W2: Marketing Attribution with Markov _ by James Kinley _ Towards Data Science.pdf

Question 1

What is marketing attribution?

Answer 1

Marketing attribution measures the value of campaigns and channels that reach potential customers.

Question 2

What are the limitations of traditional attribution models?

Answer 2

Traditional models are rules-based and may not accurately distribute credit for sales events among touchpoints.

Question 3

What was Cloudera’s initial approach to marketing attribution?

Answer 3

Cloudera initially evaluated the Shapley value model but needed a solution that scaled better for numerous touchpoints.

Question 4

How do Markov attribution models work?

Answer 4

Markov models use a graph to represent buyer journeys, with nodes as touchpoints and edges as transitions, calculating transition probabilities.

Question 5

What is the Removal Effect in Markov attribution models?

Answer 5

The Removal Effect measures a campaign’s importance by the change in success rate when it is removed from the model.

Question 6

How is campaign attribution calculated in Markov models?

Answer 6

Campaign attribution is calculated using the total value, the campaign’s Removal Effect, and the sum of all Removal Effects.