Your Marketing Reports Are Lying: What to Track Instead

In November 2025, one of our clients had the best traffic month in their history.

At least, that’s what Google Analytics said.

891 sessions.

Nearly double the normal baseline.

Big spike.

Up and to the right.

The kind of chart an agency can drop into a monthly report, circle in green, and spend ten minutes congratulating itself over.

There was just one problem.

62.4% of it was bots.

We know because we checked.

And then we did the thing that’s apparently radical in this industry:

We told the client.

No “record-breaking traffic” slide.

No victory lap.

No taking credit for hundreds of automated visits from servers halfway around the world.

We filtered the garbage out.

And when we did?

The real numbers were smaller.

They were also actually useful.

Here’s what happened — and what business owners should be looking for in their marketing reports instead.

How Do Bots End Up in My GA4 Reports?

Starting in September 2025, GA4 properties were hit by a documented wave of automated direct traffic originating from China.

It wasn’t some sophisticated attack against our client.

It wasn’t a competitor.

Nobody in a dark hoodie was sitting in a basement furiously refreshing a garage door website in Fort Wayne.

It was noise.

Automated traffic getting counted alongside real humans.

For our client, Fort Wayne Door & Dock, the wave ran from October 2025 through January 2026.

Here’s what the numbers looked like:

• October: 67.6% of reported sessions were unengaged bot traffic — the worst month by percentage
• November: 891 total sessions reported; 556 were unengaged — 62.4% fake
• December: 960 sessions reported; 52.7% were unengaged — real engaged traffic: 454
• February 2026: the wave cleared and the numbers normalized

Here’s the uncomfortable part.

If we sold total sessions, that stretch was a gift.

“Traffic nearly doubled!”

Boom.

Invoice attached.

And most business owners would never question it.

Why would you?

The chart looked exactly like what everyone hopes a marketing chart will look like.

But the traffic wasn’t real.

So the growth wasn’t real either.

We filtered the bots and reported what actually happened.

And the honest story turned out to be better than the fake one.

Bot-adjusted engaged sessions grew 21% from the pre-spike baseline to spring 2026.

Organic search sessions grew 16% year over year.

That’s real growth.

Defensible growth.

Growth you could actually use to make a business decision.

The inflated version was bigger.

And worth absolutely nothing.

Why Do Agency Reports Always Look Good?

Because Most Reports Are Built From Metrics That Are Easy to Make Look Good.

This doesn’t always require dishonesty.

That’s what makes the problem interesting.

Sometimes it’s just selection.

The vendor gets to decide which numbers go into the monthly report.

So naturally, the report tends to feature numbers that:

1. Usually go up

2. Are easy to influence

3. Are difficult to connect directly to revenue

Impressions.

Total sessions.

Reach.

Engagement.

Followers.

Keyword movement.

Clicks.

All perfectly legitimate metrics.

And all capable of doubling while your schedule stays empty.

That’s the trap.

A metric can be technically accurate and still tell you almost nothing about whether your marketing is working.

During the bot wave, GA4 really did record those sessions.

The number wasn’t fabricated.

It was just useless without context.

That distinction matters.

Here’s the Part Nobody Says Out Loud

The Person Grading the Homework Is Usually the Person Who Did the Homework.

Think about the normal agency relationship.

The agency runs the campaign.

The agency picks the KPIs.

The agency builds the dashboard.

The agency interprets the numbers.

Then the agency tells you whether the agency did a good job.

That’s a strange system when you say it out loud.

Again, you don’t need a cartoon villain here.

No agency owner needs to sit in a dark room saying:

“How can we trick Steve into thinking Facebook is working?”

The incentive already exists.

If impressions are up 60% and booked jobs are flat, which number is easier to put in 48-point type on slide three?

Human beings like numbers that make their work look successful.

Marketing people are still human beings.

Usually.

So good measurement needs a harder standard.

The report shouldn’t exist to prove the marketing team worked hard.

It should help the business understand what happened.

What Marketing Metrics Actually Matter?

Start With the Numbers That Get Harder to Bullshit.

Here’s the measurement framework we use.

