When it comes to analyzing conversion pathways, Google Analytics is a tool you’re looking for. Why is that? First of all, it’s a tool that aggregates other types of data as well, so using it for the analysis of multi-channel funnels is simply very comfortable. Secondly, if properly configured, it guarantees accurate results. Thirdly, Google Analytics, unlike other tools like Yandex Metrica, doesn’t slow the website down (don’t forget that every additional tracking code will influence the speed of the website).

As mentioned before, obtaining reliable data depends heavily on proper configuration. If done correctly, analysis becomes truly effortless.

So where to start? As per usual when it comes to marketing – with a strategy. It’s going to help us greatly with Google Analytics configuration. The first step is setting a model for the customer journey. If we find out that our persona first encountered the brand on Facebook, looked up the website while considering the purchase and then bought the product by clicking on an email – we just found a scenario for channel grouping.

It’s important to remember that Google Analytics in its standard configuration has a built-in channel grouping system, i.e. ‘social media’, that aggregates entries from different social media channels. If we know that the persona uses different social media during awareness and consideration stages, we could create our own Google Analytics channel group that will distinguish them. It’s similar in the case of other acquisition channels – paid and organic traffic, newsletters, etc. 

A proper channel determination should be therefore preceded and affected by the strategy of our activities. We need to be aware of how our persona behaves while shopping to be able to identify if our marketing efforts are delivering the expected results.

But simple channel grouping isn’t enough. Google Analytics needs to be able to distinguish particular types of traffic. We will have to add UTM parameters to each and every link that we share outside of our domain. This is the case with all types of actions, not just advertising ones. Redirecting users from your blog to Facebook? You should make sure the link has the proper parameters that Google Analytics will easily recognize as redirection from a given medium

In that case, it is helpful to create a template of link parameters dedicated to various types of activities you undertake. Thanks to this, the exact same parameter will be used each time for the given activity, and – as a result – the traffic will be read correctly by Google Analytics.

Time for analysis! We should analyze our shopping paths in terms of each channel’s goals that were established in the strategy. If social media is responsible for building our brand awareness, we should not expect it to generate the same number of conversions as, for example, Google Shopping activities that are in charge of closing sales. In short, let’s consider a channel in relation to its goal.

In the context of multi-channel paths, it is also very important to analyze the length of the path – for some brands the path will take an average of 3 days, for others 30 days. Each type of product or service has a slightly different length of the purchasing process. Based on this data, let’s examine how channels work in directing users to the next stages of the customer journey (for example, does a given channel support redirection within the first day, as we assumed?). And it is equally important to configure conversion windows – if a given channel should lead to a purchase within 5 days, it is not a good move to open the conversion window for 90 days (assuming we have the right size of the list of recipients;)), because it will cause a big mess in our data systems.

As noted in the beginning, the strategy is the first step to a reliable analysis. We cannot, however, trust it blindly and refuse to verify it. On the contrary! Analysis of multi-channel paths should also help us spot potential mistakes in our strategy, assess the time and profitability of our activities and notice changes in the users’ behavior. So let’s not be afraid of running A/B tests. They will be helpful for verifying whether a particular channel should be responsible for closing sales – or maybe another one will prove to be more effective.

In conclusion, let’s focus on the accuracy of the configuration and on holistic analysis that takes into account both our assumptions and the users’ behavior (it has not been decided once and for all! The online environment is changing rather dynamically). Let’s keep testing and – of course – let’s not forget about the most important issue, which is the return on investment!