| Proceedings | Social Programme | List of Participants | Practical Information | Speakers | Exhibition |
Are
databases still protected?
Laurie Kaye, UK
The legal protection of databases is a vital subject for the industry to address. Have we got the right balance in protection for investment in databases between owners, publishers and users? I hope to show that there exist a number of threats to this balance. Yesterday, EADP put in a submission to the Commission who are looking at the scope of database protection and who published a very one‑sided paper that actually asked the question, do we need this protection? There are other threats from new service providers. Google, for example, is aiming to make the whole of the published world searchable and indexable. This raises all kinds of concerns for publishers. For legal purposes a directory is a database, so I will use the terms interchangeably. Data protection, as the William Hill case that I will present to you shows, has significant and immediate business implications.
I will give you a quick guide the Database Directive. I will briefly tell you the facts of the cases that the European Court looked at and explain the Court’s decision. I will then focus on how this impacts on you and what you should do.
A database can be defined as a collection of independent works, data or other materials arranged in a systematic or methodical way and accessible by electronic or other means. This definition is platform neutral.
A database has to be a collection of works. For example, a directory containing a mixture of statistical information and other added value content is certainly a database.
a. Types of protection
The Database Directive was adopted in 1996 and implemented throughout the EU by around 1999. It focuses on the sui generis right of databases. A database has two potential types of protection: copyright in the structure and in the investment in the content, the sui generis right. If, in addition, you have copyright material, there is extra protection.
b. What was meant to be protected?
The cases looked at substantial investment – money, time and resources – in either obtaining the contents of the database, verifying it or presenting it. There is no reference to creating data. The issue of creating data – obtaining information and adding value to it – is the area that has the biggest question marks.
The right prevents other people from extracting or reutilising the whole or substantial part of your directory. Extraction means simply taking or copying by any means possible, electronic or non-electronic. Reutilisation means making the data available to anyone else on a website or in a print product. It lasts for 15 years. It is a renewable right if, over time, you refresh the database so as to create a new investment. You, as the publisher and investor, own the right. It is an EU‑only right. It does not exist in the US.
a. British Horseracing Board vs William Hill
So far, this sounds like quite a strong protection. Where is the problem? The European Court of Justice (ECJ) looked at four cases. I will focus on the British Horseracing Board vs. William Hill. There were three other cases involving the use of fixture lists by the Premier League in the UK and the equivalent in Greece, Sweden and Finland.
The British Horseracing Board is the entity in the UK that runs the UK horseracing industry. It has a large database that includes all the information about horses, runners and riders. It costs about £4 million to run. The British Horseracing Board generates around £17 million from their database. It distributes the data through a number of sources. William Hill was a licensee through a third party called Satellite Information Services. This contract did not give William Hill the right to use any of the information for running its online betting site. This is exactly what it did. In breach of contract, it used a limited amount of data – the name of runners, the date, the time and location – to run their website. The British Horseracing Board, which had a substantial investment, sued them for infringement of the database right. The Board were very confident. They thought this is exactly what the sui generis right of substantial investment in obtaining, verifying and presenting data must be about.
The case started in the UK courts, but they could not work out the answer and sent to the European courts. The ECJ said that creating data is not protected. The Database Directive protects obtaining third party data. The British Horseracing Board, as a result of operating its activities, which is to create the fixture list, created data.
b. What qualifies for protection
There are two aspects to this. The first is the so-called ‘spin-off’ theory. If the database is created as a spin-off to your mainstream activities, it does not qualify for protection. The investment in the creation of the database is protected, not the contents inside it. Secondly, they said the term ‘obtaining’ means seeking out independent materials and collecting them in a database and not the resources used in the creation. The ECJ has a view of a database as a passive collection of data that exists as opposed to created data. This raises a question about whether added value databases are protected or not.
When the case went back to the UK courts, they ruled that the British Horseracing Board’s investment in creating the initial list of runners and riders is, indeed, creating not protecting and thus outside the scope of the database directive. It follows that check data after you have created it does not count either.
