Multi Channel (shm)Attribution (sh)Modeling

Traditionally, conversions are calculated and attributed to the “last click”, or the source of the visit that resulted in a conversion. The main problem in this model is that the user has probably had several previous interactions with your site / brand before deciding to come and buy something. These prior visits have somehow influenced the users to trust your brand, become more familiar with it, and come to the site to buy.

There are attempts to quantify and even assign a dollar value to each of those interactions. There are models of attribution, where there are different ways of giving weights and importance to each touch point. This is an exercise in futility and an impossible task.

Assuming that only previous online interactions have resulted in the purchase is one big fat error. People don’t only interact online, they can physically visit shops selling your product, they might see them in action with their friends, interact with your sales people, and a million other ways to get to know you more.

The way multi channel attribution seems to work is by measuring all the visits to the site, leading up to the purchase. This also misses an important point. People can interact with you online on a deep level without visiting your site, especially on social media, other blogs and sites, or maybe if your products are listed on a different independent site, where people get the real reviews.

So there is a huge amount of interaction happening offline that you will never be able to measure (at least not these days!), and there are many other online interactions that are not happening on your site that you cannot measure either.

Even if we come up with tracking methods that get us much closer to tracking as many interactions as possible, I still think we cannot get to attribute conversions to a a set of sources, especially when many are in play.

What if someone got completely convinced to buy a product by watching a great video presentation or documentary? How can you measure the effect of that? Even that same video that was the tipping point for many people, influenced them in different ways. Some people were convinced by minute three because they look at numbers and saw good financials at the beginning. Some people had to watch it until they saw human beings talk about using that product, and were convinced then. Good luck measuring that!

People use different devices and we are buying new devices as frequently as we are buying new clothes, another factor that complicates the equation even more.

The concept of attribution therefore is an impossible thing to achieve, not only because of technological limitations, but because the factors influencing buying decisions are numerous, complex, they interact with each other, include human emotions (quite complex), and people are different and react differently to a different combinations of touch points.