1 Objectives
The project’s scoping aim is to create an open, scalable data-to-policy pipeline for European music ecosystems. This aim entails three distinct objectives, which connect to the three phases of the project execution: - O1. Map the data landscape will be accomplished in the first set of tasks (T1.1, T2.1, T3.1, T4.1) - O2. Bridge data gaps will be accomplished in the second set of tasks (T1.2, T2.2, T3.2, T4.2) - O3. Empower stakeholders to take data-driven actions will be accomplished in the third set of tasks (T1.3, T2.3, T3.3, T4.3) Each objective will be realised across four work packages targeting the four pillars of the Feasibility study for the establishment of a European Music Observatory: WP1. Economy of music in Europe; WP2. Music diversity and circulation; WP3. Music, society, and citizenship; and WP4. Innovation and future trends. The project’s innovations will be amplified through the project management and implementation work packages (WP5 & 6). The objectives are designed to advance the state of the art with regard to data collection, policymaking, and business practices in the music industry. Both the objectives and the outputs will be optimised for transferability to other cultural and creative industries. O1. MAP the data landscape. The project will identify extant and potential data sources, specify appropriate data collection methods, and develop policy-relevant indicators that capture the significance of the data at hand. Objective 1 meets the expected outcomes of providing new/improved methodologies for measuring the value of music and developing indicators to represent this value. It will be realised in T1.1, T2.1, T3.1, and T4.1, and reported in D1.1, D2.1, D3.1, and D4.1.
For readability, we start with the end-user objectives that we describe in greater detail, and then we show our intermediate objectives that will take us to the desired policy outcomes.
1.1 O3 Empower stakeholders to take data-driven actions
Empower stakeholders to take data-driven actions will be accomplished in the third set of tasks.
STATE OF THE ART – 1 paragraph each on economy of music in Europe; music diversity and circulation; music, society, and citizenship; and innovation and future trends. AMBITION – 0.5 page – “preview” of the relevant methods and impacts
The gradual transition of the music economy from the mechanical licensing and public performance model to the streaming model has two unfavourable outcomes. Streaming revenues in most market segments (independent artists, emerging national markets) have never replaced the lost mechanical licensing and public performance revenues—streaming often reduced the market shares of local stakeholders and let to an unprecedented concentration of revenues in the hands of few (and usually non-European) artists and labels. Second, this lead to an over-reliance on live music revenues, particularly in the emerging and future markets of Europe. This model turned out to be not sustainable when an external shock, Covid-19 pandemic forced live performances to halt.
Our ambition is to adjust all the existing intellectual property valuation models that were earlier recognized for music valuation to the new business models and realities of a streaming-based recorded music industry. From an economics point of view, this means adjusting valuation models to account for zero-price platforms. In music economics, the greatest amount of consumer use has a zero price, either because of an unlicensed use, or because somebody else (for example, a radio station or a freemium streaming platform) pays for the use. When the user does not pay, there is no market transaction and accounting trail—practically neither the quantity of the use or the price of the use, the two most important variables of any economics model, is available with certainty. In fact, the zero-price platform model, which is often analysed with examples of AirBnB, LinkedIn, Facebook, or Youtube, is a very important problem in current competition economics and intellectual property economics and generally thought to be originating from the music industry itself. Commercial radio, and later commercial and music television were the first pioneers of a model where users do not pay for music but listen to commercials instead. We will use statistical techniques to estimate both harmonised quantities and shadow prices on zero-price platforms. We will borrow techniques from advanced financial economics to bring price transparency into streaming.
In music policy, diversity and the maintenance of a healthy audience and market share for the local, national and European works/recordings and artists had always been an important cultural policy goal. In the centralised curation model and regulation of radio and television—where a music editor chose the music for everybody—had its policy toolkits. Until the 1990s most broadcasters were state-owned and the state as an owner set targets for them. With the increase of commercial radio, then television, local content regulation forced broadcasters to maintain enough airspace for local content. Strengthened by education policies in music appreciation they helped local music sales through discovery. Our WP Diversity develops toolkits that work in streaming, too, where most curation is taken over by autonomous (AI-driven systems), playlisting is decentralised, and often fully personalised. Touching on the state of the art in applied musicology, xxxxxx. The overlap with our economics toolkit is not coincidental: a higher visibility of independent, small-country or European repertoires in live and recorded performance, or streaming in economics terms means a higher competitiveness or higher market share. While the data-driven action may be different for a cultural policymaker than for a royalty valuator, the underlying data about use, value and market share is in many cases the same.
