Author: Jeff Swartz: Chief Media Office and Partner at Blink Advertising and Principal at Qujam
A commonly used Demand Side Platform (DSP) for display, pre-roll video, and OTT/CTV advertising is Simpli.fi. Its targeting, reporting, and overall capabilities make it well-suited for highly targeted programmatic advertising strategies. While various DSPs are employed across our range of services, the emphasis Simpli.fi places on transparency, value, and quality aligns strongly with our own core values. This alignment has been invaluable as we’ve worked to address a complex issue:
As advertising professionals, we believe in rigorous inquiry and tackling challenging questions. One of the most frequent client concerns in recent years has been: “Why do Google Analytics and Simpli.fi reports seem to differ?” Through investigation, we’ve observed that Google Analytics (GA) does, in fact, often show reporting inconsistencies when compared to data from Simpli.fi (and, more broadly, programmatic digital advertising reports). This has led us to extensive research, including reviewing articles and videos, consulting with experts, and conducting internal experiments on our own campaigns, all in pursuit of a reliable method for reconciling these two data sets.
It’s important to state upfront that we did not find a simple, definitive formula to make Google Analytics and programmatic advertising data perfectly align. However, this process yielded significant insights into improved data interpretation, essential data cleansing techniques, and a clearer understanding of the reasons behind these discrepancies.
Note: While this discussion references Simpli.fi, many of the principles are applicable to other DSPs as well.
Geographic and Attribution Imprecision: Google Analytics, while a valuable tool, does have limitations, as even Google acknowledges. A primary limitation, particularly relevant to this discussion, is that Google Analytics tracks every single user accessing a website. Some might consider this comprehensive tracking a positive attribute. However, we argue that it’s not entirely beneficial. Google Analytics includes non-human activity, such as bots and test clicks generated by ad campaigns. The inclusion of these non-human interactions can distort all Google Analytics data, leading to skewed results.
Furthermore, Google Analytics often struggles with precise geographic and channel attribution. Because it records every “user,” it sometimes makes educated guesses about a user’s location and entry channel. This can result in misallocation of data. For instance, if GA detects traffic originating from the United States but cannot pinpoint the specific location, it might assign the traffic to a central location, like Coffeyville, KS. Other geographic inaccuracies can also compromise the accuracy of location data in Google Analytics.
Similarly, users can be incorrectly assigned to an attribution channel. For example, a user whose origin is unclear might be attributed to a programmatic advertising campaign, even if they arrived through a different channel.
Privacy regulations also introduce challenges to accurate data capture within Google Analytics. For instance, tracking can be significantly hindered by Safari and certain apps due to their privacy-centric designs. Despite these obstacles, GA still records all on-site activity and makes its best estimate of traffic sources when it lacks definitive information. This reliance on estimation contributes to inaccuracies in Google Analytics data related to advertising campaigns, especially considering the prevalence of Safari among Apple users and the substantial portion of display advertising traffic occurring within apps.
In contrast, Simpli.fi adopts a different approach to data reporting by excluding data from test clicks, fraud prevention measures, and instances of missing data. Their philosophy centers on reporting only data that reflects genuine human interactions. Clients are not charged for non-human impressions. Essentially, Google Analytics operates with a broader scope, while Simpli.fi employs a more selective and rigorous approach, focusing on data deemed more reliable.
Test Clicks: Simpli.fi’s data reports exclude test clicks, whereas Google Analytics includes them. Simpli.fi and its publishers incorporate test clicks to ensure campaign quality and functionality throughout the campaign’s duration.
The inclusion of test clicks in Google Analytics contributes to discrepancies between the two reporting sources, most notably by inflating user and new-user counts. In a recent analysis (as of April 2021), it was determined that a significant percentage of GA users attributed to Simpli.fi campaigns were, in fact, test clicks. This inflation of user numbers in Google Analytics consequently affects bounce rates and average session durations, presenting a skewed and less favorable picture.
Consider this example of an average landing page session: If a page has 75 genuine users with an average session duration of 45 seconds, and Google Analytics reports 25 additional users (totaling 100 users) who are actually test clicks with 0-second sessions, Google Analytics will incorrectly report 100 users with an average session duration of 33.75 seconds.
