Attribution Part 6 - Channel Attribution Models
In the previous two posts I have talked about single source attribution models where all of the credit for a sale is attributed to one touch, and multi-touch attribution models which split credit across multiple touches based on their position related to the sale.
Rather than being position based, the Custom Attribution Model looks at your marketing channels and weighs the influence of each on the sale, regardless of position in the journey. In the example below, you can see that the blog gets 15% of the credit for the sale regardless of where it falls and how many touches it gets. Unlike the previous models, this approach allows you to add specifics for your industry, marketing channel factors and typical buyer behaviour to your model.
This additional data makes this model more accurate but because the position of the channels is not measured, it is best for businesses with a very defined customer journey. Additionally, it requires a lot of data to get the weightings right and so can be difficult to build, explain and use.
Like the custom attribution model, the Weighted Multi-Source Model assigns different weights to each marketing channel then also accounts for the position in each customer’s journey.
In the post on multi-touch attribution, I had a small rant about paying Google to get a well nurtured lead back to your site. With the weighted multi-source model, you could weight a Google Ad high in the early stages of the customer journey and low or even negative towards the end.
This is the most nuanced model we have looked at so far. It allows you to apply all of the information you have about your industry, your business, your customers, the various marketing channels and the customer journey, but with more difficulty in building, explaining and using the model.
We have now looked at nine different models but what we are really trying to analyse. We have hundreds or thousands of people buying from us. Each one on a unique journey with different start points, number of steps and different end points. All while we are changing our business with different product sets, pricing strategies and sourcing models both regionally and globally. On top of this are external forces that are largely beyond our control like weather, the economy and political uncertainty. None of the nine models are really equipped to handle it all, which is why in the early 2000’s a group of representatives from trade associations, agencies and brands got together and developed Uniform Marketing Measurement (UMM).
UMM is an algorithmic approach using your customer data, business data and external data to report on the success of your marketing activities. This new approach was heralded by Steven Levitt author of the book Freakonomics as having “the power to cut through the Gordian knot of advertising and marketing complexity.”
“...the power to cut through the Gordian knot of advertising and marketing complexity.”
Unfortunately, UMM has gotten a bad rap over the years, even falling into Gartner’s Trough of Disillusionment until 2022 or beyond. In reality, UMM was ahead of its time. In order to work, it needed cloud computing, machine learning and organised business data. All of which was more than a decade away from its 2005 launch.For years, only the biggest companies could adopt UMM and then only on the biggest projects. It did nothing to help the marketers on the ground. Even today with easy access to the building blocks needed for UMM algorithms, they are still very costly to develop, hard to maintain and very hard to defend in the board room.
UMM will be the wave of the future for some businesses but it is certainly not for everyone right now. In the meantime, we still have to answer the fundamental questions; “how do I hit my target” and “what is the best way to spend my marketing budget? In the next post I will introduce an alternative to traditional attribution modelling that can be done by any marketing team, answers the fundamental questions and quickly delivers actionable information that will transform your business.