The impact of recommendation algorithms on the UK’s music industry: creators’ survey results
Published 9 February 2023
© Crown copyright 2023
This publication is licensed under the terms of the Open Government Licence v3.0 except where otherwise stated. To view this licence, visit nationalarchives.gov.uk/doc/open-government-licence/version/3 or write to the Information Policy Team, The National Archives, Kew, London TW9 4DU, or email: psi@nationalarchives.gov.uk.
Where we have identified any third party copyright information you will need to obtain permission from the copyright holders concerned.
This publication is available at https://www.gov.uk/government/publications/research-into-the-impact-of-streaming-services-algorithms-on-music-consumption/the-impact-of-recommendation-algorithms-on-the-uks-music-industry-creators-survey-results
About this document
As part of the government’s response to the Digital, Culture, Media and Sport Committee’s Second Report of Session 2021-22, Economics of music streaming (HC 50), the CDEI was asked to conduct research into the recommendation algorithms used by streaming services.[footnote 1]
During this research, the CDEI surveyed creators to explore their understanding of, and feelings about, these technologies. This document pulls together the information and questions given to respondents, and the results of the survey.
Some additional text has been added to guide users of this document. Some of our questions allowed respondents to provide free form answers, we have decided not to include these here as a means of ensuring the anonymity of respondents.
This survey was used, alongside consumer polling, academic research, and stakeholder interviews to inform the CDEI’s research in this space. The full results of this research were published in a report that can be found here.
We are grateful for all those that took the time to respond to this survey.
Section one: context given to respondents
The future of music streaming: survey for creators on recommendation algorithms
The Centre for Data Ethics and Innovation (CDEI) is researching the impacts that algorithmic recommendations on music streaming services are having on the music ecosystem. This survey is looking to hear directly from creators and will be crucial in informing any future recommendations and interventions in this space. More information on how we use your data can be found in our privacy notice.
This survey will take under 10 minutes to complete.
Section two: questions on recommendation systems
Recommendation systems
On most music streaming platforms, consumers receive recommendations of new songs, artists, and playlists based on their past listening habits and other pieces of data. A number of statements about how these recommendations are made follow. Please select the statements that you think are correct (please select all that apply).
| Statement | Number of respondents (102) |
|---|---|
| Algorithms make specific recommendations to individual users. | 84 (82.4%) |
| Humans make specific recommendations to individual users. | 14 (13.7%) |
| Playlists are curated by algorithms and recommended to users. | 72 (70.6%) |
| Playlists are curated by editors and recommended to users. | 58 (56.9%) |
| Playlists are curated by editors, but the order that a user hears the songs on them is recommended by an algorithm. | 32 (31.4%) |
| I had never considered how these recommendations are made. | 4 (3.9%) |
| I am unaware of this feature on music streaming platforms. | 3 (2.9%) |
| Don’t know. | 5 (4.9%) |
Please tell us how much you agree/disagree with the following statements from 1 (strong disagree) to 5 (strongly agree):
| Statement | 1 | 2 | 3 | 4 | 5 |
|---|---|---|---|---|---|
| I feel that recommendations on music streaming platforms have enabled new consumers to discover my music. | 23 (22.5%) | 16 (15.7%) | 29 (28.4%) | 21 (20.6&) | 13 (12.7%) |
| I think that streaming services give me sufficient information about how recommendations are made to consumers. | 60 (58.8%) | 25 (24.5%) | 12 (11.8%) | 2 (2.0%) | 3 (2.9%) |
| I would like more detail as to how recommendations are made on music streaming platforms. | 2 (2.0%) | 1 (1.0%) | 8 (7.8%) | 20 (19.6%) | 71 (69.6%) |
| I have changed the way I make music to try and increase the likelihood of it being recommended. | 44 (43.1%) | 20 (19.6%) | 21 (20.6%) | 11 (10.8%) | 6 (5.9%) |
| I (or my representatives) have changed the way I release music to try and increase the likelihood of it being recommended. | 20 (19.8%) | 6 (5.9%) | 21 (20.8%) | 30 (29.7%) | 24 (23.8%) |
Which of the following streaming platforms have you or a representative uploaded your music to? Please select all that apply:
| Platform | Number of respondents (102) |
|---|---|
| Spotify | 90 (88.2%) |
| Apple Music | 88 (86.3%) |
| Tidal | 74 (72.5%) |
| Amazon | 82 (80.4%) |
| YouTube | 86 (84.3%) |
| IDAGIO | 16 (15.7%) |
| Napster | 41 (40.2%) |
| Qobuz | 24 (23.5%) |
| SoundCloud Go | 36 (35.5%) |
| Don’t know | 5 (4.9%) |
| I/my representative has never uploaded my music to a streaming platform. | 6 (5.9%) |
If you or a representative has uploaded your music to one of these platforms, how well do you think the metadata options on the service represent the music being uploaded from 1 (Not well at all) to 5 (Extremely well)?
| 1 | 2 | 3 | 4 | 5 |
|---|---|---|---|---|
| 21 (22.1%) | 31 (32.6%) | 35 (36.8%) | 7 (7.4%) | 1 (1.1%) |
Section three: questions about algorithmic bias
Algorithmic bias
When recommendations are made to consumers, often the recommendation is based on data about them and their listening habits. There may be a risk that certain music, artists, or playlists are prioritised over others, and that the prioritisation of these artists is skewed or unfair against certain groups, for example by gender or genre.
