Attribution modeling is the process of coming up with a set of rules on how you will assign credit to the different touch-points in your marketing strategy. This set of rules applies across campaigns as well as channels or devices. The purpose of finding an attribution model is so that you can be able to optimize your touch-points to increase conversions and your return on interest. It applies across the marketing divide, even for SaaS vendors.

Attribution Models

There are different product and service providers. This diversity informs the different marketing strategies that are employed. While some might only focus on social media ads, others might also invest in television commercials, billboard ads, emails, or newsletters. The simpler and fewer the touch-points that you use, the simpler your attribution will be. This is common for businesses whose marketing and sales process is simple and short.  First and last click attribution models can therefore be used in this case.

Some of the most common attribution models include:

attribution-models

Source: pierrelechelle.com, https://www.pierrelechelle.com/tracking-customer-acquisition-saas

However, for SaaS vendors, the marketing and sales process is not that simple. There are different touch-points. These are classified into three broad categories:

  • Creating awareness
  • Generating leads
  • Converting leads to sales

The first and last touch-points are clearly defined. They are those that create awareness and lead to the sale or sign up. However, the midpoints which are lead generators should also get credit since they contribute to the conversion. With this in mind, some Software as a Service vendors use the W-position-based attribution model. This assigns 30% of the credit to the first, last and middle touch-points. The remaining 10% is distributed equally among the other touch-points on the customer’s journey. Although this makes sense in light of the three main steps in the customer’s journey, it is still user-driven.  It’s based on the marketer’s perspective of which touch-point generates leads most effectively. Therefore, you might not actually optimize all of your touch-points to yield the best results. For this, you would need a more complex model.

Cohort Analysis Attribution

Cohort analysis attribution is relatively complex. It follows a group of people who joined the bandwagon at the same time. For example, this could refer to people who become paying customers or sign up in the same period. When it comes to software as service provision, cohort analysis is paramount. There’s a danger in generalizing the number of conversions in a given time period. This ignores the fact that customers sign up or start paying at a given time period. By the time you have gotten the total figure in the next time period:

  • Some customers have dropped out.
  • Others have retained their position.
  • New ones have come on board.

Therefore, to measure the effectiveness of your campaigns, you have to think about these different customers at every time period. This is only possible if you use attribution modeling that will factor in cohort analysis.

Complex Multi-Touch Attribution

There are other complex multi-touch attribution models that can be used by vendors of software as a service. These models would help to measure the effectiveness of online lead campaigns. The Markov model is one of these. It uses the removal principle to measure the effect of a certain touch-point. If, for example, your marketing campaign involves social media, emails and paid per click ads, you would remove one touch-point from your model and assess how it affects the total number of conversions. Repeat the same for all the touch-points in order to come up with data-backed rules for assigning credit. This helps to achieve your goal, which is to optimize every touch-point.

Conclusion

SaaS vendors are better off using multi-touch attribution models that are data-driven. Although they are complex, they are worthwhile. At the end of the day, attribution modeling should be able to go beyond user perception and yield actionable models.

0 Comments

Leave a reply

Your email address will not be published. Required fields are marked *

*

Log in with your credentials

Forgot your details?