report

Ask what would have to happen for next month’s report to look worse. A report that cannot answer that is measuring things that only go up, which means it is not measuring performance at all.

Every honest reporting structure has a way of showing a bad month. If you have never seen one, you are not being shown the whole picture.

The Test That Takes Two Minutes

Open your last three monthly reports side by side.

If you cannot tell which month was the best one without reading the commentary, the numbers in them are not measuring anything. That is the whole test, and it works on any report from any agency in any channel.

A final habit that costs nothing. Keep the first report you were sent and compare it with the most recent one. If the metrics changed between them, ask why. Reporting that quietly swaps which numbers it features is the most common way a declining engagement keeps looking healthy, and the substitution is almost never announced.

The reverse is also worth noticing. A supplier who adds a metric that makes their work look harder, or who starts reporting something they previously left out, is usually being straight with you. That is rarer than it should be and it is worth saying so when it happens.

The shortest version of all of this: a report should be able to make you uncomfortable. If it never has, it is not reporting on the work, it is reporting on the relationship.

That is a low bar and a surprising number of monthly documents fail it. Asking for one number that went the wrong way is usually enough to find out which kind you are receiving.