Know your Customers Better with Path Analysis




How do your channels really interact? Which channel is an opener and which one is a closer? Answers to these questions have traditionally been produced by a qualified guess. We have a better solution - actual reconstruction of paths, based on collected data. Findings resulting from this reconstruction are presented in Path Analysis. 

**The Path Analysis section can be found on the left-side bar of the Roivenue application under the "Marketing" dropdown arrow

In the previous section we introduced the data driven attribution modeling feature of Roivenue. In order to be able to calculate these models, we need to track full conversion paths of your customers, which is done by custom measurement setup implemented during the onboarding phase. Then, once we have these paths fully mapped, more opportunities open up. We can now learn about true length of paths and what they are comprised of. Knowledge gathered from paths is summarized in the Path Analysis feature. Therefore, this feature only works with implemented enhanced attribution.


One of the most important numbers on this initial dashboard is the average path length, located on the bottom left corner of the dashboard. The longer the average path of your customer, the more expensive each conversion will be (with all other aspects equal). This also closely relates to your product - if you are selling kitchen tables, your paths are going to be longer on average compared to a company selling sunglasses.









For each channel, we can see how many times it had customer interactions within the time period and their share of total interactions. Also, for each channel you can see if an interaction is understood as a click or an impression - impressions are available for the Roivenue Brand plan. This table already gives you a hint about which side of the conversion path each channel leans to. But to get the full, visual picture, navigate to the neighboring tab - Position in Path and Path Length.






With this dashboard, you now have all the evidence about real steps taken by your customers. It is time to throw away your preconceptions about the role each channel plays in your channel mix. With proper measurement setup, channel dots should organize themselves in a quasilinear relationship - channels with larger average position tend to be in longer paths, therefore they tend to be located in upper right quadrant.
Next tab, Conversion and Non-Conversion Paths breaks down paths by length and distributes channels by their positions. Each line represents a different path length - line 2 contains 2 touchpoint long paths. The Channel bubble on the right side signals channel tendency towards last touch, while the left side shows otherwise. For example, if you have been looking on line 3 which had 10 occurences, a channel on the very right of this line would be on third position of every path out of the 10, while each of these paths had exactly 3 touchpoints. Clicks and impressions are again divided - click channels are full dots, impressions are circles.


All is set in a familiar setting, with a working right side filter. Path Analysis introduces several tiles with information about paths relating to the business unit. Information such as average path lengthaverage path cost and path mROMI are displayed.







Underneath these tiles, three views presenting deeper path insights can be found. Participants contribution breaks down path related metrics such as interactionsunique pathstypes of conversions and first touch / last touch position by channels.











Position in Path and Path Length scatterplots channels is based on two criteria - average position in path and average path length. Each channel is displayed as one dot on the scatter plot, while the size of each dot depends on its number of occurrences, or how many times has the channel been touched by a customer. If the dot is displayed as hollow, this channel is considered as an impression and not a click.















To take a deeper look into what each of these lines is comprised of, you can open its detail by clicking on the row. In the picture below you can see paths with length of 3 by channels, ordered by the number of channel interactions. The first column displays the amount of cases when the channel was on the first position in the path, second column on second position and so on. To learn about number of interactions per position, mouse over the selected column to find additional information in the tooltip.

If you are interested in even more detailed overview of your customer paths, your Client Success Manager can provide you with path data by specific channels. On this dataset you can explore real actions of past customers and evaluate various paths by their conversion rate and ROI.