Analytics tools today must meet high standards. In what is a relief to those who geek out about this, the industry seems to have moved on from limiting their view to only last-touch attribution and calling it a day. That isn’t to say this type of view wasn’t helpful. But, it did mean analytics teams needed to explain time and time again how most ad platforms use a conversion window, and having an internal system set up for last touch would mean the number of conversions showing would differ from what the ad platform said. This would typically lead to burning a lot of resources trying to figure out what number was ‘real.’ All this additional effort and review really derailed the point of marketing analytics – finding meaningful patterns in the noise to drive business.
Luckily, as analytics systems advance so does their ability to present more complex information in ways that are easier to understand. This is especially fortunate given that some of the most common and simple-sounding questions about marketing efforts often have complex logic underlying the answer they are providing. Think of the question: “What things that we are doing are working the best?”. At face value, a last-touch model providing conversions could answer that question. But, that frequently misses the real intent of the question “what are we doing that is contributing to pipeline and eventually won deals”. This version of the question feels much more focused on figuring out every touch point that contributed to a conversion and trying to connect them in a meaningful way. This process has become known as journey analytics.
Journey analytics is inherently a very data-heavy practice. By definition you are trying to look at every touch point you’ve had with a prospect or customer and connect possible dots between them. Up until fairly recently, this was more or less unattainable by all but the most well-equipped organizations. Not only did you need rock-solid data, but you needed a team capable of handling that data. The end result, if done well, was the ability to actually see which touch points matter most and whether the order of these touch points mattered. Knowing this information could help craft a very effective marketing strategy. Think of it this way:
Content piece A generates leads at a 3% CVR
Piece B generates leads at a 2% CVR
Piece A then B also generates leads at 2%
Piece B then A generates leads at 5%
If you were just looking at last-touch efforts the conclusion would be that piece A is superior to B and that cutting piece B should improve conversion rate. While this is true, it fails to understand that showing B prior to A increases conversion rates even further. It seems simple with only 2 pieces of content but imagine trying to solve this for 20 or 100. Things get very complicated, very quickly…
Fortunately, a lot of tools have started shifting towards event-based collection systems and expanding their software to accommodate and take full advantage of what an event-based system offers – especially the ability to do multi-touch attribution much better. This improvement to multi-touch really is key because it allows analysts to make a timeline of a visitor’s interactions sequentially, as well as look backwards down that timeline to identify which point, or series of points, seem to be the most influential in generating leads or purchases. This ability to look back at individual journeys, known as Reverse Path analysis, is far more impactful than it may seem. Until this type of analysis was possible, most analysis aimed at determining which touch points were most impactful relied on aggregates and averages. IE: purchasers who experienced touch point D converted at 3%, those that experienced touch point E converted at 2%, and those who experienced both converted at 4%. It failed to see if there was a difference between what order the touch points occurred in, especially when the touch points were many and occurred over numerous sessions. Now, with modern tools like Indicative analysis like this comes standard and is given it’s own GUI elements to allow for much easier manipulation of the data being looked at.
This granularity in data really can make all the difference when it comes to developing strategies. Rather than doing a lot of complex math to assume what order touch points occur in and whether that matters, you can fairly directly show what effect it has in many of the available tools right out of the box. Data engineers will be able to save a lot of time on manipulation and model development and focus on what matters most to the rest of the company – answering questions and guiding the team towards what works best.