Not every business will have every piece immediately.

But this is the direction the reporting should move.

Engaged Sessions — Not Just Total Sessions

We still track total sessions.

We just don’t let them drive conclusions by themselves.

An engaged session gives you a better signal that an actual human did something meaningful.

Bots almost never qualify.

That one change filters a lot of noise by definition.

It was also the first reporting adjustment we made for Fort Wayne Door & Dock during the bot spike.

Total traffic stayed visible.

It simply stopped being treated like a victory condition.

That’s an important difference.

Google Business Profile Actions

For a local service business, a customer does not always need to visit the website.

They can find you in Google.

Read reviews.

Look at photos.

Check your hours.

Tap the phone number.

Call.

Done.

That customer may never appear as a website visitor.

But they are very real.

Which is why we track things like:

• Phone calls
• Direction requests
• Website taps
• Other meaningful Google Business Profile actions

In the AI-search era, this matters even more.

Customer demand can move through search without producing the same website session we used to expect.

If your report only watches the website, part of the customer journey becomes invisible.

Confirmed Calls

Not:

“Estimated call volume.”

Not:

“Phone-number clicks.”

Not:

“Possible conversions.”

Actual calls.

For Fort Wayne Door & Dock, we cross-check organic calls through two independent systems:

• Google Business Profile call tracking
• CRM organic attribution

That’s how we can say the campaign produced 367 confirmed organic phone calls across five months and defend the number.

Every one of them.

That is much more useful than:

“Organic visibility increased.”

Great.

Did anybody call?

Qualified Leads

A lead is not automatically valuable because a form was submitted.

Your cousin asking for a job through the contact form is technically a conversion.

So is spam.

So is somebody 400 miles outside your service area.

So is a vendor trying to sell you office supplies.

The business does not need more “conversions.”

It needs more qualified opportunities.

Reporting should eventually distinguish between:

A lead.

A real lead.

A qualified lead.

Those are different things.

Treating them as the same number is how marketing reports get very impressive while sales gets increasingly pissed off.

Booked Jobs

Now we’re getting somewhere.

How many leads turned into actual work?

Not clicks.

Not inquiries.

Not phone calls.

Jobs on the schedule.

For a local service business, this is where marketing starts getting much harder to hide behind.

If a channel generates 100 leads and books five jobs while another generates 40 leads and books 18?

Which one do you want more of?

The lead report says the first one won.

The business says otherwise.

Cost Per Booked Job

This is the final boss of honest marketing metrics.

How much money went in?

How many actual jobs came out?

This number is harder to track.

You need:

Call tracking.

A functioning CRM.

Clean attribution.

Marketing and sales talking to each other.

People actually updating the system.

Which is exactly why a lot of marketing reporting stops several steps before it.

The closer you get to revenue, the more departments have to cooperate.

But that’s also where the numbers become worth making decisions from.

Revenue

Eventually, marketing has to connect to money.

Not every sale can be perfectly attributed.

Customer journeys are messy.

Someone may:

See an ad.

Search your name later.

Read three reviews.

Visit the website.

Ask a neighbor about you.

Call from Google Maps.

Then hire you two weeks later.

Attribution will never be perfect.

That doesn’t mean we stop trying.

It means we use the best available evidence and get as close to revenue as the system allows.

Because your bank does not accept impressions.

I checked.

The Vanity Metric Test

Ask One Question: “What Would I Do Differently Because This Number Changed?”

Impressions are up 40%.

Okay.

What decision does that change?

Website traffic is up 35%.

Great.

Did the right people visit?

Did they contact you?

Did anything sell?

Followers increased by 2,000.

Cool.

Did those people buy anything?

A useful metric should help you make a decision.

Spend more.

Spend less.

Fix the offer.

Change the targeting.

Improve follow-up.

Hire.

Stop hiring.

Expand.

Cut something.

Do more of what works.

If the metric cannot reasonably influence a business decision, it probably doesn’t belong at the top of the report.

That doesn’t make it useless.

It makes it secondary.

Big difference.

A Green Arrow Does Not Mean “Good”

Marketing dashboards love green.

