In quantifying the influence marketing channels and campaigns have on a customer’s decision to buy, some marketing platforms will ask its customers to choose for a ‘marketing attribution model’. Here’s why this doesn’t make sense.
Visibility into what influences customers allows marketers to optimize their labor for the purpose of conversions, and not just attracting random visitors. By comparing different marketing channels and campaigns, it gets easier to understand the entire conversion path and where best to invest resources. But the big question then is: which portion of the conversion path is most important?
Is it the part where you attract new visitors (so-called “first touch”)? Or maybe the part where you manage to get that lead back (so-called “assisting touch”)? Or is it the part where you finally get your lead to buy (so-called “last or converting touch”)?
The big question is: which portion of the conversion path is most important for your business?
In most cases, this decision falls into the hand of the customer to tell which portion is most relevant for his/her business.
Choosing a Marketing Attribution Model
By most accounts, the traditional way to express which portion of the journey is most valuable is by asking for a ‘marketing attribution model’. Such model has evolved over the years, and can be divided into 2+1 categories:
- Single Source Attribution (also Single Touch Attribution) models assign all the credit to one touch, mostly the first or last touch (model). CRMs and marketing automation tools often show this in one single field.
- Fractional Attribution includes linear weights (every touch weighs in equally), time decay (weight increases for next touches) , customer credit (based on past experiments), and multi-touch / curve models (marketer sets weight)
- Algorithmic (or Probabilistic or Data-driven) Attribution uses statistical modelling and machine learning techniques to derive probability of conversion across all marketing touchpoints. This can then be used to weight the value of each touchpoint preceding the conversion. This is the model used by journy.io. Specific probability algorithms are used to analyze all of the different paths in your account (both non-converting and converting) to figure out which touchpoints are most responsible for conversions, in each phase of the journey. The big advantage of this model is that it isn’t based on predefined assumptions. Thus it doesn’t require any customer input!
The Data-driven Attribution Model used by journy.io doesn’t require any customer input.
Why choosing doesn’t make sense
As mentioned earlier, most platforms will ask its customers to indicate which model they see fit for their business. Here are the different reasons why we —at journy.io— choose not to ask customers, adopting a data-driven model:
- People don’t know: It is extremely hard to understand the complete flow of your customer base, let alone know which portions most contribute to attracting [which type of] customers. So asking often means guessing without any assurance for truth!
- Each channel and campaign deserves respect! In choosing a model such as first touch, drip channels and campaigns will never receive proper weight to what they contribute to conversions. Simply because they will never generate a first touch. (One needs to ask for an email address to start a drip campaign...)
- One journey is not the other. Since different types of customers typically show different behaviours before buying, one campaign will work well as first touch for some, while others —through e.g retargeting— may see that same campaign primarily in their assisting touches.
- Data-driven attribution offer better granular results. Even if one would be certain about the model to choose, these models provide very strict weights to the business. In first touch model, everything which generates a first touch is ‘GOOD!’ (white, or digitally 1) and the rest is ‘BAD!’ (black, or digitally 0). As the rest of the world seems grey, so needs to be the attribution score we give to channels and campaigns and content. And it is the business itself, with its proper unique visitors data that need to set the tone of your model, not some random assumption!