Best Practices for Using Google Analytics to Measure Programmatic Advertising

When Google Analytics was introduced in 2005, it was designed to determine the effectiveness of a SEO or SEM campaign. Despite the growth and changes in digital marketing, many still rely on Google Analytics to determine the success of a campaign. While there are points of measurement that do help, the correlation between website metrics and programmatic may appear differently.

Join us for some best practices in using Google Analytics to help tell the story of your advertising campaign including:

  • Locating discrepancies between platform reporting
  • Better attribution through UTM codes
  • Identifying and explaining test clicks and secure sites





Ann Kraus
Hello and welcome to the webinar series. I’m Ann Kraus, a trainer here at These webinars are designed to give you some tools to help sell, and provide you some insight into the technology. Today you’re going to get a bit of both! You may be familiar Google Analytics and think of it as a mainstay for measuring the success of a digital campaign. You may also be aware that as popular as it is, there are often a lot of discrepancies associated with Google Analytics.

So today we will take a deep dive into the world of Google Analytics, how to use it effectively, and how to guide your advertisers in reading the results.

As usual, there is a handout for this webinar that can be found in the resources section. You might want to print that and follow along. We also have the transcript and the deck in there, and a best practices document that you can print and keep handy.

OK- let’s get started.

Let’s start with the basics. This is for those of you who have heard of Google Analytics but aren’t really that sure about it. So, what is Google Analytics, anyway? Think of it as like an analytical “report card” for websites. It gives insight on what pages are getting the most traffic on a web site, for instance, and how long web visitors spent on certain pages. It also reports on the source of traffic to the site- meaning where was a user before they came to a website?

Google Analytics was designed in 2005, mainly to help advertisers determine how SEO (That’s Search Engine Optimization), or SEM (Search Engine Marketing) campaigns were doing. It worked very well at tracking visitors from desktop and laptop browsers
Since its inception, it helped businesses judge not only the effectiveness of certain digital campaigns, but also how a website itself was performing.

An important thing to know is that Google Analytics relies on website cookies – which, in 2005 wasn’t a big deal.

Lately though, we’ve been hearing a lot about cookies with the heightened interest in privacy and regulations. We are now seeing messages on websites that tells us “this site uses cookies…” and we have to accept the message in order to continue. But what exactly is a cookie?

According to Google,

“A cookie is a small piece of text sent to your browser by a website you visit. It helps the website to remember information about your visit, like your preferred language and other settings.”

So, when someone visits a site, or clicks on a PPC ad, Google Analytics can report how long they stayed on the site and which keywords performed the best, all thanks to cookies. Remember, it is a report card!

So it was 2005… Google Analytics was introduced as a free measurement tool (it remains free today by the way) that relied heavily on cookies.

Then the iPhone came out in 2007, and the app store launched in 2008, with the initial 500 apps. Just to put this in perspective, there are over 2 million apps in the app store now, and the average person has between 60 and 90 apps on their phones.

So what do apps have to do with Google Analytics? Well, browsers like chrome, firefox, internet explorer and safari use cookies – meaning Google analytics can easily track visits that come from these browsers. Meanwhile, the 2 million apps that we use DON’T use cookies.

So, Google Analytics was essentially only measuring desktop traffic effectively. Any ads that were appearing in apps were not being measured accurately by Google Analytics because the app environment is like a paleo diet: “Cookie-less”.

To this day, the app environment is cookie-less and still causes issues with Google Analytics reporting.

This is really important to understand because here At, close to 70% of our ads serve in app. This results in advertisers using Google Analytics and then getting frustrated and cancelling their campaign because the “numbers don’t match”, or “ads were served in the wrong location”!

We do need Google Analytics for advertisers to understand the overall impact of a campaign. But, we also need to be the “interpreter” between Google Analytics reporting and reporting. To help you with this, here are several best practices that you should share with your advertiser.