Four statements relating to bias in recommendations follow. On a scale of 1 (Strongly Disagree) -5 (Strongly Agree), please state how much you agree/disagree with these.
| Statement | 1 | 2 | 3 | 4 | 5 |
|---|---|---|---|---|---|
| I am concerned about bias in recommendation algorithms leading to tracks from certain artists, or certain labels, being prioritised over other creators. | 0 (0.0%) | 1 (1.0%) | 10 (9.8%0 | 18 (17.6%) | 73 (71.6%) |
| I am concerned about bias in recommendation algorithms leading to music from certain demographic groups (e.g. different ethnic groups, genders) being prioritised over others. | 2 (2.0%) | 4 (3.9%) | 27 (26.%) | 21 (20.6%) | 48 (47.1%) |
| I am concerned about bias in recommendation algorithms leading to music from certain genres being prioritised over others. | 0 (0.0%) | 1 (1.0%) | 14 (13.7%) | 28 (27.5%) | 59 (57.8%) |
| If I felt as though I was being treated unfairly by platform recommendations, I would know who to speak to. | 65 (63.7%) | 14 (13.7%) | 10 (9.8%) | 4 (3.9%) | 9 (8.8%) |
Section four: questions about further research
Further research
A better understanding of how algorithmic recommendations are affecting music consumption could be gained through voluntary submissions by creators of data about themselves, including demographic data such as gender identity, age, and ethnic group. One proposed way of doing this would be through a trusted third party data holder called a data intermediary. This data intermediary would hold this data and enable trusted researchers to use it to test for bias in their algorithmic recommendations.
On a scale of 1 (Strongly Disagree) - 5 (Strongly Agree), how much do you agree/disagree with the following statement: I would be interested in providing my data to a data intermediary in order to facilitate research into bias.
| 1 | 2 | 3 | 4 | 5 |
|---|---|---|---|---|
| 11 (10.8%) | 3 (2.9%) | 29 (28.4%) | 26 (25.5%) | 33 (32.4%) |
On a scale of 1 (Strongly Disagree) - 5 (Strongly Agree), how much do you agree/disagree that additional research into the impact that algorithmically-generated recommendations are having would be valuable?
| 1 | 2 | 3 | 4 | 5 |
|---|---|---|---|---|
| 2 (2.0%) | 2 (2.0%) | 9 (8.8%) | 17 (16.7%) | 72 (70.6%) |
Section five: self-identification questions
About you as a creator
Below are some questions about you as a creator. Answering these questions will help us to understand any differences in opinion on music streaming algorithms between different types of creator, genre, record company and publisher, and therefore significantly contribute to our research into meaningful transparency and potential bias.
| Description | Number of respondents |
|---|---|
| Songwriter/Composer | 50 (49.0%) |
| Vocalist | 5 (4.9%) |
| Instrumentalist | 14 (13.7%) |
| Engineer | 2 (2.0%) |
| Producer | 20 (19.6%) |
| DJ | 0 |
| Rapper/MC | 1 (1.0%) |
| Other | 10 (9.8%) |
If you define yourself as a performer (i.e. vocalist or an instrumentalist), would you define yourself as a featured artist (e.g. a primary artist or guest/secondary performer) or a non-featured artist (e.g. a session musician)?
| A featured artist | A non-featured artist |
|---|---|
| 73 (92.4%) | 6 (7.6%) |
What musical genre(s) do you most closely identify with as a creator? Please select up to three.
| Genre | Respondents (102) |
|---|---|
| Afrobeat | 2 (2.0%) |
| Ambient | 4 (3.9%) |
| Classical (Western) | 20 (19.6%) |
| Classical (Non-Western) | 3 (2.9%) |
| Desi | 0 |
| East Asian | 0 |
| Electronic Dance Music | 25 (24.5%) |
| Folk | 17 (16.7%) |
| Grime | 0 |
| Hip Hop | 8 (7.8%) |
| Indie | 28 (27.5%) |
| Jazz | 11 (10.8%) |
| Latin | 3 (2.9%) |
| Metal | 4 (3.9%) |
| Pop | 39 (38.2%) |
| Punk | 6 (5.9%) |
| R&B | 3 (2.9%) |
| Reggae | (2.0%) |
| Rock | 39 (38.2%) |
| Soul | 7 (6.9%) |
| Traditional (English, Irish, Scottish, Welsh) | 8 (7.8%) |
| Traditional (Non-UK and Ireland) | 2 (2.0%) |
| Other | 26 (25.5%) |
To What extent does music contribute to your income
| Statement | Respondents |
|---|---|
| Music is my only source of income | 39 (38.2%) |
| Music contributes significantly to my income (over half my total income) | 18 (17.6%) |
| Music contributes a small amount to my income - I mostly rely on other forms of employment/income | 30 (29.4%) |
| I currently do not earn money from music | 9 (8.8%) |
| I’d prefer not to say | 6 (5.9%) |
If you are a featured artist, which of the following describes you in relation to record companies/labels?
