Conversion segments in Analytics that will make you look as experienced digital marketer
Sometimes we all are asking ourself questions about real impact of different digital marketing efforts on our ROI, or on our conversions. But, there is always one follow up question: How can I see this information in my Analytics account?
In analytics we have one easy and very important possibility. Yes, I am thinking about Multi Channel Funnel reports – but, this post is about something else! MCF reports are great for insights about all different channels and their influence on our goals. But, there are some cases when it is very difficult to see certain information that we need.
For example, let us say that we started display campaign and now we want to see its impact on total number of conversions thru different channels and their channel path. For example, we also want to see how many people have remembered our brand (display banner) and came directly to our web and converted.
One way to see all that is to use Conversion segments in Google Analytics. It is in one way similar to creating new segments in Analytics. But, when looking at Multi-channel Funnels reports we can not use our custom segments.
This post will show some of ways, or should we say – ideas, in which we can use this conversion segments. Those ideas will generally be about pay-per-click campaigns and their assistance on other channels. This ideas can also be used for all other channels as well.
- Conversion Segments for better channel insights
- Things to look out for when setting up custom conversion segments
- User Defined Conversion Segments in Analytics
- Impact of Display campaign on brand strength
- Different aspects of Assisting interaction and segment that must always be used
Conversion Segments for better channel insights
But, let us start from beginning. Conversion segments can be find just at top off any multi-channel report.
After we have clicked on that “button” we now have 2 options. Default segments on left, and user defined segments on the right.
We can right away see that we have several options from default segments. Some of this segments are good but we are not going to describe them thoroughly here. We are going to create new segments that are going to be of a more use to us than default ones.
When we choose to create new conversion segment we are presented with something like this:
I am going to steer you attention to first option -> types of interaction. Although they have self-explanatory names, there is something that you must keep in mind when working with assisting interaction.
But first: any interaction = interaction that has happened on any step in channel path resulting with conversion/goal.
For example, if we choose from default segments “Any interaction is referral” we could see something like this:
As we can see – referral in this case can be first interaction in channel path, can be last interaction or can be middle interaction in a path to conversion.
First interaction = interaction that has happened on first step in channel path resulting with conversion/goal.
For example, if we choose from default segments “First interaction is direct” we could see something like this:
As we can see – direct in all this cases is always first in channel path. If we look at second example from picture above we can see that direct is first step but also a last step. When we select “first interaction” then it is only important that “direct” (in this case) is first step in any path.
Last interaction = interaction that has happened on last step in channel path resulting with conversion/goal.
For example, if we choose from default segments “Last interaction is Organic Search” we could see something like this:
As we can see – organic is last step in every channel path shown above. But, pay close attention to first and last example in picture above.
Organic search in first example is the only step that has resulted in conversion. But, he is also a last step in that channel path. And, yes, he is also a first step. Now, let us have a look at last example.
There are 2 interaction in which both of them are “organic search”. But, first step, or should we say first interaction, is “organic search” and, second step is also an “organic search”.
We can get even better results by using different filters and options in user defined conversion segments. But, before we go to examples we first must go to some information about different “source” and “medium” of your web traffic.
Things to look out for when setting up custom conversion segments
When we are setting up a custom conversion segment we are going to use different “source” or “medium” according to our needs (or what we want to show). And here is great chance for you to make a mistake.
From picture above it is very clear that we have AdWords campaign targeting search and display networks. If we look at their source/medium we are seeing “google” as source and “display” & “cpc” as “medium”.
But, if you want to see impact of some other PPC campaigns that are not adwords – then look at their “source” and “medium” labels (names) in analytics and use them when creating new conversion segment.
For example, picture shown below is showing source/medium for Display campaigns:
See anything different than before? Yes, we can see Google “source” but “medium” is cpm! Also, there is “dfa” source which is label for DoubleClick.
Besides “medium” and “source” we also have several other options we can use for our filters. There are some great options if you want to see impact of your adwords campaigns, ad groups, keywords, other campaigns (utm anyone?), ad format, ad content…
But, there is also option to choose Multi Channel Funnel default channel grouping.
Whatever we are going to use – we must be very careful. We must first investigate all data that we have in analytics and only then when we are very certain – we are proceeding with creating different user defined conversion segments.Only when we are fully informed on all sources of data can we start creating a conversion segment. Click To Tweet
This is even more important when we have custom campaigns with custom utm tags.
User Defined Conversion Segments in Analytics
Because we want to have better insights for our different channels, and marketing efforts that are pushed to different channels, we are using several different filters. Majority of examples shown here is in some way connected to AdWords or PPC campaigns. But, ideas behind them can be used for any other sources or needs.
