Chapter 4 Innovation

Our Demo Music Observatory wants to help with free, frequently published and validated indicators for the industry to participate in business innovation, and to find innovative solutions for the COVID-19 crisis.

4.1 AI & machine learning

AI plays a very important role in the music industry. We believe that probably half of the legal music sales in the world are driven by AI algorithms and machine learning. This places most industry players in a very uncomfortable position. They are using sales platforms such as YouTube, Spotify, and Apple Music, where their music and its associated data and metadata feed into proprietary and unknown black-box algorithms which to a large extent determine their future success or failure.

Most music stakeholders, such as labels, publishers, self-published bands and even smaller collective management societies are too small to have access to advanced data engineering and data science teams. While presence on the online platforms means that music has sufficient data and metadata (and our Demo Music Observatory shows that it is very easy to find further supporting data for modelling), these capacities are beyond the organizational and financial reach of most music industry players.

What makes our approach unique is that we try to bring down the costs and entry barriers of AI to the level of music organizations.
In large organizations where music is present, typically 80-90% of an AI project cost is related to data processing. In small organizations that are buying some or most of the data, the actual AI modelling cost is only 5%, because all data acquisition and data processing costs are incurred before a single machine learning model can be deployed. We are trying to save a lot on data aquisition, internal data costs and data processing so that deploying AI, forecasting, valuations, impact assessments or better reporting remain financially feasible.

We are committed to develop one AI model with our partners and make it freely available connected to this website on an interactive web application. We want to demonstrate to our partners that playing a non-passive role with the AI revolution is within reach through our open source, open collaboration based reproducible research approach.

4.2 Forecasting

Forecasting models are very similar to AI and machine learning models: they need large amounts of flawlessly processed data. We are committed to show that our data is flawlessly processed, and we are committed to demonstrate to our partners that we can deploy a useful, quantitative forecasting tool similar to those that are used by banks and large corporations for free in an interactive web application.

4.3 Valuing music

When CEEMID was founded, most of the CEE stakeholders had little or no experience in valuing their intellectual property and setting its prices. Over the years it turned out that this is not a regional phenomenon, but very typical for all smaller national market players even in mature markets. While copyright administration usually follows national jurisdictions, many prices are set outside of national markets, or the CEE region in the form of various agreements (for example, in the case of traditional mechanical licensing) or in the case of global digital platforms. The nationally set prices are often set by collective management organizations. These prices are differently regulated in the region, and often have not been changed since gaining national independence or transition to the market economy.

While all EU member states are participating in international agreements that regulate royalty setting principles, this international law does not create an unambiguous ex ante price regulation. In the absence of clear legal mandate for an ex ante price regulation, most price disputes go to competition authorities and special competition courts which have an EU-wide jurisdiction for ex post price reviews. Since competition law is directly applicable EU law, some cases are sent to the Court of the European Union for preliminary rulings.

In our view, several legal precedents, which are binding now in all mature and emerging EU markets, were decided unfavourably for the music industry because of the weak evidence produced by music rightsholders.2 This jurisprudence was set out in the context of background music use in hotels, restaurants and catering, but may be applied to any licensed form of music. In these cases, the EU Court set out that royalty prices should be equal in the European single market, subject to objectively verifiable differences in social, cultural and economic conditions.]

Stakeholders in the creative and cultural sectors are micro-, small and medium enterprises, and often non-profit organizations who are lacking professional skills to fight a competition case. On the other hand, royalty payers are often large corporations who do have the expertise to enter price negotiations or competition disputes with big legal teams. Because the ex post nature competition disputes results only in downward price modification, the fact that competition law applies already starts a downward spiral in prices.

CEEMID has been used in competition cases, but we believe that the music industry should be more proactive, and not only react to such reviews, because they by nature can only decrease royalties. It is a common and bad practice that industry players are setting their prices to some average European value. Whenever there are downward changes, the average goes down, and the industry enters into a downward spiral. The current, one-sided application of competition law is a recipe for disaster, and will lead to a race to the bottom. This is not only a threat in the less developed markets, where usually both royalty levels and total royalty collections are very low, but in the richer markets. Because of the methodologies adopted by the EU Court, royalty levels are compared across the EU and unjustified differences can be levelled out by the Courts.

