Digital analytics of advertising channels is an integral part of working on marketing campaigns.

One of the most effective approaches in analyzing advertising campaigns is the study of performance indicators across various segments.

In practice, the following popular breakdowns are most commonly used:

• Analysis by advertising campaigns;
• By keywords and ad groups;
• By device type (smartphones, tablets, computers);
• By geography (cities and countries);
• By demographic characteristics (gender and age).

As a rule, during the analysis, the specialist divides the traffic among segments and evaluates the difference in cost per lead (CPL) or cost per order (CPO). Then, appropriate adjustments are made.

Such analysis can provide the specialist with a large amount of useful information, which, if properly considered, can have a significant positive effect.

However, practice shows that standard segments are often insufficient, and many analysts/marketers start to wonder:
What else can we look at in advertising campaigns to increase the efficiency of the attracted traffic?

Let's figure it out!

1. Analyze the effectiveness of advertising by city size

If your advertising campaigns include a wide list of cities (especially if you are targeted to several countries at once), the usual analysis of traffic by location often doesn't work.

Typically, this is due to the fact that the statistics are excessively "spread out" over many geos; such data does not allow you to accurately measure the effectiveness of the long list of locations.

Practice shows that in such a situation, analysis by city size is effective.

Of course, such analysis can be difficult and takes some time to prepare, but it's practical benefits can be very significant.

For example, in the screenshot below (from a real analysis) we can see a significant difference in conversion rates between different size cities (so we can make corrections to the geo-targeting of the advertising campaign):

This report is usually created manually.

Recommended steps to create the report:

• Download the statistics of the effectiveness of advertising campaigns by city;
• For each city, enter the specific range of its population;
• Then, using the aggregated tables tool, create the report to analyse how campaign effectiveness varies by city size;
• If we notice a significant difference in effectiveness between different sized cities, we adjust the geo targeting in ad. account (e.g., keep only cities with effective ranges).

The most commonly used ranges (according to our experience):

• Up to 100 thousand people
• From 100 to 500 thousand people
• From 500 thousand to 1 million
• From 1 to 3 million
• More than 3 million.

Of course, these ranges can be corrected depending on the countries in which our online ads are shown (for example, the ranges for a small Czech Republic would be completely different from the ones for a large country).

Analysing by city size has proven its effectiveness many times before, and we recommend doing it regularly.

2. Explore advertising metrics by browser, screen resolution, and operating systems

While analysis by device type (mobile devices, computers, tablets) is standard and often used, analysis by operating system / screen resolution / browser is much less popular.

At the same time, these analyses can be very useful, for example, they allow you to identify technical problems with the site for different OS versions, browsers, screen resolutions (for example, finding a reduced conversion rate in the Opera browser, we can check the site display in it).

In addition, some advertising systems allow you to make adjustments to the "Android/IOS" segment; we know that these operating systems often show different conversion rates and CPL. This is an extremely useful feature.

For example, in the screenshot below (generated from Google Analytics 4), we see a significant difference in the performance of different segments of operating systems ( maybe it's necessary to check the performance of the site on Android devices):

3. Analyze the effectiveness by landing pages

Analysis by landing pages can be useful and very illustrative.

Imagine: a certain list of landing pages attracts a large traffic volume from several advertising campaigns, each of which is characterized by its own features (geo-targeting, keyword list, topics and interests, creatives, etc.).

The diversity of advertising campaigns can "blur" the received statistics. A breakdown by landing pages can become a starting point for adjusting advertising campaigns.
Thus, it may turn out that there are a number of landing pages that show higher performance than the rest; having analyzed this, a specialist can try to find out the reason: it may be a semantics issue, it may be a specific product category, sometimes technical problems are found on the low-performing pages.

For example, in the screenshot below we see problematic landing pages with low conversion rates:

This is a convenient and effective method. If standard advertising traffic reports did not provide new information, we recommend using additional breakdown by landing pages as one of the first analyses.