First simple example is when first interaction is AdWords and any other interaction is organic. We want to see impact of our AdWords campaigns on “organic” traffic.
For this Conversion segment we have decided to use “source/medium” filter for first interaction. Because we want to look at impact of AdWords campaigns their source is “google” and their medium is either “display” (for campaigns targeting google display network) or “cpc” (for campaigns targeting google search network).
And, we have also included “any interaction” is for “medium” containing “organic”. Some of the results when looking at conversion paths are shown below:
We can now see all conversion paths in which first step was from adwords campaign AND any other step was also from organic search. In this way we can see what impact AdWords campaigns have on organic traffic to our website and their uttermost impact on conversions.
We see that this segment is around 0.32% of all conversions. But, what if we would like to see only cases where first interaction is our AdWords campaigns AND last interaction is organic search?
We have only changed “last interaction” from “any interaction” for “organic”.
We now see only cases where first step was adwords campaign and last step was Organic search.
We can now compare these two conversion segments. When any other step was organic we had 128 conversions (0.31%) but, when we set up organic search as last step in conversion path we had 70 conversions (0.17%). As we can see – these information can help us in getting the big picture about performance and impact of different channels.
Impact of Display campaign on brand strength
Sometimes we want to know just how really good are our display campaigns. To help us find out answer to that question we can look at examples described below.
First, let us say that we want to see if people who clicked on our display ad are coming to our website directly (and in what paths).
In this example we are using analytics account where display campaigns have medium “cpm”. Also, for last step we have included “medium” (none) which is marking for direct traffic.
There are some interesting info here. We could say that there are people who have bookmarked our website, or directly inserted web address in browser, after they have clicked on our ad. Afterwards they have come to our website and converted. We could say that using this conversion segment we can see impact of display campaigns on “direct” channel.
But, what if we have active AdWords display campaigns along with active brand search campaigns and we want to see if display campaigns have any impact on increasing clicks on AdWords search campaigns (brand keywords)?
In second filter we have included AdWords campaigns who have “brand” in their name. Also, we have decided that it is last interaction.
Different aspects of Assisting interaction and segment that must always be used
Google analytics is sometimes very strange when you look at it deeply. But, it is also very helpful. The thing is that as we know more and more about it, as we learn more about it – we can look “with different perspective” on same data as before.
Assisting interaction is one of those things. For example, let us look at all assisting interactions from our display campaigns (whose medium was “cpm”).
Pretty easy to set up. But, if we look at our top conversion paths – we could see something like this:
In first path we actually see that display was last step. Also, in second path we see that display was only step.
But, everything is ok. This is just a way that analytics is counting steps (interactions). For example, we have first interaction that is from “display” campaigns and afterwards we have another interaction from display from which we have conversion!
First interaction (from display) is actually assisting interaction to second interaction (from display). So, we could have case where multiple interactions are from same “channel” and where last interaction (from that same channel) ended in conversion which makes all previous interactions actually assisting interactions.
The (not-so-perfect) solution is easy. We just need to exclude last interaction from same channel. For our example it would look like this:
After we have implemented this segment the result could be something like this:
Now, when we know a little more about how analytics is counting and assigning different types of interactions – we can proceed to make conversion segment that can always be used. This segment is of great help when looking at assisted conversions in analytics.
This is typical assisted conversion report in analytics. But, can we make it even better with our conversion segments? We can certainly try!
We have created segment in which we are trying to include channel paths where any interaction (assisting and first) is from CPC, but, excluding those where last interaction is from CPC.
Pay close attention to number of assisted conversions for “paid search” and sum of all last click conversions for all other channels. Interesting, isn’t it?
We have made segment that is showing “real” number of assisted conversions. Now, word “real” is in apostrophe because nothing is wrong with default way in which analytics is counting conversion. But, in this way – we are seeing more logical connections and impact of our paid search campaigns.As we know more and more about Analytics - we can look at it with different perspective. Click To Tweet
In same way we can make another segment where we can see impact of search and display campaigns on other channels.
In this example (for this client) we have included search campaigns and his display campaigns.
Of course, the way you are going to include filters for your segments is up to you. Instead of “medium” you can use names of your campaigns, or certain keywords, or whatever you want. The thing is that in this way you can more clearly see impact of you digital marketing efforts/channels.
We did not include all segments that we are using in this post. It would be too big to read if we did that. All we can say is to start testing with different segments and see what is best for you. Possible combinations are numerous.