The only way out from this is the long and hard and legal way: to create primary valuation, and CEEMID had been able to collect, process and provide data for all types of valuations in the past. Our data in the past 5 years were used all pricing methods reviewed by IFPI’s public Valuing music document (PwC 2008).

  • We used the hedonic price method to review and correct tariffs for restaurants and hotels, instead of relying only on international comparison in Hungary and Slovakia;

  • A complex market comparator model to value broadcasting rights and private copying in Hungary and Croatia;

  • We used experimental methods in disputed public performance payments in Hungary;

  • We used hypothetical evaluations when no other data was available in Hungary, Slovakia and Croatia.

Our valuations must follow the international fair valuation principle, and must rely on objective data3. In valuations we commission surveys and collect data following Eurostat’s ESSNet technical recommendations, whenever applicable.

We are committed to make indicators available as credible, peer-reviewed, open data for fair valuations. We will engage copyright and neighboring rights associations and their representative bodies on how to do this best, and support the industry with better pricing.

4.4 Economic impact assessment

While classical music and national folk music have long merited public support, jazz and popular music have been seen, correctly, as market activities. In many countries of Europe, the popular perception of music is that it is granted, publicly supported art form, whereas in reality the music industry is more heavily taxed than the average sectors, an in fact, public support is dwarfed by tax payments.

A more focused industry cooperation may be necessary in the forthcoming modernization of the EU VAT regulation which will change the VAT levels and brackets. This could be a once-in-quarter century opportunity to argue for more just taxation of music in the whole European Union.4

Economic impact assessments are crucial parts of public policy and advocacy reports on better taxation, or in the design of better creative industry strategies and economic policies. Our methodology enables music and other creative stakeholders to create impact assessments for economic and cultural policy consultations. For example, in the Slovak national music industry report Správa o slovenskom hudobnom priemysle we showed the relative economic impact of car manufacturing and a music festival is similar in Slovakia, however, the music business creates a many times bigger tax base (Daniel Antal 2019b). A similar argument could be made in Hungary and Croatia where the tax treatment of music is one of the biggest obstacles to growth.

We made iotables, a data processing, integration and calculation tool for economic impact assessments, peer-reviewed, open source and available for comparison with national statistics office calculations, for example, on value-added and employment multipliers5. The use of this software requires programming skills, an understanding of input-output economics and national accounts. We are currently looking for smaller users, such as concert halls, granting authorities or individual festivals to use our data, and to create tools such as web-based apps that can make this software more user-friendly and standardised, helping smaller players who do not employ specialists to make their own impact assessments, which is often required by granting authorities for larger funding applications.

We are looking for a partner to develop one indicator: for example, a tax or GVA multiplier that we can report here for all EU countries and make available as open data.

4.5 Automated reporting

This website will be a use case of automated reporting. The website will update every day, and refresh itself with new data and citations. We believe that we offer a unique selling point to an industry that is dominated by small enterprises without large research organizations. We would like to automate tasks in data acquisition, processing, documentation, citation, pre-modelling and pre-publication, so that individual researchers can work as if they were surrounded by a team of research assistants.


Antal, Daniel. 2019b. “Správa o slovenskom hudobnom priemysle.”

PwC. 2008. “Valuing the Use of Recorded Music.” IFPI PricewaterhouseCoopers.

  1. Probably the most important, EU-wide applicable royalty setting precedents were set in the CEE region in AKKA/LAA vs Konkurences padome and Léčebné lázně Mariánské Lázně v OSA↩︎

  2. The Commission Regulation (EU) No 1254/2012 adopted into EU law the International Financial Reporting Fair Value Measurement Standard, which must be used in all EU member states.↩︎

  3. On this important policy topic, refer to the thematic website of the European Parliament or the European Commission↩︎

  4. In the bibliography see Daniel Antal (2020a)↩︎