Marketing Engineering - Assignment #2 Marketing Engineering

1. Estimate the parameters b1, b2, b3, b4 b1 | 2.119 | b2 | 0.405 | b3 | -0.460 | b4 | -0.744 |

2. Interpret the b3 estimate based on its magnitude and sign. Do you think streaming service and download service are complementary or substitute to each other?

If we consume complementary goods together, utility for consuming tends to be larger than sum of utilities for consuming each goods respectively. For example, if we consume coke and chicken together, our utility will be bigger than sum of utilities from consuming coke and chicken respectively. On the other hand, if we consume substitute goods, utility tends to be smaller than sum of utilities for respective consumption. b3 is in the equation about U3 which is utility for bundle of streaming and downloads. So we can consider b3 as complementary or substitute quotient.
In this model, b3 was estimated as negative number(-0.460). It means that U3 (utility for consuming ‘streaming service’ and ‘download service’ together) is smaller than sum of U2 and U3 (sum of utilities for consuming ‘streaming only service’ and ‘download only service’ respectively) by 0.460 when we put aside the price. So I think ‘streaming service’ and ‘download service’ are substitute to each other.

3. The data are from ourselves. Do you think the estimation resulting from #1 makes sense? (i.e., are the results somehow consistent with our preference?). Please briefly explain
When we compare the V1, V2, V3 (in which we exclude the epsilon 1, 2, 3 from U1, U2, U3), we can see that V1 is always larger than V2, V3. It means that epsilon excluded utility for streaming only service is always larger than that for download-only service or streaming+download sevice. So model says that it is rational to choose streaming only service. But as we can see in the data, our choices were diverse (even though number of choices on streaming-only service is biggest as 274). So I think estimation results from #1 is not consistent with our preference