Our Music, society, and citizenship work package can be seen as a supporting policy toolkit for improving the economic value, the diversity and competitiveness of xxxxx. It is centred around making use of the quantitative cultural sociology model of cultural (music) participation. Measuring participation is necessarily broader than measuring consumption. Consumption is a market category, and usually entails sales transactions and an accounting trail. As most music use (at least in enjoyment hours) is free, and so are many forms of active participation (amateur playing and singing, and informal learning practices; liturgical use music), quantitative surveys must be used that can be harmonised with existing survey programs (such as Eurobarometer, EU-SILC, and AES) and also with the existing transactional databases of music (about radio/television licensing or ticket sales.)
The European music industry is dominated by freelancers, micro and small enterprises, yet it must apply technological and business innovation to business challenges posed by global data monopolies on platforms that it uses for selling music, or the Covid-19 pandemic that disrupted the live music part of the ecosystem. Our WP Innovation focuses on the neglegted aspects of metadata. The new information systems rely on metadata to put data into work and build useful information out of it. This metadata, often called ‘data about data’ is central to make NFT/blockchain or autonomous AI systems work in a reliable, trustworthy, and ethical manner. This metadata is the focus point of our innovation toolkit, because without proper metadata, modern music technology will not work for, but against the European creators. Making these innovation work requires advanced information technology. Because our Consortium embraces the standards of reproducible research standards to its full extent, including computational reproducability. We develop software tools that enable to music ecosystem of enterprises, policymakers and researchers to solve economics, diversity, educations, and societal valuation problems with our high-quality, open-source softwares and web applications.
Following the best practice on Open Policy Analysis, we have placed a slightly extended, versioned, authoritative copies of the state of the art with our objectives on the Zenodo open science repository’s Digital Music Observatory repository community.2
1.1.1 Ambition
Our Proposal is made in the spirit of scientfic reproducibility, interoperability, and transparency. We aim to make already existing industry-data, such as data held by GESAC and CISAC in their societies, distribution and in-store promotion industry metadata, public data, open, but not readily accessible (reprocessable data) and primary, standardized data collection support key business and policy goals of the music ecosystem in a way that can be reproduced.
To make sure that our results are transferable from one European country to another, or one metropolitian area to another, we will use the highest, Level 3 standards of the Open Policy Guidelines, which build on the highest level of scientific reproducability3. For example, the valuation report on Bulgarian music is contained in a clearly documented dynamic document that contains the assumptions, software code to read in the data from the data sources, the processing codes makes them standardized and interoperable, if they are not well processed at source, perform the modelling, create the visualizations, document citations, and place it in the same file with the conclusions. Our “live policy documents” with a press of a button are reading in new information, change model outputs and visualizations, update the bibliography, place the results on the Zenodo open science repository with an authoritative copy and DOI. See one live policy document in making for the industry investigation of the UK Competition & Market Authority here.4
Our WP Music Economy will show how to improve the valuations of music on zero-price platfors and bring transparency to streaming, also highlighting the trade-offs in investing into music content and its proper documentation. We will bulid on existing results in Hungary made by Artisjus to be transfered to probably the poorest national collective management society, Muzikautor in Bulgaria. Artisjus and SOZA were able to increase its revenues by about 40%-50% with improved valuation—we hope that opening up this methodology to open scientific improvements and inquiry, and making it fully interoperable with data standardization will yield monetary benefits for Bulgarian authors by the end of our project, and provide hands-on, complex policy template that is fully reproducible, even in terms of computational reproducability,i.e. the algorithms and software used to arrive to the value of music in use on Bulgaria’s radio, televions, streaming platforms and in unlicensed copying.
Similarly, in WP Diversity we will build on a detailed Feasibility Study5 created by SOZA and Reprex to measure and promote the Slovak national repertoire on streaming platforms. Setting proper diversity targets in cultural diversity policy is essential for the survival of small European labels, publishers, and artists of small-language member states. We will demonstrate with open database standards, full data interoperability, user-friendly and well documented software code the transferability of our results from Slovakia to Lithuania and Bulgaria, but our Consortium will remain open during the timeframe of the project to assist any nationally representative music export office or other stakeholder organization to assist them with replication.
