Why We Prefer Markov Models to Last Click for Attribution Modelling

So many decisions; which one to make?

You’ve integrated your data streams from your marketing channels. You have access to holistic raw data about the consumer interaction journey. But what attribution model are you going to use to base your decisions on the significance of each marketing channel?

There are a number of different options. You can go for “first-click wins”, where you place all the value on successful conversion into the first online interaction. You can go for “last-click wins”, where the value is placed on the last online interaction before conversion. You can do “even allocation”, and give equal credit to every step from the beginning to the end. Or many other approaches. These models are heuristic; they use a specific, straightforward approach in understanding the complex marketing cycle, which accelerates analysis at the expense of accuracy.

Here at Windsor.ai, we drill deep into data, ensuring that the best insights can be extracted from the information available to you. Our experience has led us to believe that Markov models are the best option from those available for attribution modelling. A Markov model is a probabilistic model, which focuses on specific calculations of the chance that an interaction in one channel will transition to a different state, such as a conversion.

Let’s compare last click to Markov models and then explain why we prefer Markov models for attribution modelling.

What is a last click model?

In a last click model, all the credit for a conversion goes to the marketing channel the customer interacted with directly before making a purchase. All other marketing channels in the chain are given zero credit.

Let’s take the example of a web hosting company, that we’ll call Brand X. They have a number of marketing channels: an active Twitter account that links to another channel (an informative blogpost on the company’s website), television advertising, videos on YouTube, and Google banner ads displayed on a variety of websites.

Let’s say that a potential customer performs a Google search on web hosting. They click an organic search result taking them to Brand X’s website. They browse through the site, reading about the various subscription options and domain names available as well as details about security and ease-of-setting-up website management systems such as WordPress. They watch embedded YouTube videos to learn more about the offers.

They don’t make a purchase yet. They think about it for a few days.

During this time, they find the occasional banner ad from Brand X promoting a discount for new subscribers while they are casually browsing the web. After seeing a few of those banners, they decide to click on one, and then purchase a subscription.

Under a last click model, all the credit for the purchase will go to the Google-linked banner display ad.

What is a Markov model?

For attribution modelling purposes, a Markov model is a way to chart out the customer interaction cycle. You specify individual states in that process as vertices in a chain. You calculate the probability of transition from one particular state in the chain to a different state in a chain (for example, the probability of a visit to a Twitter page to lead to a click to the company’s main website, as well as the probability of that Twitter page visit being the last interaction (leading to a failed conversion attempt).

Let’s go back to our friend Brand X. They peek at the data from their customer interaction history. Let’s say that they have five marketing channels as defined below:

   ChSEO: a search-engine-optimized website, to allow for high ranking in organic search results.

   ChYT: YouTube videos, including ones embedded in the website.

   ChTW: an active Twitter account.

   ChDIS: display ads on other websites via a Google-based ad channel.

   ChTV: television advertising.

You see that there are, say, five different customer “journeys”:

   ChTV -> no conversion

   ChTV -> ChSEO -> conversion

   ChSEO -> ChYT -> ChDIS -> conversion (as in the example given above)

   ChTW  -> ChDIS -> ChSEO -> conversion

   ChSEO -> ChYT -> no conversion

From here, we can calculate the probabilities of customers finding themselves at different parts of the customer journey, and transitioning from those states to another state in the Markov chain. Here’s our chart showcasing that:

Success! We have a straightforward, mathematical breakdown of the chances of a consumer transitioning from one state to another. We know:

• Which channel consumers will start at;
• Which ones they will transition to; and
• Which channels will consumers be in directly preceding a non-conversion or a conversion.

Now let’s plan it out in a Markov chain:

How does this demonstrate the superiority of Markov models?

Let’s go through the same customer journey we looked at earlier:

   ChSEO -> ChYT -> ChDIS -> conversion

In a last click model, we attribute the Google display ads with 100% of the credit for the conversion. However, in the Markov model we see the real value that the search-engine-optimized website has for driving conversion; we see the value it has in driving traffic to the banner ads, as well as its leading role in a number of successful consumer journeys.

The last click model completely brushes aside the value the website has in getting paying customers. Thus, a company relying on this model will be overvaluing one marketing channel while dramatically undervaluing another marketing channel.

The Windsor.AI Advantage

Windsor.AI’s Attribution Insights solution generates easy-to-read Markov models to gain real insight into how your marketing efforts are performing. Contact us for details!