Muse is a music discovery service that uses algorithms and user preferences to recommend songs, albums and artists to music lovers looking to expand their music libraries. The app will also use machine learning, made possible by recent updates to mobile operating systems, to learn from the ways in which users consume music on the service.
The app will present songs to the user and then list the streaming services the song is available on. The user will select the service and be redirected. Rather than play the song natively in the app, the terms of the two largest streaming services (Apple Music and Spotify) stipulate that their APIs cannot be used in apps that are monetised, as Muse is. (See, Legal Considerations, Native Playback)
One particular cause for concern may be that the user will find no reason to use Muse if they are eventually redirected to the streaming service anyway but the algorithm is the key to ensuring relevant music is recommended and the algorithm can only be found inside the Muse app. The app will be developed to provide as seamless an experience as possible, minimising the risk that the user will avoid Muse because of this.
MuseSense is the name of the algorithm created specifically for Muse will be integrated into the app. Users will download the app and register an account. From here, users are guided through the setup process and will be asked some short questions about the music they like. These questions will be broken down into categories similar to the following list;
Ideally, the algorithms were to then take this information, analyse it and match it to a music database built from using a Spotify API however Spotify's Terms Of Service prohibit this. More detail on this can be found in ‘Legal Considerations'. For this reason, custom algorithms are necessary but it is important to consider that these are both costly and time-consuming.
Muse will not alter any part of the user's Apple Music account, other than providing an affiliate link to sign up for an account and will not add songs to a playlist but will link users to that song in the Apple Music library where they can then add this to their library or any playlist they choose - this is due to restrictions on monetised apps using Apple and Spotify APIs (See, Legal Considerations, Native Playback).
Additionally, the app will process music matching off-device to take advantage of server power.
Due to the need to remain current and be able to utilise the most efficient and effective APIs to maintain the best user experience possible - Muse will be launched on a single platform and will be expanded over time if necessary. I have decided that the launch platform will be iOS due to high rates of install. According to Apple internal statistics, over 90% of its users are using either of the two latest iOS versions. Android, however, has much lower figures with only around 53% of users installing one of the two latest versions.
Apple's large smartphone market share has allowed the organisation to create a unified OS (iPhone, iPad and iPod Touch all use iOS) so app design/development and branding can all be carried over easily to multiple devices. In addition, Apple also offers Apple Music to Android users and so in the future, it would be simpler to carry over the same features - the infrastructure of the service should be built first.
Artists or their teams can register for an artist profile. This will be publicly available, similar to Apple's Connect platform and can be used to promote the artist, their performances or merchandise. The artist will be asked to assign themselves to the categories that will be used to match users to music. They can do this for their music as a whole, by album or even by song. This is particularly useful for artists who like to experiment or do not have one distinct style of music.
Muse will be built as an app solely to discover music and algorithms will be wide-ranging to reflect this. For example, the Muse algorithm will take into account not only the basic data such as artist and album when recommending a song but will also, for example, analyse the tempo (beats per minute) of songs among other more detailed characteristics. This extra analysis will help provide further data used to enhance MuseSense over time to take into account whether a user likes particular sub-genres or a particular mood.
Currently, competitors offer specific human-curated playlists but the recommendation features tend to focus solely on artists similar to those the user has listened to or artists most popular on the service, even when recommending a genre. In addition to the typical considerations a music discovery app will make, MuseSense will take into account the differences between artists within the same genre such as styles, chord progressions (to ascertain what mood the song could be categorised as) and eventually investment in the algorithms will provide Muse the time and justification to look at aspects such as loudness, ‘danceability' and energy. These attributes are similar to those used by Spotify's API but will be more fine-tuned for MuseSense to ensure the service provides the experience the user wants.
