Here’s A Good Way To Audit & Fix Your GEO In Google Analytics

One of the things Google Analytics is not good at is finding out the location of a website visitor that’s on a cell phone not connected to wifi. In most cases, the IP is associated with a cell phone tower in another location.

I’m sure if you’re in California you see a lot of San Fransisco Traffic. In the Northeast, you will see a lot of Boston & New York Traffic. The good news is most of these visitors are actually in your area.

If you are a car dealer and haven’t figured out by now that you are a brick and mortar business and location matters, you probably shouldn’t read on.

There are two ways to identify your tower cities in your area. You can look at the dimensions “Network Domain” & “Service Provider” combined. I prefer Service Provider because the Network Domain will show a lot of “not set”, which means Google Analytics couldn’t resolve the IP address.

For this example, I’m using a car dealer located in New Hampshire. This dealer does ZERO paid digital marketing, so I thought this would be the best example of raw traffic.

It’s important to know this because if you are paying for traffic you will see these cities comes up. Google Ads uses an entirely different GEO mechanism (for lack of a better word) than Google Analytics. I’ve seen vendors get humiliated by angry customers because of this.

Identify Mobile Users Not Connected To Wifi

I’m going to show you two ways do do this:

Google Analytics

Go to your cities by clicking Audience, Geo, Location, then click city. Now select “service provider” as a secondary dimension.

To make this easier to read, click “advanced” next to the search icon

Copy and paste this for the white box:

t-mobile.*|verizon.*|at&t.*|sprint.*

Now click apply.

Depending on your area, you may have other providers but this will be a good start.

Google Data Studio

I prefer data studio because I can display data so it’s understood more. 72% of this dealer’s traffic came from “another state”, so here I eliminated New Hampshire and showed all suspected cell phone towers.

We can go one step further and rewrite the cities, like this:

WHEN ROI Cell Service Provider = "Verizon Wireless"  AND City = "Boston" THEN "Cell (Boston)" 
WHEN ROI Cell Service Provider = "Sprint"  AND City = "Boston" THEN "Cell (Boston)" 
WHEN ROI Cell Service Provider = "T-Mobile"  AND City = "Boston" THEN "Cell (Boston)" 

You’ll end up with this:

Notice Boston and New York each are listed twice?

Note the traffic quality difference for New York. So now when you look at all cities, you will know the difference.

To avoid confusion, I’ll change the name of the city to “Cell Phone” because it’s irrelevant where Google Analytics thinks they are located. I kept the state column on so I could see the entire picture. I have found 409 active users and every cell phone tower.

Here is a look at a random user. There’s more to show, but let’s keep this simple. This is an active shopper that was in 3 different locations in 5 days. He or she started with Google Organic. Now imagine if it was a paid source? You would think that you were getting clicks from a City you don’t typically sell cars to. Brattleboro and Swanzey are only 35 minutes apart. Brattleboro is in Vermont, and only 35 minutes away from the Dealership.

The way I measure traffic quality, GEO plays a big part of it. Now I can include these 409 visitors in the equation.

Here is an example of what I would have missed from the amount of measurable leads from another dealer during the month of January:

That’s more than a small amount considering these cell phone tower leads.

Let’s test.

Desktop Only

Mobile Only

This helps me when I study the pump in / pump out reports.

This isn’t any type of “proprietary software”, this is an example of data you already have. You just need to find it.

Pump In / Pump Out Reports

This week I’m working on a pump in / pump out report. I love these, but they are so hard to get from dealers! These should be shared with any agency or vendor that is doing anything for you.