In WP Music, Society, and Citizenship we will focus more on local sustainability and audience planning. Building on our experience in survey recycling, survey and output harmonizations, we will show that
Our Consortium partners have alreadly created6 open-source software for the music industry that had been tested in real-life policy advocacy and business cases7, and scientific uses related to piracy research8.These software tools solve important problems for WP1, WP2, and WP3. While developed with a clear music industry focus, they have found thousands of users in the open research community worldwide for other purposes, too.
In WP4 we will further improve them to work as a software ecosystem, and whenever possible, add web-based application interfaces with the ambition to make them useable for music organization that do not possess in-house R&D, IT, or data science capacities following the ISOxxxxx standard on software application usability.
1.2 O1. Map the data landscape
O1. Map the data landscape will be accomplished in the first set of tasks (T1.1, T2.1, T3.1, T4.1)
In the last decade, the evidence-based policy movement gained significant traction in Europe as well as globally. Its focus has been to increase the rigour of the evidence generated, to improve the credibility and understandability of evidence created for policy purposes. As evidence-based policies often rely on scientific evidence, the evidence-based policy movement went hand in hand with the efforts to increase the transparency and reproducibility of scientific research (See: (Munafò et al. 2017) and in an EU context (J 2015; Commission et al. 2020; European Commission and Directorate-General for Research and Innovation 2020).)
We will use an interdisciplinary methodology of economics, royalty accounting, quantiative finance, and socio-economic research in copyrights and blockchain for achieving the goals in WP1, WP4, and partly WP3. The goals of WP2 will include socio-legal research and quantiative musicology, and WP3 quantiative music sociology. All WPs will deliver data via WP5 that uses reproducible scientific and policy research supported by data science and computer science.
In our work, we will improve several quantitative methodologies used in music copyright data management and cultural statistics, among others, to produce improved indicators for business, policy, and academic use. Throughout the project, we will follow the Eurostat guidelines on creating new indicators (Eurostat 2014, 2018; Angelova-Tosheva and Müller 2019), which will ensure broad consensus-forming among stakeholders around the objectives and methodology of the improved measurements. We will create key business indicators (for individual companies), city-, regional-, nationalü and pan-European/EU indicators, but focus on the national level, because cultural policy and creative industry policy is at this level in the EU subsidiarity setting. Because the Slovak Republic is currently developing a detailed policy indicator set where SOZA is providing indicators, and EUBA’s researchers place scientific input, we will place an emphasis on this country, as we are most likely to get the widest level of business and policy user feedback. We will practically use the production side statistics, following xxxxx
Our ambition is to fill the data gaps with data that is created with similar methodology, user testing, quality control, documentation and dissemination standards that Eurostat or national statistical offices do, and balance the legal of legal mandates and large institutional capacity with relying on the quality control of the scientific community and the open-source software/algorithm development community, and the flexibility of working with voluntary and well-targetted data collection in select countries and segments. To do this, we rely on the best methodologies developed first in the United States, then implemented on EU and at last on national levels
Our consortium has a very strong background in economics and quantiative finance, and experience with royalty accounting and valuation, with much prior experience in putting this into scientific, business and policy use. Our focus will be on methodological difficulties that we have been presented in our researchers work in the past years. Our first focus is to modify the existing valuation models of the recorded industry, and intellectual property in general, with the increasing use of the so-called zero-price platforms, like YouTube, Facebook, radio, unlicensed copying, torrenting, where people do not pay directly for music. Our Consortium Members have gradually extended the music comparators model xxxx, and created econometric models that followed the evolving jurisprudence of the Court of the European Union9. we aim to standardize this method, using mapping xxxx, and transfering and improving the Hungarian/Slovak model to Bulgaria in a way that can be replicated in any European country.
To go back to our previous example, if an Italian label with an emerging artist receives 15% more euros in June than in May, currently, there is no way to know if a) there artist was doing better b) it was played in markets where the per stream rate is higher c) it received accrued income from some export markets that had been growing since February d) the artists was played in Japan and the euro rate changed favorable to the label. The currently available market information (annual or monthly total or average revenue) is not useful, because streaming revenues are extremely skewed—rendering average or median values useless to represent a ‘typical’ Italian song, artist, or label.