Furthermore, user traffic generated from ad previews within a campaign is recorded by Google Analytics but not by Simpli.fi. This occurs when advertising professionals test ads within the DSP dashboard during campaign setup and optimization. These brief interactions also negatively impact bounce rate and average time-on-site data in Google Analytics. This type of traffic often appears as “Referral” traffic in Google Analytics, attributed to sources like eastads.simpli.fi, westads.simpli.fi, or centralads.simpli.fi, leading to our first data cleansing recommendation.
Data Cleansing Tip #1: When analyzing Google Analytics data for Simpli.fi campaigns, exclude “Direct” traffic with UTM codes containing “scrub” and “Referral” traffic from sources such as eastads.simpli.fi, westads.simpli.fi, or centralads.simpli.fi, as these represent test clicks or ad preview activity during campaign setup.
It’s also worth noting that the volume of test clicks tends to increase when new creative ads are introduced to a campaign. Campaigns with frequent creative changes will generally have a higher number of test clicks compared to campaigns with less frequent changes. This is due to the extra testing conducted by Simpli.fi, publishers, and the advertising agency to ensure the quality and functionality of new creative. These additional test clicks further contribute to inaccuracies in Google Analytics data.
Cookie Complexities: Google Analytics relies on cookies for data tracking. While a detailed explanation of how cookies and Google Analytics interact with digital ads is readily available elsewhere, it’s important to acknowledge that a significant portion of programmatic advertising (especially banner ads) appears within apps. Apps present challenges for cookie-based tracking because they are generally cookie-free environments. Despite this, in-app traffic is still recorded by Google Analytics, which then estimates the traffic’s origin.
Cookies can also lead to other issues, such as double-counting users. If Google Analytics “cookies” a user from a UTM code, and that user returns to the website, they might be attributed to the last known campaign and counted twice. This leads to our second data cleansing recommendation.
Data Cleansing Tip #2: While excluding in-app inventory from programmatic campaigns is an option, it’s generally not recommended. Testing this approach with a multi-targeted banner ad campaign showed that it did not significantly improve data accuracy in Google Analytics. However, it did negatively affect campaign performance by limiting inventory. Specifically, the Cost Per Thousand (CPM) increased slightly, the Cost Per Click (CPC) increased substantially, and the Click Through Rate (CTR) decreased. In this case, the trade-off was not advantageous.
Differences in Data Interpretation: Google Analytics and Simpli.fi inherently present data from different perspectives. Front-end DSPs, like Simpli.fi, primarily focus on performance metrics related to user activity before reaching a website (e.g., impressions, clicks, view-through rates). While some conversion data may be tracked within the DSP, the emphasis is on pre-website interactions. Conversely, Google Analytics primarily focuses on website performance, analyzing user activity after they have arrived. A click and a user are distinct entities, so direct comparisons between the two data sources are often inappropriate.
This discussion aims to provide guidance on interpreting data effectively. However, it’s essential to recognize that ad clicks are not the sole determinant of programmatic advertising success. In reality, most people do not click on ads.
Consider that the average national banner ad CTR is quite low. This means that the vast majority of ad impressions do not result in clicks. Does this imply that these non-clicked impressions are valueless? Certainly not. If that were the case, non-clickable forms of advertising, such as TV, radio, print, out-of-home, transit, and even some digital formats, would be considered ineffective. While it’s understandable that clients sometimes fixate on click data, it’s crucial to remember that a significant portion of advertising value lies in the exposure generated by those impressions.
It’s a misconception that clicks are the only measure of campaign success. There are other key (and often free) data points to consider when evaluating programmatic digital advertising effectiveness:
While sales data is a valuable key performance indicator, it’s important to remember that advertising facilitates the opportunity to sell; it does not guarantee a sale. Increases in traffic and leads, without a corresponding increase in sales, may indicate issues within the sales funnel itself, rather than a failure of the advertising.
In a recent geofencing display campaign for a B2B client, an analysis of Google Analytics versus programmatic data highlighted the importance of looking beyond clicks.
The client compared prospects who were targeted by the geofencing campaign with those who were not. The results showed that targeted prospects were significantly more likely to become leads. Additionally, there was a substantial increase in paid and organic search traffic to their website during the campaign period, further demonstrating the campaign’s impact beyond direct clicks.
Our goal is to achieve optimal results for our clients. We are committed to thorough research and utilizing effective advertising and creative solutions. Please contact us if you have any questions.