Specifically, Muse aims to put user experience over profit with a view to giving customers such a good service that they continue to use the service. The larger services will generally promote artists with more fans such as Adele, for example. This is useful if an artist already has a large following but for artists who are not as popular or have yet to reach a certain success. Muse will provide an opportunity to reach listeners that are likely to specifically enjoy whichever type of music listed on their artist profile. This way, artist profitability is both not a factor in the success of an artist or their music whilst they are simultaneously being promoted and profiting from the increased exposure. This is one way in which sign-ups can be encouraged and promoted to new artists. Larger artists who are keen to embrace new technology might also sign up to the services as the exposure could still benefit them however the technology will ensure they are not recommended any more than any other artist using the service.
Soundcloud is another similar service but is smaller in terms of brand recognition. According to a Wired article, the service has 175 million users monthly. Interestingly, SoundCloud has been most popular with those wanting to release or consume rap/hip-hop music or electronic music. This has proven popular as evidenced by the number of users however SoundCloud is a free service and additionally, according to the same article, has to pay up to 70% of profits to the owner of that song. For this reason, they are limited. This is important to Muse as it is evidence of the danger of expanding into hosting/streaming music.
2. Business Strategy
Essentially, MuseSense is the product and the app is a facilitator for commercialising the algorithm. This is important to consider as this could eventually become the main source of revenue and Muse should be prepared to either license or sell the algorithm.
Advertisements will be built into the app - initially, Muse will utilise a third-party SDK for efficiency reasons. The ads will be persistent but take up a small section of the screen to ensure uninterrupted use of the app by users. The revenue from these adverts will provide funds to keep the app running whilst partnerships are built to ensure long-term higher value income.
The chosen provider is AdMob by Google. It is hard to accurately forecast for an advertising framework as there are no upfront costs but the provider will keep a percentage of revenue generated from ads within Muse and the income generated is similarly hard to predict due to the cost-per-click structure on which advertising frameworks run.
AdMob provides an analytics feature which can be useful for maximising the revenue generated in the early stages of the business. This initial funding will be key to keeping a steady cash flow.
AdMob also allows Muse the opportunity to target adverts - this could be particularly useful for a time when alternative sources of funding are found and Muse can be more specific about which adverts it wants to show.
In addition to this, affiliate links will be used. Affiliate links are a form of profit-sharing that allows people to provide specialised links to content and the provider/owner of the content will pay a share of the revenue generated from the exposure. Each service has different rates and systems;
Apple Music pays affiliates per account sign-up. The amount depends on the type of account as Apple Music offers three types currently (Single, Family and Student Memberships).
The prices of these and the amount Muse will receive, pre-tax, from each signup are as follows;
Included in the table are calculations for how much Muse will receive per 100 signups and could be useful for revenue forecasts.
Artists can pay to have their music promoted but they will only be promoted to people that like the category they fit into - making the cost of advertising lower and more suitable for smaller artists trying to gain some exposure. The nature of targeted advertising also makes the experience better for the user - suggesting only songs that they are likely to be interested in. This differs from competitors in that the generally advertise, for example, Drake to all of their users despite the possibility that some of them will not be interested in rap music or Drake as an artist.
In future, it may be necessary or beneficial to form strategic partnerships. Apple or Spotify, for example, could build in our algorithms to their service - at which point Muse would become an official partner. Apple currently does this in the form of Shazam, a music recognition service. Siri is the digital assistant for Apple devices and when asked will listen to the song being played and try to match it to the Shazam database. Interestingly, Apple recently confirmed it had acquired Shazam. This could be interpreted as proof that an eventual acquisition by Apple or Spotify is not impossible.
According to an IFPI report, a recording industry body, music streaming grew by 60.4% whilst downloads (songs bought digitally on stores such as iTunes) fell by 20.5%. This boom in streaming marks a milestone, the first year to include figures following Apple's expansion into the streaming market with its Apple Music offering which only serves as a sign that the streaming market will continue to grow. To tap into this $7.8 billion market, Muse will be introduced as a companion app to streaming services to help users find the most relevant music.
Muse is first and foremost a music discovery and recommendation service. The other services have recommendation features but have had to heavily market human-curated aspects of their business as the algorithms' choices have proven unpopular with users of the services. Slant, a community-run recommendation website, ran a poll for its users to vote on whether Apple Music was a better choice for streaming music - SoundCloud won by a large margin. The rationale was
“Explore tab on SoundCloud shows trending music that can be filtered based on genre/tag. There's a constant stream of new music.”
