## Solution

Paragraphs highlighted in green indicate diagrams or tables that shall be shared in the “Case exhibits” section.

Paragraphs highlighted in blue shall be verbally communicated to the interviewee.

Paragraphs highlighted in orange indicate hints for you on how to guide the interviewee through the case.

**Question 1: How would you structure this problem?**

Do not share any data points with the interviewee, yet. Prompt the interviewee to ignore business usage (e.g. offices, clubs, restaurants, etc.) and only think of consumer use.

**Proposed solution**

**Number of people in this world**
- Number of people per continent

**Number of people on the platform**
- Adoption rates by continent
- Number of people on the platform = adoption rates * number of people per continent

**Hours of listening per week**
- Segment hours of listening by weekday and weekend based on purpose
- Weekday (other structures are possible)
- Listening while commuting
- Listening while at work
- Listening in spare time

- Weekend (other structures are possible)

- Average hours of listening per week = 5 * weekdays + 2 * weekends
- Total hours of listening per week = average hours of listening per week * number of people on the platform

**Hours of listening per year**
- Total hours of listening per year = total hours of listening per week * 52

**Question 2: Using your structure, can you estimate the number of hours listened to Spotify on average per month?**

Share that the global population is 7.5 billion and that the interviewee may assume a 20% adoption rate globally.

Share Table 1 with the interviewee.

**Proposed solution**

**Number of people on the platform**
**Hours of listening per week**
- Weekday = 1 +2 +1 = 4 hours per day
- Weekend = 1 hour per day
- Average hours per week = 4 * 5 + 1 * 2 = 22 hours per week
- Total hours per week = 22 hours per week * 1.5 bn users = 33 bn hours per week

**Hours of listening per year**
- Total hours of listening per year = 33bn * 52 = 1.716 bn

**Question 3: What variables is your analysis most sensitive to?**

Tell the interviewee to think of the problem conceptually and to not do any calculations. The interviewee should provide clear reasoning.

**Proposed solution**

- The analysis is most sensitive to the % adoption of people on the platform; a 1 percent point increase would skew the analysis by ±5%.
- The analysis is also sensitive to hours assumed; the interviewee should point out that the analysis is more sensitive to assumptions in weekdays than weekends.

**Question 4: If your friend wants to earn €2000 per month from royalties on this platform, how many streams does he need per year?**

Share the following data after the interviewee asks about costs and revenues:

- He/she can ignore any costs or investment. Additionally, there are no transaction or intermediary fees.
- The streaming platform pays €4 per 1000 streams (assume that all plays are qualified streams).
- The average length of a song is 2.5 minutes.

**Proposed solution**

In order to earn €2000, ignoring any costs, your friend needs to have €2000 / €4 * 1000 = 500.000 streams per month (or 6 million per year). An excellent candidate would touch on % of total streams, as a sanity check. Total length = 6 million * 2.5 = 15 million minutes per year (or 25.000 hours), which is ±0%.

**Question 5: Is it realistic for an average artist to achieve this?**

The interviewee should not perform any calculations but touch on the long-tail concept (distribution of plays on streaming platforms). Push the interviewee to think about the distribution.

Share Exhibit 1 after the interviewee talks about distribution and ask the interviewee to draw some inferences (again, no calculations).

**Proposed solution**

- The average number is highly skewed due to the long-tail concept.
- To outperform the average (have more plays/streams than average), one needs to be amongst the 80-90 percentile due to the long-tail phenomenon.

**Question 6: If your friend asks you about the conclusion you made, what would you tell him?**

**Proposed solution**

My analysis shows that people listen to the number one platform for 1.716 bn hours per year, which is 1.144 hours per person per year on average. Even though that number is high, to earn the mentioned salary you need people to listen to your music for 25.000 hours per year, which is negligible as a %. It is important to point out that 80-90% of all streams are on the same artists; this is called the long-tail effect. That means that the average is highly skewed and that obtaining 25.000 hours per year is a potential challenge. It requires further research into the average number of plays per track.