Business Analytics in Practice

Presented as a webinar in Ultimatern event held by Universitas Multimedia Nusantara on April 24, 2021. This session is targeted at undergraduate students to give more illustrations of real-world cases before starting internship programmes in business analytics related role.

Transcript

1. Business Analytics in Practice Elvyna Tunggawan April 2021
2. How do I get here? Sep 2012 - June 2016 Undergraduate study in Information Systems June 2016 - Feb 2020 Data Scientist March - Nov 2020 Master of Professional Studies in Data Science Dec 2020 - present Data Scientist
3. Outline • What is business analytics? • Why is it needed? • Case studies
4. What is business analytics?
5. Business analytics is the process of using quantitative methods to derive meaning from data in order to make informed business decisions. Source: Business Analytics: What It Is & Why It's Important | HBS Online
6. The University of Auckland Nanyang Technological University
7. The University of Hong Kong Massachusetts Institute of Technology
8. The heart of business analytics is a strong analytical thinking
9. You’ve applied that as well • Deciding which smartphone to buy • Planning out travel itineraries • Choosing a major on the university • … and many more
10. Companies also have similar questions • Marketing: ◦ Which marketing channel should we focus on? ◦ Will this marketing campaign increase our sales? • Business / product development: ◦ Will this kind of business attract a new customer segment? ◦ Should we provide more payment methods? • … and many more
11. Companies without analytics be like:
12. Four types of business analytics (Gartner, 2012)
13. First thing first ...
14. “No worries, we have the data”
15. Source: paulkukk.com Beware of the data silo
16. Source: bi-insider.com Handling various data types
17. Source: Dilbert Comic Strip
18. Well, let’s assume the data are available and ready to use
19. Companies usually start measuring numbers and produce tons of reports
20. Source: • Dashboard examples: The good, the bad and the ugly • 10 Dashboard Design Errors And How To Avoid Them
21. Source: Dilbert Comic Strip
22. … are these reports enough to answer a business question?
23. Business questions are not straightforward
24. Before diving into the data, always start with question(s) 1. Understand the end goal 2. Understand the current situation 3. Explore, generate hypothesis, and formulate the solution
25. Working at an e-commerce company
26. What is the most important metric to monitor?
27. Are these what you need? • Traffic • Time on page • Session duration • Bounce rate Beware of vanity metrics various numbers that may look satisfying, but do not answer or provide action items to the business
28. Remember how e-commerce sites work
29. General steps on e-commerce sites
30. These steps can be seen as a funnel Source: smartinsights.com
31. Evaluating marketing campaigns
32. Situation assessment So far, we frequently send promotion emails that include a banner image of a female. After conducting further user research, we figured out that most of our existing customers are females between 18 and 30 years old. Acknowledging the customers demographic, we think that using a banner image of a male might increase the click rate of the emails -- hence, increasing the potential customers in the website. How to evaluate this hypothesis? Illustration of banner images (Source: Digital Marketing Institute)
33. Why don’t we just observe the historical data? It’s not that costly -- we already have the data! However, the data may not be comparable • The campaigns were conducted at a different point of time on different group of users • Subject to bias and confounding variable
34. Randomised experiment (source: Wild About Statistics)
35. 1. What is the success criteria of the experiment? 2. Who is the subject of the experiment? 3. Do we have a particular target population? 4. How large is the required sample size to achieve a conclusion? It determines how long we should run the experiment Experiment design is important! Source: Udacity A/B Testing Course
36. Run the experiment and analyse the results!
37. Source: somewhere on Facebook
38. Statistics help us to deal with uncertainty
39. "Data is what you need to do analytics. Information is what you need to do business." - John Owen
40. Useful learning resources • Avinash Kaushik's blog • Lean Analytics Book – Use data to build a better startup faster • Factfulness: Ten Reasons We're Wrong About the World – and Why Things Are Better Than You Think by Hans Rosling • Udacity A/B Testing Course
41. Look for useful references and create a high-quality presentation Find funny memes to be included in the slides Credits: @DoraemonHariIni
42. Any questions? @vexenta linkedin.com/in/elvyna speakerdeck.com/elvyna