Let’s start with the Google Analytics pixel.

Just like pixel placement, Google Analytics requires a tag be placed on all pages of an advertiser’s website. This tag will load asynchronously with the page- meaning if the page is requested to load, the tag will then get the request. So, when a person visits the web site, the page (and the tag) are requested from the server that is hosting the site.

But if the page loads too slow for instance, the Google Analytics code might break away so the page can load. Nice, right? Well, yes, for the page load, but not for the reporting. It’s important that the advertiser places the pixel on ALL pages of their web site, and as high up on the page as possible. The higher the pixel is placed, the better chance of it “making the cut” and loading if there is an issue with a slow server.

Now, why would a page load slow in the first place? Well, it could be for a variety of reasons, like poor wi-fi or large photo or video files. When this happens, Google Analytics reports show a session, but not any of the time spent on the page. The advertiser is left with a report showing either a high bounce rate or low page views.

I know I know…. We need to pump the brakes a bit here.

A “session” is a period of time when someone is on a site. This resets every 30 minutes. That means that if you are surfing a website, then take a 15-minute lunch break and come back, it’ll register as the same session. However, if your lunch break is over 30 minutes, when you come.

“Time on page” is how long someone is on an individual page of a site. So, a session can register and low time on a page can register, but if during that session the user visited many pages for many minutes, and the server loaded the pages slowly, then all of those page views may not get counted.

“Bounce rate” is often considered “bad” but hold that thought before you decide. Bounce rate is determined by how many sessions only had 1 page view. Or said another way, a person visited the site, didn’t click around on anything while they were there, and then left. This is a tough one, because in many cases, a person only need visit one page of the site! Maybe they just need a phone number, an address, or the hours of operation for instance. Just because they got exactly what they were looking for without diving deeper into the site, that isn’t necessarily a bad thing. We will take a further look at bounce rate later when we discuss test clicks.

So how does all of this pertain to campaigns anyway? If a user clicks on a targeted display or video ad from inside an app, or maybe they don’t click at all, but go to the advertiser’s site later, will not be seen as the source of that traffic because Google Analytics cannot tell exactly where the visit came from (specifically in the case of mobile app traffic where there are no cookies) or the server or pages were slow (this happens more than you might imagine), or the traffic got there organically simply because they remembered the brand or business name.

Now we need to talk about secure and non-secure sites.

When an advertiser’s ads are clicked on, it’s important that the traffic is driven to a “secure” website. You can tell if the site is secure when there is an HTTPS at the front of the URL. If there is no “S”, the site is not secure.

This matters because unsecured sites may not measure Google Analytics correctly. Going back and forth between secure and non-secure sites has been known to break the code as well. Whether or not a site is secure or not can all be checked prior to launching the campaign. And for the record, making a site secure isn’t terribly complicated or expensive. Advise your advertiser to contact their web hosting provider to upgrade to a secure site. Besides Google Analytics working better, there are SEO advantages to having a secure site as well.

We often here about “landing pages” or “splash pages” that advertisers want to build for their campaigns. This is a classic tactic that has been successfully leveraged with paid search for over two decades.

It sounds like a great idea, right? Drive the advertising campaign traffic directly to a specific page where the advertiser has a special phone number, unique form or distinct Google Analytics code, so they can measure the traffic from their targeted advertising campaign separately from their other online campaigns.

The problem with this as it relates to targeted display is the low CTR that accompanies most of our campaigns. While paid search often sees CTR from 2% to 7% or even higher, targeted advertising averages around .1%. Or said another way, only about 1 in 1000 ads get clicked.

But thanks to the imagery of display and video, our campaigns have something that a ppc ad does not – the ability to promote a brand or a business name for future recall. This 999 who don’t click are still getting the message and that message often influences them – even if they don’t stop what they are doing the moment that the ad is served to click on it.

People who are influenced by targeted advertising often remember the ad LATER, when they are doing other things, and that is when they seek out the advertiser’s website.

But because they aren’t clicking an ad with a special URL that goes directly to the landing page where the advertiser is measuring results, instead googling the business or going directly to the advertiser’s home page, that visit isn’t recorded on the landing page. So, the home page gets a page view and a session gets recorded, but the metrics will not reflect that the visit came from the ad because the landing page wasn’t actually visited.

So, if this is all happening and it not being measured correctly, is there anywhere in Google Analytics that record these “LATER” visits?

The answer to that is yes, but with a caveat. These visits will most likely be accounted for in the “Direct” traffic.

Now, you would think that “direct” would only be traffic that came directly to the site, you know, from literally typing in the website URL into the browser’s address bar. But, Google Analytics ALSO uses this “bucket” when it can’t identify the source of the traffic. So, this “direct” classification can also include mobile traffic from apps (those cookie-less environments), and traffic that might have come from a non-secure site. Or said another way, direct traffic isn’t always direct.

Exhausting isn’t it? It’s kinda like trying to figure out how fast you’re driving in a foreign country. Is 100 kilometers an hour fast? It sure sounds fast, but it’s actually about 63 miles per hour. Translating Google analytics (which was built specifically for SEO and PPC) to work with targeted display and video is a lot like that!

But wait. There’s some good news too. I want to introduce you to your new best friend. The UTM code.

This code is critical for getting some proper attribution on your programmatic targeted display and video campaigns. Here’s why: because Google Analytics reporting can be inconsistent with programmatic reporting, the UTM code you provide will show up in Google Analytics as a source! This will get us out of the direct bucket and into the referral bucket. The referral will list “” as the source allowing you to show your advertiser that attribution.

It really is not that challenging to get a UTM code either! Google Analytics has a page where you simply enter in a few fields and it spits out a fancy new URL that has been modified with the UTM code. So, anything with this code in Google Analytics can be attributed back to your campaign. If no UTM codes are used, zero sessions will be attributed to It really needs to be a best practice. No actually it should be an “always” practice for advertisers who rely on Google Analytics. It will save you so much grief, and will help your advertisers see the value of the campaign.

So rather than ending up in the DIRECT “Bucket” you ards can be seen in Google Analytics as an additional source and ill carry all of this information with them to the reporting. A good new best friend, right?

To access this URL builder- just type in Campaign URL Builder in Google, and it should take you to the URL Builder page
There is a link to the page in the resources section of this webinar. There are also many you tube videos and additional articles on the UTM codes.

Now let’s look at another important difference between Google Analytics and… how we each handle Geo-Locations. This is one of the most common discrepancies that gets noticed in reporting. But again, there is an explanation. Understanding this will go a long way with your advertisers!

Google Analytics reports location details based on the IP address where an ad is served, while reports location details based on the IP address it receives from the exchange during the initial bid request, not the location of the user.

So, if an ad is served to a user who is in Charlotte, NC, Google Analytics will most likely see the IP as being in Charlotte. However, that same ad being served in Charlotte through a programmatic exchange will show the IP location of the exchange used to serve the ad, not the local IP assigned to the device at the time. That could be NYC for example.

And that last bit is kind of important for many reasons. IP’s are not tied to houses, or people, or devices. They can change as one moves from a home, to a coffee shop, to a friend’s house, or to a place of business. And when you look at corporate accounts that use VPN, forget about accurate IP locations. All bets are off.

Obviously, this can create issues when the reporting shows one location on Google Analytics, and another location in reporting because the ad was served from an exchange that was not in the desired location.

I’m going to ask you to please remind your advertisers that a discrepancy in reporting does not mean that the ad was delivered in the location of the IP, but was served from there.

Your advertiser may also notice some additional Geo related discrepancies when they look at Google Analytics reporting. They’ll probably see some odd Geos in their metrics where they did not want ads to serve.

This is the result of running “test clicks”. As we test the creative and click through capabilities of ads, the metrics report from where the ads are being served. For instance, if we have test clicks being served from New York, that is where the ad looks likes it served, and any web traffic following also looks like it is coming from New York. None of this is happening. In truth, the ads are simply being tested from a server that is located in New York. or are simply test clicks.’s reporting will filter out these test clicks, but Google Analytics will not. for this reason, we recommend using our reporting to help the advertiser see accurate numbers.

More location discrepancies that your advertiser may notice happen when they are running Geo-Fencing campaigns. Since Geo-fencing is all about local, local, local, we understand how frustrating it can be to see traffic coming from somewhere that doesn’t look like it targeted locally.

The reason this happens is because captures users who visit a specific physical location and puts them into an audience. Rather than trying to target that audience only while they are in the fence, targets them with ads for up to 30 days – wherever they go! So if they leave the fence and go home, we can target them. If they leave the fence and go to Jamaica, we might target them.

Consider the BBQ restaurant that is targeting his competitors. Prospects that visit those target zones are put into the audience. As they go about their lives… attending school, going to work, going out of town for business, or going to Jamaica, they will be served ads for our advertiser’s restaurant. Those may not be served in the Geo the advertiser wants, but they are being served to the USER the advertiser wants because they became part of an audience when they entered a LOCAL Geo-fenced target zone. Of course, when this happens, Google Analytics registers the different location, rather than the target fence where the user originated.

I’m going to end this webinar with my most important Google Analytics advice and here it is: If you can, get READ ONLY access to your clients Google Analytics prior to a campaign launch. It’s really the best way to guide your advertiser through the reporting and it’ll allow you to SHAPE the story of campaign success.

You can look back a few months, or a few years and see how a site performed before you launched your campaign. Comparing that traffic to the traffic after a campaign launches will help the advertiser see the increase in page views, more searches for their brand name, hopefully some lower bounce rates, and possibly even more time on site or more pages visited per session. In other words, you can associate all the positive changes that happen to the site while the campaign is live BACK TO the campaign – even if it isn’t from clicks.

PS: For those of you who sell traditional media, it’s not a bad idea to leverage google analytics in this same way to show that your TV commercials, radio ads, print ads or outdoor advertising is having an impact by measuring some of these same things inside of Google Analytics.

OK- let’s do a recap: here are the main points that I hope you took away to better leverage Google Analytics and reporting for the renewal and upsell on those campaigns!

  • Google Analytics was designed BEFORE the smart phone, and thus BEFORE apps. serves 70% of ads through apps.
  • Google Analytics requires a cookie for measurement. That cookie- or small piece of text that communicates between websites and browsers- does not exist in apps. Since apps are “cookie-less”, no data will show in Google Analytics.
  • Make sure the advertiser’s web site is secure. If there’s not an “HTTPS” at the beginning of the URL, encourage them to get in touch with their website hosting provider right away.
  • The advertiser must have the Google Analytics pixel placed on all pages of their website, and it should be placed high on the pages so it has the best chance of loading.
  • Every piece of creative being used for the advertising campaign should have a UTM code. This will cut down on referred traffic to a site being reported in that “direct” bucket.
  • Ask the advertiser to share the last six months of their overall web traffic- or better yet- to give you “read only” access to their Google Analytics. Typically, a strong campaign lifts overall web traffic and increases branded terms within Google Analytics.
  • Make sure the destination URL goes to the advertiser’s actual site rather than a landing page or splash page that cannot be searched for, or naturally found, by a user.

Just a reminder to download the transcript and the best practices document that will be in the resources section in Bullseye. If you have any account specific questions, reach out to your account manager. If you have any training questions about this webinar or anything in Bullseye, please send an email to

Have a great day everyone.

Connect With Us

Please complete this form to receive more information.

Learn More.


Request a Demo

Complete the form to request a demo of the platform and see a sample report.

* Required