Summary: Digital Marketing And Analytics Week 6

Read the summary and the most important questions on Digital Marketing and Analytics Week 6

• 1 Clip 1 - Recommendation Systems

  This is a preview. There are 1 more flashcards available for chapter 1
  Show more cards here

• Recommendation engines run on a machine learning algorithm, what does it maximise? List the 3 things.

  1. Average order value; 
     
  2. Increase page-views; 
  3. Enhancing engagement.

• What is an example of how amazon increases the average order value?

  If you have bought something from amazon, they will recommend you to buy other products so you'll buy more in that purchase.

• Increasing the average order value leads to certain opportunities when you have got product catalogs. Which 2 opportunities?

  Cross-selling and up-selling, by showing relevant products that are viewed or purchased together on the homepage and category pages (finish the outfit... People also viewed...).

• How do you increase the number of page views and drive loyalty?

  By engaging visitors with personalised content.

• 1.1 2 Types of Recommendation Systems

  This is a preview. There are 5 more flashcards available for chapter 1.1
  Show more cards here

• What are the 2 types of recommendation systems?

  1. Content based filtering; 
     
  2. Collaborative filtering.

• What does the content based filtering entail?

  Identify items that are similar (similarity is based on category preferences) to the ones already "consumed" by current user and recommend similar products.

• What are the basic steps for the content based filtering (3 steps)?

  1. Pre-defined set of product characteristics: e.g. Genre, writer, actors; 
     
  2. Calculate similarity between products; 
  3. Select the products that are most similar to the consumed products.

• What are the pros of the Content Based Filtering?

  Simple and intuitive, useful even if we know nothing about the users.

• What are the cons of the Content Based Filtering? Mention the 2.

  1. Product characteristics can be hard to define
     
  2. Not model based so it will not take user similarity into account.

• What does the collaborative filtering entail?

  Matches users based on their behaviour and then recommends one product to the users depending on their similarity.