1.2.1 Ambition
The Feasibility study for the establishment of a European Music Observatory (in short: EMO Feasibility Study)10
Related to the music economy, the feasibility study has identified 13 data gaps, and 4 existing but publicly not available data sources. Regarding the 4 identified existing, not available data sources, the Digital Music Observatory maps the metadata of all of these. It can close half of the of the data gaps pending approval from music stakeholder (data owners), and can replicate the other data. The former CEEMID partners, i.e., Reprex, Artisjus, and SOZA have experience in collecting data, or accessing proprietary database and harmonizing data in 12 out of the 13 identified data gaps. In this project, we will focus on the Value of EU’s music sector and the Structure of the market, and some elements of Export better described and more transparent. Because statistical data collection has high fixed cost and low marginal costs (for example, gross value added, exports, and employment has the same primary or secondary data sources), we will provide a methodology and datasets in relating to 8 identified data gaps. Employment, Live music, Financing of the music sector, The impact of the not-forprofit sector on the overall economy of the music sector, Independent music companies, Neighbouring rights, Music publishing will find their way into our transferable, while the issue of Music retail will appear in several deliverables.
In the field of diversity and circulation, the _ EMO Feasibility Study_ established four data gaps, and we will provide a methodology and national use cases to fill all four of them, and in fact, provide more data, because gaps 1,2,4 Cross-border activity on radio and in streaming, and in live_music requires the separate measurement of domestic (national) use and cross-border (export or importuse). The methodological requirement for measurement is a more comprehensive ethnomusicological database than envisioned in the EMO Study, and something that we have brought to a significant, working demo stage in Slovakia with the Slovak Demo Music Database, which we will develop into a comprehensive, open source database for measuring national and international uses.
In the field of music, society, and citizenship, the EMO Feasibility Study has identified 8 data gaps, out of which we will give a very significant contribution on the EU and national levels to four data gaps, i.e., EU consumers and music, Social networks and music, Consumer patterns regarding piracy and its impact on the music sector, and even extent the social impact to enviornmental, social and governance impacts. We will add som esome important, but not comprehensive data on the training of music professionals and artists, and the scope of the not-for-profit sector in Europe.
The EMO Feasibility Study also identified 7 data gaps, or rather, data-related topic in relation to innovation. We will address all of these topics to some extent, but in this field, or aim is not to create business, statistical, or policy indicators, but to provide a deeper understanding in data management, interoperability, and metadata, which are all very important issues in music tech, trustworthy, ethical AI, NFT and blockchain, et.
Of course, we are not planning to make a competing data observatory with the planned European Music Observatory. We aim to provide a minimum viable model of creating at least a hundred useful indicators—selected from hundreds of indicator-candidates with user feedback— that goes through the unit-testing of data science and computer science, the peer-review of open-source scientific algorithm/software development, the methodological peer-review of science, and eventually user verification from music industry users. This will make sure that the indicators can be reproduced, refreshed, and placed in a future European Music Observatory with at least as good data quality as one would expect from a governmental statistical source or Eurostat. Placing our work in the future EMO require no payment or further investment—our output will be disseminated as open data, and the production code as open-source statistical software.
1.3 O2. BRIDGE data gaps
The project will collect data from multiple sources using appropriate methods, create synthetic datasets capable of populating the project indicators, and visualise and report the data collected. Objective 2 meets the expected outcomes of promoting standardised data collection on the contributions of music to the economy and estimating its impact on society. It will be realised in T1.2, T2.2, T3.2, and T4.2, and reported in real-time via the Digital Music Observatory (D5.1).
We will use the standard mapping of the music industry11 to identify the configuration of the music industry, and place critical information gathering points where value is added, jobs are created in the conversion chain of music use to industry (artist) revenue.
The working group of ESSnet-Culture has summarized the best practices in cultural statistics and xxx, that our Consoritum members have been following for 8 years. The most important takeaway is that in cases where official statistical offices have no clear mandate or resource to collect data (microenterprises) or to process existing data with traditional methodology (data products made for example from Eurobarometer, EU-SILC, AES, anonymized tax filings, data from inflation measurement surveys), and even with the recent improvement efforts both in terms of methodology and statistical regulation, the situation will not significantly improve. The only way to seriously imporove data availability in the music industry is to apply standard statistical procedures in a voluntary way with a joint effort from cultural policy organizations and representative industry organizations.
In the last years, ESSnet set up a similar working group to analyze the use of big data for statistical purposes. Because big data sources offer large volume of data, it is not a problem to create large enough samples from these sources that will represent well the observed universe, and let the Law of Great Numbers work. The problem with big data sources is selectivity and biased access, and often a very high level of sparcity12. Our of our main methodological and innovative contributions in this objective is to borrow big data methodology from quantiative fianance, where solving similar problems has many decades of scientific practice.