Google Support Links:
Author: Jeff Swartz: Chief Media Office and Partner at Blink Advertising and Principal at Qujam
A commonly used Demand Side Platform (DSP) for display, pre-roll video, and OTT/CTV advertising is Simpli.fi. Its targeting, reporting, and overall capabilities make it well-suited for highly targeted programmatic advertising strategies. While various DSPs are employed across our range of services, the emphasis Simpli.fi places on transparency, value, and quality aligns strongly with our own core values. This alignment has been invaluable as we’ve worked to address a complex issue:
As advertising professionals, we believe in rigorous inquiry and tackling challenging questions. One of the most frequent client concerns in recent years has been: “Why do Google Analytics and Simpli.fi reports seem to differ?” Through investigation, we’ve observed that Google Analytics (GA) does, in fact, often show reporting inconsistencies when compared to data from Simpli.fi (and, more broadly, programmatic digital advertising reports). This has led us to extensive research, including reviewing articles and videos, consulting with experts, and conducting internal experiments on our own campaigns, all in pursuit of a reliable method for reconciling these two data sets.
It’s important to state upfront that we did not find a simple, definitive formula to make Google Analytics and programmatic advertising data perfectly align. However, this process yielded significant insights into improved data interpretation, essential data cleansing techniques, and a clearer understanding of the reasons behind these discrepancies.
Note: While this discussion references Simpli.fi, many of the principles are applicable to other DSPs as well.
Geographic and Attribution Imprecision: Google Analytics, while a valuable tool, does have limitations, as even Google acknowledges. A primary limitation, particularly relevant to this discussion, is that Google Analytics tracks every single user accessing a website. Some might consider this comprehensive tracking a positive attribute. However, we argue that it’s not entirely beneficial. Google Analytics includes non-human activity, such as bots and test clicks generated by ad campaigns. The inclusion of these non-human interactions can distort all Google Analytics data, leading to skewed results.
Furthermore, Google Analytics often struggles with precise geographic and channel attribution. Because it records every “user,” it sometimes makes educated guesses about a user’s location and entry channel. This can result in misallocation of data. For instance, if GA detects traffic originating from the United States but cannot pinpoint the specific location, it might assign the traffic to a central location, like Coffeyville, KS. Other geographic inaccuracies can also compromise the accuracy of location data in Google Analytics.
Similarly, users can be incorrectly assigned to an attribution channel. For example, a user whose origin is unclear might be attributed to a programmatic advertising campaign, even if they arrived through a different channel.
Privacy regulations also introduce challenges to accurate data capture within Google Analytics. For instance, tracking can be significantly hindered by Safari and certain apps due to their privacy-centric designs. Despite these obstacles, GA still records all on-site activity and makes its best estimate of traffic sources when it lacks definitive information. This reliance on estimation contributes to inaccuracies in Google Analytics data related to advertising campaigns, especially considering the prevalence of Safari among Apple users and the substantial portion of display advertising traffic occurring within apps.
In contrast, Simpli.fi adopts a different approach to data reporting by excluding data from test clicks, fraud prevention measures, and instances of missing data. Their philosophy centers on reporting only data that reflects genuine human interactions. Clients are not charged for non-human impressions. Essentially, Google Analytics operates with a broader scope, while Simpli.fi employs a more selective and rigorous approach, focusing on data deemed more reliable.
Test Clicks: Simpli.fi’s data reports exclude test clicks, whereas Google Analytics includes them. Simpli.fi and its publishers incorporate test clicks to ensure campaign quality and functionality throughout the campaign’s duration.
The inclusion of test clicks in Google Analytics contributes to discrepancies between the two reporting sources, most notably by inflating user and new-user counts. In a recent analysis (as of April 2021), it was determined that a significant percentage of GA users attributed to Simpli.fi campaigns were, in fact, test clicks. This inflation of user numbers in Google Analytics consequently affects bounce rates and average session durations, presenting a skewed and less favorable picture.
Consider this example of an average landing page session: If a page has 75 genuine users with an average session duration of 45 seconds, and Google Analytics reports 25 additional users (totaling 100 users) who are actually test clicks with 0-second sessions, Google Analytics will incorrectly report 100 users with an average session duration of 33.75 seconds.