| Statement | Respondents (79) |
|---|---|
| I’m currently signed to a major record company/label (e.g. Universal, Sony, Warner) | 2 (2.5%) |
| I’m currently signed to an independent record company/label | 16 (20.3%) |
| I was previously signed to a major record company/label but am not anymore | 5 (6.3%) |
| I was previously signed to an independent record company/label but am not anymore | 7 (8.9%) |
| I am a self-release or DIY artist | 45 (57.0%) |
| Other | 4 (5.1%) |
If you are a songwriter/composer, which of the following describes you in relation to publishers?
| Statement | Respondents (85) |
|---|---|
| I’m currently signed to a major publisher | 4 (4.5%) |
| I’m currently signed to an independent publisher | 18 (21.2%) |
| I was previously signed to a major publisher, but am not anymore | 2 (2.4%) |
| I was previously signed to an independent publisher, but am not anymore | 9 (10.6%) |
| I am self-published | 45 (57.0%) |
| Other | 7 (8.4%) |
Section six: demographic questions
Demographic questions
Finally, it would be extremely helpful for our research if you could answer the following four demographic questions, so that we can analyse our results by age, region, gender identity, and ethnic group, and understand if there are any significant differences between groups. These questions are all required, but each has a “prefer not to say” option.
What is your age group?
| Age | Respondents (102) |
|---|---|
| 18-24 | 5 (4.9%) |
| 25-34 | 15 (14.7%) |
| 35-49 | 50 (49.0%) |
| 50-59 | 5 (4.9%) |
| 60-60 | 13 (12.7%) |
| 65+ | 8 (7.8%) |
| Prefer not to say | 6 (5.9%) |
In which UK region is your main base?
| Region | Respondents (102) |
|---|---|
| Northern Ireland | 3 (2.9%) |
| Scotland | 5 (4.9%) |
| North-West | 4 (3.9%) |
| North-East | 3 (2.9%) |
| Yorkshire & Humberside | 5 (4.9%) |
| Wales | 4 (3.9%) |
| West Midlands | 4 (3.9%) |
| East Midlands | 2 (2.0%) |
| South-West | 7 (6.9%) |
| South-East | 19 (18.6%) |
| Eastern | 4 (3.9%) |
| London | 30 (29.4%) |
| Channel Islands | 1 (1.0%) |
| I’m not based in the UK | 11 (10.8%) |
| Prefer not to say | 0 |
Which of the following best describes how you think of your gender identity?
| Gender Identity | Respondents (102) |
|---|---|
| Female | 20 (19.6%) |
| Male | 75 (73.5%) |
| Non-binary | 2 (2.0%) |
| Prefer to self-describe | 0 |
| Prefer not to say | 5 (4.9%) |
How would you describe your ethnic group? Choose one option that best describes your ethnic group or background.
| Ethnic Group | Respondents (102) |
|---|---|
| White - English/Welsh/Scottish/Northern Irish/British | 66 (64.7%) |
| White - Irish | 9 (8.8%) |
| White - Gypsy or Irish Traveller | 0 |
| Any other White background | 14 (13.7%) |
| Mixed/Multiple Ethnic Groups - White and Black Caribbean | 1 (1.0%) |
| Mixed/Multiple Ethnic Groups - White and Black African | 1 (1.0%) |
| Mixed/Multiple Ethnic Groups - White and Asian | 0 |
| Any other Mixed/Multiple ethnic background | 1 (1.0%) |
| Asian/Asian British - Indian | 0 |
| Asian/Asian British - Pakistani | 0 |
| Asian/Asian British - Bangladeshi | 0 |
| Asian/Asian British - Chinese | 0 |
| Any other Asian background | 0 |
| Black/African/Caribbean/Black British - African | 1 (1.0%) |
| Black/African/Caribbean/Black British - Caribbean | 0 |
| Any other Black/African/Caribbean background | 1 (1.0%) |
| Arab | 0 |
| Prefer to self-describe | 4 (3.9%) |
| Prefer not to say | 4 (3.9%) |
-
House of Commons Digital, Culture, Media and Sport Committee, Economics of music streaming: Government and Competition and Markets Authority Responses to Committee’s Second Report, 2021. ↩