This is interesting considering that Apple is a much larger company and SoundCloud does not have the financial backing that Apple Music does. It is of note that the sample size was extremely small, a total of 85 votes however, if these results are indicative of the wider customer base they are key to illustrating what consumers look for in a streaming service.
As stated many times, Apple Music and Spotify are the two most popular streaming services. The two are similar in that they both offer two distinct tiers - one free and one paid. The difference is, however, that Spotify's free tier is ad-funded and Apple Music provides a trial that only allows users up to three months before they are required to pay for an account however there are absolutely no adverts anywhere in Apple Music. These two approaches have different benefits but as Muse will not take direct payment from users, the strategies do not apply.
Apple has also focused on exclusivity as a marketing strategy - Drake released an album on the service earlier than anywhere else and Apple maintained an exclusivity window.
The nature of the services is also different; whilst the two services do have recommendation features, the main focus of the services is to host and stream music. The features are relatively well hidden. Muse will not have any other focus than music discovery and recommendation and so the right marketing plan could ensure the long-term success of the service.
Bandcamp is the most direct competitor but the focus of this website is more to fund and sell music. Muse only refers listeners and will not provide any money exchange facility whatsoever. A partnership with Bandcamp could eventually be possible though, enabling the purchase of merchandise in exchange for encouraging artists signed up to Bandcamp to also register for Muse.
Soundcloud, also mentioned above, is a smaller service but it is used by up and coming artists, particularly rap artists. In this respect, it is a competitor. To manage this, Muse could focus on incentivising SoundCloud artists to move over to the platform or at least to run the two channels alongside one another. SoundCloud also has a premium offering but Muse should avoid this as the message of the service is about providing music to all and I would want to avoid creating classes of customer.
Tidal is a streaming service with the benefit of having some of the biggest musicians as its owners. These owners all own stakes in the business in exchange for exclusivity and expertise. Beyoncé released Lemonade, the most popular album of the year on Tidal. MIDiA research in December 2016 found that in the UK, Australia, America and Canada, only 2.7% of its users used the platform weekly, indicating a need to engage its subscribers more. The same research also asserts that Tidal could have just 1% of the global subscription market, though Tidal says that figure is inaccurate as it has 4 million users versus the 1 million active users the research considers.
Amazon offers streaming services; Prime Music and Amazon Music Unlimited. These two services rely on the same library but some songs such as newer songs or albums which the artist prefers to remain paid to guarantee better royalties are only available on Music Unlimited. This is an interesting strategy taking the best of Apple Music and Spotify and combining them but both platforms have a relatively small market share. Nevertheless, Amazon's offerings will still be integrated with the app owing to the relatively high commission rates.
There are three particular markets Muse will target initially. These are;
18-25 years, young adults
According to Statista report, 96% of people ages 16-24 own a smartphone.
Additionally, a LOOP study found that 51% of 15-19-year-olds listened to streaming music versus 24% of users of all age groups, evidence that a larger proportion of young people use streaming than any other age group. The same report corroborates the Statista report, finding that this age group spent 41% of the total listening time using a connected device such as a smartphone. This research highlights the huge potential for growth in the market aimed at young people and their reliance on a smartphone to get the latest music. The convenience of listening anywhere means they invest more time in music discovery and consumption.
Music Industry Employees (A&R)
A&R (Artists and Repertoire) is a role in the music industry that involves raising the profile and general management of an artist. This is a particularly good area to focus on as artist sign-up is going to be key to maintaining a steady and fresh stream of new music to recommend to Muse users. Building relationships and involving A&R in Muse will both introduce their clients to Muse and will hopefully introduce the executives to new artists they might find on the service. A&R executives can also help broker artist advertisement deals to ensure a fair price for all parties.
“From an A&R perspective, the goal has become finding artists that have a great following–not artists that just have talent.” REF(http://www.thembj.org/2015/03/ar-in-a-digital-age/) Raising the profile of the artists signed up to Muse will also raise the profile of Muse, leading to more registered users.