1.3.1 Ambition
Our ambition is to design data pipelines, with appropriate methods, that will result in high-quality data that can be processed into simple and composite indicators and a easy-to-understand visualizations. We will follow, whenever possible, established statistical guidelines and techniques. In other words, we will employ statistical production procedures like national statistical offices and Eurostat’s harmonization efforts, but with significant differences. National statistical offices and Eurostat (together, the ESSnet) are state organizations and follow mandatory statistical regulations. These regulations are a result of careful judgement of member state interest, institutional capacities, policy priorities, and collection costs. Our Consortium is not bound by these regulations. The advantage of this is that we can align our priorities entirely with our stakeholders’s needs (as collected in xxxx) and known from sources like the EMO Feasibility Study. The disadvantage is that we can rely only on voluntary participation, and voluntary data sharing. For example, we have the possibility to simplify the mandatory enterprise surveys (which target companies with at least 10 employees and therefore almost never collect from music businesses and NGOs), but we must find strong national partners to reach them in a way that allows a representative data collection, or at least, the statistical correction of collection biases. We will rely on the policy recommendations from Finland, and the experience of the former CEEMID consortium that managed nationally representative data collections in Hungary, Slovakia, Austria, and Croatia. We will apply, whenever applicable, the same rigorous statistical unit testing and documentation standards that Eurostat uses. However, not being official statistics, or quality control will be different. We hope that this will lead not to lower, but higher data quality than official statistics. The application of the Open Policy Analysis Guideline, and the methods of academic data science, coordinated by the TURKU and applied by REPREX offer very strict quality control. Our statistical processing code is subject to blind peer-review on CRAN to make sure that our code and its documentation meet standards. The openness of the code allows further scrutiny in case of ‘suspicious’ user experiences—TURKU and REPREX reply to all ‘issues’ raised by users. Fur Because we do not only collect data and create datasets, but for each indicator we write an open-source program, our work is fully replicable. This means, for example, that a future European Music Observatory can just run our open source (free) software and refresh our indicators. We aim to give a similar or higher user experience and quality then Eurostat. Our data is documented with SDMX compatible codebooks, and DublinCore and DataCite (mandatory and recommended) metadata, ensuring full interoperability. They are available in tidy format for easy importing for single users, and in our Eurostat-like Rest API on the Digital Music Observatory. We replace the authoritative copies of official statistics with authoritative copies placed automatically on Zenodo and OpenAIRE. Our data analysis with the data makes sure that the data is usable. The data analysis of the Consortium remains open, per OPA, and goes through scientific peer review.
References
See Music Economy, Diversity and Circulation, Music, Society and Citizenship, and Music Innovation.↩︎
The Open Guidelines state this as “Code is clearly documented into a dynamic document, or open notebook. No spreadsheets.”↩︎
We created this submission to the market inquiry of the authority (Competition & Markets Authority 2022) using the dynamic document concept. The “live document” can be viewed here, with the version authoritative copies are held on the Zenodo identified by separate DOIs (for example, 10.5281/zenodo.6088844. The entire document with producing data processing and software code is available on the Github repository. This technology enabled us to create the Central European Music Industy Report harmonizing data from 12 countries in the region and further 8 countries for benchmarking.↩︎
Short description here (Turku, Reprex)↩︎
short description Artisjus, SOZA↩︎
The regions software that makes regional data comparable in Europe across countries and years was used in the creation of piracy research: (Bodó 2020).↩︎
The jurisprudence of the Court of the European Union, particularly OSA v Léčebné lázně Mariánské Lázně (InfoCuria 2014) and AKKA/LAA vs Konkurences padome (InfoCuria 2017) must be considered when making market comparisons, respecting the agreement made closing the Commission v CISAC case (InfoCuria 2013).↩︎
Feasibility study for the establishment of a European Music Observatory (European Commission et al. 2020, pp31–38).↩︎
The standard mapping of the music industry was developed in the US ((Hull et al. 2011), and was adopted by the Joint Research Centre Institute for Prospective Technological Studies (Leurdijk and Ottilie 2012)). Artisjus and its neighboring rights societies created a more granual report in Hungary and Slovakia (Antal 2015a, 2019b), and a special version for focusing on the unlicensed, “free” use of music in Croatia (Antal 2019a), and eventually the comparative report to finish the original mission statement of the CEEMID project .↩︎
An overview of methods for treating selectivity in big data sources (European Commission. Statistical Office of the European Union. 2018).↩︎