Green means up.

And up means good.

Apparently.

Except:

Cost per lead up 42%.

Green arrow?

Bad.

Traffic down 15%, booked jobs up 30%.

Red arrow on traffic?

Who cares?

Impressions down, calls up.

Which one wins?

Calls.

Every time.

The direction of the number does not tell you whether the number is good.

The business context does.

That sounds obvious.

And yet entire monthly reporting systems are built around turning metrics green.

What Your Marketing Report Should Actually Tell You

Not Just What Happened. What It Means.

A useful report should answer questions like:

• What changed this month?
• Why do we think it changed?
• Did the change matter to the business?
• Which channels created real opportunities?
• Where did leads get lost?
• What produced booked work?
• What cost increased?
• What became more efficient?
• What needs attention?
• What are we doing next because of what we learned?

That last one matters.

A report should lead to a decision.

Otherwise it’s just a scrapbook of numbers.

Some very colorful numbers.

But still.

What Should You Ask Your Agency Next Month?

Open the next report.

Skip the first few slides.

Then ask:

How many qualified leads did this produce?

How many became real opportunities?

How many booked?

What did each booked job cost us?

How much revenue can we trace back to this?

What changed from last month?

Why?

What are we changing because of it?

Then ask one more:

Is there anything in this report that looks better than it really is?

That’s a fun one.

A good agency should be able to answer it.

Because there is almost always something.

A weird spike.

An attribution issue.

Seasonality.

Bot traffic.

A channel claiming credit twice.

A number that looks amazing until you compare it to another number.

Good reporting doesn’t hide those things.

It explains them.

The Honest Story Was Better Anyway

This is the funny part of the Fort Wayne Door & Dock bot story.

We didn’t actually need the fake traffic.

Once the noise was removed:

• Bot-adjusted engaged sessions grew 21%
• Organic search sessions grew 16% year over year
• Google Business Profile visibility improved
• AI citations grew
• Calls increased
• The business generated 367 confirmed organic calls across five months

Those numbers were already good.

We didn’t need 62.4% robot traffic to make the campaign look successful.

And even if the real numbers had been worse?

We still would have reported them.

Because the point of measurement is not to make us look smart.

It’s to tell us what to do next.

You cannot improve a system if everyone is pretending the system is already winning.

This Is Bigger Than GA4 Bot Traffic

The Real Problem Is Measuring Activity Instead of Outcomes.

The bot spike is a funny example.

Mostly because:

“62% of your record traffic was robots”

is objectively a pretty good marketing story.

But the larger problem has been around much longer than the bot wave.

Businesses have spent years being taught to judge marketing by:

Traffic.

Reach.

Followers.

Impressions.

Clicks.

Rankings.

Engagement.

When the numbers they actually care about are:

Calls.

Qualified opportunities.

Booked jobs.

Customers.

Revenue.

Profit.

The first list helps explain the second.

That’s useful.

But the first list should never replace the second.

That is the distinction.

Your Dashboard Should Make the Business Less Confusing

If you’re spending money on:

SEO.

Google Ads.

Facebook Ads.

Content.

Email.

A website.

CRM.

Lead generation.

Sales.

Whatever else somebody convinced you to buy last quarter.

You should eventually be able to answer:

What is working?

What is not?

Where is money leaking?

What should we fix next?

If adding more dashboards somehow makes those questions harder to answer, congratulations.

You have created enterprise software.

The point is clarity.

One system.

One scoreboard.

Numbers close enough to revenue that everyone has to play the same game.

That is what reporting is supposed to do.

Keep Reading

Want to See the Actual Campaign Behind the Numbers?

Read: How AI Search Is Changing SEO for Fort Wayne Businesses — Five Months of Real Data, Month by Month →

See the full month-by-month story behind the bot spike, organic calls, Google Business Profile performance, rankings, and AI visibility.

Read: Google’s AI Overviews Cut Local Website Traffic — Our Calls Went Up Anyway →

See why fewer traditional website clicks do not automatically mean fewer search-driven customers.

Read the Fort Wayne Door & Dock Case Study →

See the full results behind the numbers.