The Multi-touch Attribution Model is a marketing measurement method that accounts for all touchpoints on the customer journey and assigns a certain amount of credit to each channel so that marketers can see the value that each touchpoint has in driving a conversion. There are several multi-touch attribution models available to marketers that look at user-level data, i.e., the impact that user-level events (clicks, impressions) have on the overall goal. Each of these models scales ad effectiveness differently.

1. In First Interaction Attribution also known as “First Click,” one single click/interaction receives 100 percent of the credit. First Interaction attributes the entire conversion to the business’s first interaction with the customer.

2. Last Interaction Attribution also known as “last click” or “last touch.” As the name implies, this model attributes 100 percent of a lead’s conversion to the last interaction the company had with them before they converted.

3. Attribution in Linear Form the credits for a conversion equally between all of the interactions the customer had with the business using a Linear attribution model.

4. Time decay attribution - Is similar to linear attribution in that the value is spread out over multiple events. However, unlike linear attribution, the Time Decay model considers when each touchpoint occurred, Interactions that occur closer to the time of purchase are given more weight. The first interaction is given less weight, while the last interaction is given the most weight. Simply put,the credits deteriorate over time, in here we undertook that the channel credit decays by half every seven days.

5. Markov chains - A probabilistic model of the buyer’s journey represented as a graph, with nodes representing different channels/touchpoints and connecting lines representing observed transitions between them. The number of times buyers have switched between states is converted into a probability, which can then be used to assess the significance of each.

Based solely on the current interaction, Markov chains calculate the probability of one interaction leading to another. This model requires pathing of the data, which demonstrates the order in which a customer encountered various marketing channels and whether the journey resulted in a conversion.

The Removal Effect allows us to quantify the contribution of any individual channel to conversions. This is accomplished by removing the channel entirely from the graph and observing the effect on conversions. The greater the impact, the greater the value assigned to the channel. It is necessary to repeat this process for each channel in order to assess the impact on conversions and, ultimately, quantify the value of each channel. To compute the Removal Effect, we first compute the probability of all converting paths.

This is a digital marketing data.The data set contains over 586,000 marketing touch-points from July (2018)

Features description:

• The number of conversions attributes to each channel for each model is shown in the bar graph. We learn more about the relative significance of various marketing channels by examining the graph and, in particular the markov model in comparison to the other approaches.

• Facebook and paid search are the most essential channels for generating conversions in the first touch, last touch, and linear touch models, while Instagram and online display are the least significant. However, the Markov Model indicates that Instagram is significantly more crucial to our conversions than the Heuristics attribution models.