4. Analyze advertising campaigns by "new vs. returning users"

Conversion rates between first-time and repeat visitors can vary significantly.

Measuring this difference will help you understand if you need to separate campaigns for new and repeat users. Sometimes the difference is so large that it makes sense to manage bids separately for new/repeat audiences.

In Google Analytics 4, you can use the "New/Returning" parameter to measure this segment:

Having discovered a significant difference between the segments under consideration, we can not only manage them through bid adjustments (this is the standard approach), but also create separate campaigns where targeting is strictly limited to either new or returning users.

5.Segment the semantics of advertising campaigns by thematic groups and analyze the metrics for each group

Similar to cities and landing pages, very often the conversions by which you measure search ad campaigns are "spread" between keywords (or search queries). This makes analysis more difficult.

By grouping keywords into sense groups and measuring the performance of each group individually, you can find segments that perform significantly better or lower than the average campaign results.

Usually the work is based on downloaded key phrases: each keyword belongs to a certain sense group, after which a summary report with performance metrics is generated. Of course, this is quite a time-consuming process, but the result often makes it possible to significantly improve the quality of advertising campaigns.

What attributes can be used to organize key phrases into semantic groups? There are a large number of variations, for example:

• Existence of "hot" additions in the key phrase ("buy", "price", etc.);
• Phrases with geo and keywords without geo;
• Phrases with and without brand names;
• Phrases with a high level of specificity (e.g. exact model name) and other (broader) phrases;
• Phrases grouped by different product categories;
• Phrases with higher and lower number of words.

The variants of sense groups can be very numerous, and it is up to each specialist to decide which variants to use at any given case.

Finding semantic groups with high metrics, you can, for example, bring key these phrases i*n separate advertising campaigns*.
In contrast, keywords from the group with low metrics can be stopped.

6. Analyze the effectiveness of search advertising based on the ad's position in search results

In a range of businesses, the conversion rate in search campaigns can vary significantly depending on the place from which the ad click was made: from the first place or below.

Such dependence is characterized by "urgent" themes (car tow truck, opening doors, etc.), as well as any themes with a very high level of competition in the search advertising results.

It may turn out that the increased conversion of ads from the first positions will compensate the high cost per click, and it is needed to maximize the volume of impressions in the upper part of the search advertising results.
The reverse situation may also be true, when savings from a lower cost-per-click for out-of-top ad place easily compensate for lower conversion rates.

If you find that performance is affected by ad place in search results, this should be taken into ad account while setting bids.

In the example below (from the Google Ads report), we see a significant difference in conversion rates between the top and other positions:

7. Examine the IP address ranges for advertising traffic

Analysis by IP addresses can be useful if you are worried about "attacks" by bots.

Advertising systems always claim that their bot filtering is extremely effective due to automated defense algorithms. As practice shows, this is true, but sometimes these algorithms fail (we recorded the last case of such a problem at the end of 2023).

If you identify potentially problematic IP ranges, you can exclude them (some advertising systems allow you to do this) or temporarily disable those campaigns / ad groups that bring in the bulk of clicks from bots.

It is important to remember that manual IP address verification works only in quite simple cases, when boat traffic works on the basis of primitive algorithms.

For example, in the below screenshot, we can see that there are several IP address ranges that raise suspicions and require additional research:

8. Investigate the "unknown" segments in advertising accounts

The "unknown" ("undefined", " undetermined") segment is most often tracked in demographic parameters ("gender" and "age").

This is an extremely interesting segmentation to analyze, because very often the "undefined" segment demonstrates effectiveness that varies significantly from the average performance of an advertising campaign.

In addition, in rare cases, an excessively high percentage of conversions from the "unknown" segment is a sign of an overload of bots on your site.

It is important to remember that in analytics systems (such as Google Analytics 4) "unknown" segments are not either present in the demographic reports or are calculated differently than in the advertising cabinets. It is recommended to measure the availability and impact of this segment directly in the ad systems.

