More Than Vocabulary

Stamats Insights

Share On  

Understanding the Relationship Between Attribution Modeling and Social Media Optimization

In the final analysis, the only thing that matters is results.

This is why seasoned marketers are constantly looking for tools that will allow them to better separate marketing wheat from marketing chaff. In other words, they want to know what’s working, and why, and what’s not.

To help everyone get a better handle on these concepts, this blog and our next will focus on two terms that are driving much of the conversation:

  • Attribution modeling
  • Social media optimization

This blog will look at the idea of attribution modeling. The next will look at social media optimization.

Attribution modeling

In the not too distant past it was simply called mROI, or measuring marketing return on investment. We looked at simple campaigns with simple goals: Did we get the class? Did the amount generated by the annual fund increase?

Attribution modeling (sometimes called multichannel attribution), involves the use of statistics, sometimes very high-end statistics, to determine and even allocate proportional credit to complex campaigns that feature multiple online media.1

We are no longer interested in just knowing whether or not we got the class. Now we want to know the effectiveness of specific messages and channels within the larger campaign, especially those messages that converted prospects from one stage to another.

Rather than looking at broad trends, attribution modeling analyzes users (either as individuals or in segments) to measure the value and contribution of specific marketing tactics within a larger marketing initiative.

For attribution modeling to be successful there must be a high degree of confidence in how the data was collected and whether it was correctly tabulated.

And then, of course, there is the analysis. In the past, this process would have required a team of data scientists, advanced computing ability, and days (or weeks) of analysis. Today, this process is greatly enhanced and hastened by advances in processing and the use of machine learning. At Stamats, we employ machine learning algorithms that assess how effective each marketing activity is at bringing about a desired result (response, inquiry, application, etc.).

Finally, because of the volumes of data that are available, analysts often use dashboards and scorecards to present the data to increase understanding and application. Scored activities with clear impacts on student behavior provide a data-based guide for effective marketing plans.

Attribution modeling is ideal for evaluating online campaigns with multiple messages and channels in play. It is also extremely helpful when you (or the client) is comfortable with continually (often daily) refining the marketing strategy. Finally, as noted earlier, attribution requires a high degree of numerical fluency.


1The ease and timeliness with which data can be collected makes attribution modeling much more effective in the online space.


In May, learn about the relationship between attribution and social media optimization in Part II of this blog series, More Than Vocabulary.

About the Author