Fitness is a huge industry, with one report finding that 1 in 7 people are registered with a gym, a rise of .6% between 2016 and 2017. Additionally, the total value of the market is now £4.7 billion, up 6.3% on 2016. Tapping into this market, as Nike has done with the Apple Watch Nike Edition, could be profitable. Building a relationship and utilising Muse features like BPM-matching could allow users to make running playlists. This information could be used to build partnerships with equipment manufacturers or gyms to provide music solutions for those relying on music to enhance their activity or mood when exercising.
3. Legal Considerations
Copyright laws prohibit downloads of any materials without permission from or payment to the rights holders of content. As Muse does not store any music and rather redirects users to stream from the service of their choice, no legal protection is required. As importing the Spotify API is against the ToU, songs matched from Spotify will be played in the app via a built-in redirect function.
“Streaming SDAs. Use the Spotify Platform to develop and distribute non-commercial Streaming SDAs that comply with the Branding Guidelines (i) for private personal use (ii) on Approved Devices (iii) by Spotify users who are subscribed to the Premium Service (as defined in The Spotify Terms and Conditions of Use).” 
To avoid any legal disputes, Muse will instead rely on a first-party, Muse-owned algorithm. To prevent competitors from copying/altering the algorithm, developers will be hired as employees and as such will be prohibited from sharing any information on the algorithm that would reduce competitive advantage.
To protect the secrecy of the algorithm, the logical step would be to patent it however under EU Law , software is not patentable;
Article 52(2)(c) “schemes, rules and methods for performing mental acts, playing games or doing business, and programs for computers;”
For this reason, it is advisable to implement a practice of Non-Disclosure within the organisation and its contractors.
Muse will rely on redirection to the streaming services the user chooses. This is because Muse is a commercial app with the sole purpose of using these platforms to recommend music and does not perform any functions reasonably considered separate from the streaming services. The two largest platforms, Apple and Spotify do not allow the use of their APIs for commercial services but affiliate links are permitted and so redirection is the most cost-effective workaround. I have decided that as these are the two largest services, it would be unnecessary to integrate any other services initially. This may change, however when the resources are available.
One thing Muse may consider though is to integrate the 30-second previews offered by iTunes as these can be used in commercial apps.
Below are the specific sections and guidelines from Apple and Spotify.
This stipulation is made in the Apple App Store Guidelines.
4.5.2 Apple Music.
“your app may not require payment or indirectly monetize access to the Apple Music service (e.g. in-app purchase, advertising, requesting user info, etc.).”
“Streaming SDAs. Use the Spotify Platform to develop and distribute non-commercial Streaming SDAs that comply with the Branding Guidelines (i) for private personal use (ii) on Approved Devices (iii) by Spotify users who are subscribed to the Premium Service (as defined in The Spotify Terms and Conditions of Use).”
4. Ethical Considerations
All other identifiable and traceable data, such as user preferences, will be anonymised on-device using tokenisation technology. All data leaving the device will be run through a separate algorithm which can be built by Muse that identifies this type of information and removes it from the data being passed to the matching servers.
Customer data that has to be kept such as login data and in-app activity will be stored on Muse servers but will also be tokenised. This is to protect customer data and minimise the chance of the information being stolen and used for purposes of fraud and theft.
5. Security Considerations
As with any online service, data integrity is paramount. This is especially important from a PR perspective as seen with the Uber data hack. Attempts to cover the hacker resulted in a backlash from customers and the general public. For this reason, it is important that Muse protects customer data - this protection is also important in terms of Muse brand image; user experience over profit.
As stated above, the data stored about customers will be completely anonymous. This means that should anyone wrongly or unlawfully gain access to the Muse server, the data cannot be used to identify the user.
The internal systems however need to withstand attacks to a higher degree. They will contain the algorithm in addition to information on payments from affiliate links and other operating expenses. In addition to being stored separately from other information, firewalls and strong authentication will be applied to protect the data.
...(download the rest of the essay above)