Below is an example of what the "Undetermined" segment looks like in Google Ads reports:

By discovering significant differences in the metrics of the "unknown" segment compared to the average values, we can make the necessary adjustments in advertising campaigns.

9. Examine the placements of display advertising campaigns

Analysis in the context of the platforms where media advertising was shown (websites, apps, Youtube channels, etc.) is actively used by almost all experienced marketers. This section describes not the analysis by placements itself, but the recommended way to apply its results.

In general, placement research is a really useful and necessary type of analysis.
As usual, problem placements, identified during the research of statistics, move to the "negative placement" (advertising display on these placements is not allowed).
In the majority of cases, problem placements are clearly non-relevant sites / apps (for example, resources with children's games), or placements for which a significant budget was spent without receiving targeted conversions.

The problem with this method is that for advertising campaigns with a wide reach and a large traffic volume, such placement sorting can take a lot of time, and all this time you will spend your budget inefficiently.

**The other way is often more effective.
**It requires media campaigns to "work" in pairs.

The first ad campaign works normally and is used primarily to collect effective sites / apps. This ad campaign works most often with a limited budget.

The second ad campaign is designed to maximize the use of effective placements. Thus, having found that a selected site / app is effective (actively generating conversions or demonstrating good user behaviour), we *move it from the first ad campaign to the second *("hot campaign"), adding it to the previously selected effective placements.

In the "hot campaign", we maximize the positive impact of quality sites / apps, making sure our ads get maximum reach (usually by increasing budgets and bids).

Example of placement report in Google Ads:

This method requires regular iterations of analyzing and collecting quality sites / apps, but the effect of a well-built hot campaign can be very significant.

10. Analyze advertising metrics based on micro-conversions

The concept of "microconversions" can be understood in different ways. For the purposes of this section, we mean microconversions as events that are actively achieved by users valuable to the business (those who leave leads and make purchases).

The list of variations of such events is quite long, popular ones are the following:

• Long time spent on the site;
• The fact of the beginning of checkout process;
• Scrolling a long site page to the bottom;
• Visiting important pages of the site;
• Downloading files from the site.

According to our practice, micro conversions are best used to identify ineffective segments for deactivation.

The use of micro-conversions to identify effective segments should be used with caution; experience shows that it is better to identify such segments based on full leads and transactions.

A popular recommended scenario of using microconversions in practice is as follows.
Suppose you need to significantly reduce your advertising budget, but the volume of targeted leads/sales is not large enough to perform a high-quality analysis. In this case, you can select one or more microconversions, and then research how actively they are achieved in the different segments. Those segments that are not characterized by a large number of microconversions become the first in list for deactivation or bid reductions.

This method is quite simple, but it allows you to effectively identify problem segments in your advertising campaigns.

Example of microconversions split by country in Google Ads:

11. Comparative analysis of paired time periods (successful and unsuccessful)

In some ways, this type of analysis is derived from segment analysis and can be useful if you have detected an unexplained decrease (or increase) in ad. campaign performance.

So, let's imagine that you've found a dynamic in the results of your advertising campaigns that you can't explain. If, according to the Change History, there were no significant changes in the "problem" campaigns, we recommend performing benchmarking according to the following steps:

• Identify a period in which there has been a change in metrics for some time;
• Find a comparable period with good ad conversion rates (preferably, this period should be chronologically close to the first one);
• Compare the previously selected ranges in the analyzed reports;
• Sequentially split the report into different segments, trying to determine in which segments t*he distribution of traffic or conversions can have changed* between periods.

An example comparing the distribution of clicks by "age" segment over two periods:

Such a report can be extremely useful, because it clearly shows the traffic "flows" between segments and, thus, telling the specialist where to pay attention first.

Now you know how versatile the study of incoming advertising traffic can be and how many opportunities exist to improve the quality of advertising campaigns.

We hope our article was helpful to you! Share in the comments which methods you most frequently use to improve the performance of your advertising traffic. Or contact us for an analysis of your marketing campaigns :)