Furthermore, user traffic generated from ad previews within a campaign is recorded by Google Analytics but not by Simpli.fi. This occurs when advertising professionals test ads within the DSP dashboard during campaign setup and optimization. These brief interactions also negatively impact bounce rate and average time-on-site data in Google Analytics. This type of traffic often appears as “Referral” traffic in Google Analytics, attributed to sources like eastads.simpli.fi, westads.simpli.fi, or centralads.simpli.fi, leading to our first data cleansing recommendation.
Data Cleansing Tip #1: When analyzing Google Analytics data for Simpli.fi campaigns, exclude “Direct” traffic with UTM codes containing “scrub” and “Referral” traffic from sources such as eastads.simpli.fi, westads.simpli.fi, or centralads.simpli.fi, as these represent test clicks or ad preview activity during campaign setup.
It’s also worth noting that the volume of test clicks tends to increase when new creative ads are introduced to a campaign. Campaigns with frequent creative changes will generally have a higher number of test clicks compared to campaigns with less frequent changes. This is due to the extra testing conducted by Simpli.fi, publishers, and the advertising agency to ensure the quality and functionality of new creative. These additional test clicks further contribute to inaccuracies in Google Analytics data.
Cookie Complexities: Google Analytics relies on cookies for data tracking. While a detailed explanation of how cookies and Google Analytics interact with digital ads is readily available elsewhere, it’s important to acknowledge that a significant portion of programmatic advertising (especially banner ads) appears within apps. Apps present challenges for cookie-based tracking because they are generally cookie-free environments. Despite this, in-app traffic is still recorded by Google Analytics, which then estimates the traffic’s origin.
Cookies can also lead to other issues, such as double-counting users. If Google Analytics “cookies” a user from a UTM code, and that user returns to the website, they might be attributed to the last known campaign and counted twice. This leads to our second data cleansing recommendation.
Data Cleansing Tip #2: While excluding in-app inventory from programmatic campaigns is an option, it’s generally not recommended. Testing this approach with a multi-targeted banner ad campaign showed that it did not significantly improve data accuracy in Google Analytics. However, it did negatively affect campaign performance by limiting inventory. Specifically, the Cost Per Thousand (CPM) increased slightly, the Cost Per Click (CPC) increased substantially, and the Click Through Rate (CTR) decreased. In this case, the trade-off was not advantageous.
Differences in Data Interpretation: Google Analytics and Simpli.fi inherently present data from different perspectives. Front-end DSPs, like Simpli.fi, primarily focus on performance metrics related to user activity before reaching a website (e.g., impressions, clicks, view-through rates). While some conversion data may be tracked within the DSP, the emphasis is on pre-website interactions. Conversely, Google Analytics primarily focuses on website performance, analyzing user activity after they have arrived. A click and a user are distinct entities, so direct comparisons between the two data sources are often inappropriate.
This discussion aims to provide guidance on interpreting data effectively. However, it’s essential to recognize that ad clicks are not the sole determinant of programmatic advertising success. In reality, most people do not click on ads.
Consider that the average national banner ad CTR is quite low. This means that the vast majority of ad impressions do not result in clicks. Does this imply that these non-clicked impressions are valueless? Certainly not. If that were the case, non-clickable forms of advertising, such as TV, radio, print, out-of-home, transit, and even some digital formats, would be considered ineffective. While it’s understandable that clients sometimes fixate on click data, it’s crucial to remember that a significant portion of advertising value lies in the exposure generated by those impressions.
It’s a misconception that clicks are the only measure of campaign success. There are other key (and often free) data points to consider when evaluating programmatic digital advertising effectiveness:
While sales data is a valuable key performance indicator, it’s important to remember that advertising facilitates the opportunity to sell; it does not guarantee a sale. Increases in traffic and leads, without a corresponding increase in sales, may indicate issues within the sales funnel itself, rather than a failure of the advertising.
In a recent geofencing display campaign for a B2B client, an analysis of Google Analytics versus programmatic data highlighted the importance of looking beyond clicks.
The client compared prospects who were targeted by the geofencing campaign with those who were not. The results showed that targeted prospects were significantly more likely to become leads. Additionally, there was a substantial increase in paid and organic search traffic to their website during the campaign period, further demonstrating the campaign’s impact beyond direct clicks.
Our goal is to achieve optimal results for our clients. We are committed to thorough research and utilizing effective advertising and creative solutions. Please contact us if you have any questions.
Google Support Links: