Many industries have become tied to emerging technologies

With market research, analyzing consumer trends can often mean looking at the technology driving the trend. The music industry has changed more in the last 5 years than it ever has and technology is a big reason why. Streaming has become the preferred method of music consumption for many people regardless of their age. The ease, cost effectiveness, and immediacy of streaming music has made streaming services highly valued and highly profitable.

Buyer Behavior

People do still listen to music from their personal collections and CDs, vinyls and other physical media sources are still sought after by collectors. They are also popular merchandise at live shows. Live music has always been a stable source of entertainment.

Building Awareness

With artists having more platforms on which to publish their music and gain exposure (such as Soundcloud and Bandcamp), there are more ways than ever for people who enjoy live entertainment to find acts to follow. Streaming, however, is still the preferred medium for music consumption and for finding new artists.

Music Streaming

Spotify and Pandora are perfect examples of platforms that use technology to not only make music more accessible to fans, but also introduce new music to people who may not have had the opportunity or interest in seeking it out before.

These companies, as well as Google with their Google Play music platform, Apple with iTunes and the myriad of other streaming services, collect data on users’ listening trends which can be used both for marketing purposes and to make inferences on industry trends.

Consumer Insights

One pitfall to the data being collected by many music streaming services is that very little of it is qualitative. Services collect information such as search history, how many times a song is played, whether people prefer to listen to an artist on shuffle or a radio station based on an artist, but very little of that information answers any of the “why” questions that come up so often in music and in all art.

Measuring Performance

If two artists have uplifting, danceable songs, why would one person choose either one over the other? These can be addressed using qualitative surveys, however the methodology must be chosen carefully and the end goal of the research must be planned carefully before performing such an endeavor.

Big Data

Companies such as Spotify are working harder on using the data they collect to answer deeper questions about people’s relationships to their music. Recent articles have suggested that Spotify can tell people’s moods by looking at the type of music they listen to and they context in which they listen to it (location, device, etc). Other companies such as Facebook have tried to market based on information on people’s moods, as well.

The analysis of all that data is a phenomenally complex task, but the results can garner anything from an individual’s preferences based on their mood (and vice-versa) to what acts are having the most success and, potentially, how new acts can follow suit.

Conjoint Analysis

Conjoint analysis is one quantitative method that can be helpful with determining the choice practices of consumers for something like music, but it is used across multiple industries for its versatility and power. Conjoint analysis is designed to mimic the buying process by showing respondents a series of products each with multiple attributes.

Respondents are to make decisions on their preferred product which, after several selections, would show which attribute is the most influential. Several steps must be taken to make the data as valuable as possible. The attributes chosen must be attributes that actually affect consumer decisions. 

While the key and meter are attributes of songs, they are generally not deciding factors in why a person selects a song. In the same way, engine size and horsepower are attributes for cars that could be influential for some consumers, but may not be influential for a certain type of consumer.


Knowing who your respondents are going to be and what attributes you’re trying to test is vital to getting useable data.