If you are a directory publisher with databases made from information obtained wholly from a third party, you are in a good position. If you are in the position of the British Horseracing Board and publish data that you create, you are in the worst possible position. Most of you, I think, are somewhere in the middle. You may have a mix of third party data and added value or created data.
c. Quantitative and qualitative infringement
It was not all bad news. The ECJ also said that you infringe the right by taking protected data from a third party and not just from the source. The ECJ gave some guidelines about what constitutes taking a substantial part. There are two measures: you can infringe the right both quantitatively and qualitatively. The former refers to the amount of data taken and the latter the amount of investment taken. Interestingly, if you only take a small part of the investment, even if the data taken is absolutely key to their business, it may not qualify. The courts will look at whether what was taken was a substantial part. It is not entirely clear cut. It has to be a substantial part, either in terms of volume or investment or both.
The ECJ also made some comments about how far you can infringe the right by regularly taking small parts. If these small parts add up to a large part, quantitatively or qualitatively, the right is infringed.
This impacts industry in two areas: how narrow is the scope of your copyright and database right protection. If it is weak, you are in a position to take data from third party sources, free of infringement risks. Conversely, it means your competitors – both existing and emerging – can take it free of constraint. Ultimately, where does the balance lie? Where does this leave your directory and database contracts?
a. Protection
Neither the investment in creating data nor the investment in verifying this data are protected. What is protected is investment in obtaining data that already exists, investment in verifying this pre‑existing data and investment in verifying information already in your database. This should mean that subsequent presentation and verification of created data, once it has been collected, should count.
b. Hybrids
This brings us to the crux of the problem. A typical B2B directory publisher is a hybrid consisting of third-party data that has been licensed in or obtained and added-value data that has been created. Unlike the British Horseracing Board, who were in the position of only having data they had created, such a hybrid database should qualify for database right protection on the basis that it has made a substantial investment in obtaining, verifying and presenting third‑party data. Once data has been added to the database, any subsequent verifications will be covered also.
If a mixture of obtained data from third-party source and created data is taken without authority, you, as the owner can argue that your right has been infringed because at least some of your investment is protected. What is the legal position if only the created data is taken by, for example, data scraping or by a search engine? There is a risk for all database producers that they have no protection in cases where substantial amounts of created data has been taken. Whether you think this tips the balance of protection in the right or wrong way depends on your perspective. Google, for example, will be pleased to see this dilution of rights. For a database directory owner, it goes the wrong way. The Commission is now looking at this sui generis right. EADP has submitted a paper to look at this very carefully. The Commission has based its initial view – which has raised all kinds of questions about whether this right is worthwhile – on an analysis of the volume of data entries in the Gale Database Directories. This is a very one-sided view. Now is the time to support the EADP if you believe that maintaining protection through the sui generis right is important.
c. Contracts
In the UK, one of the British Horseracing Board’s customers said that as a result of the ECJ’s decision they believed they no longer had to respect their contract and pay the Board anything. The UK court said that this was wrong, that contract rules. If you have no contractual relationship, your ability to take the content will depend purely upon the legal position. Getting customer contracts right is important. Some contracts have a no challenge clause, one which says that in the event that the licensee disputes the validity of the licensor’s rights, the licensor can terminate the contract.
d. What should you do?
One man’s created data is another man’s obtained data. There is a need to separate out the creation and obtainment activities. When you are in a contractual relationship, business is as usual. However, the rights set out in the contract are important. A greater degree of definition and segmentation of rights is needed. Make sure you have a chain of title to the content. The database right protects investment. If you have outsourced this to a third party, make sure you preserve the rights in this contract. Make sure you have the right notices. Few directories, online and offline, refer to database rights apart from copyright. Support EADP’s lobbying activities. Importantly, when the EADP asks for examples, please respond. Finally, watch this space. This debate about database rights is part of a much bigger debate between, on one hand, rights owners, including the music, film and publishing industries, and on the other hand, a variety of users seeking to dilute the scope of those rights, including academic institutions, search engines, news service providers and so on. This battle is to be ignored at your peril.
From the Floor
You gave information that the European legal position was founded upon the Gale Directory Database. Can you give a little more detail about that? Why did they reach this conclusion?
Laurie KAY
The Directive required the Commission to report back to the Council every three years. The Commission was very slow in doing this. It commissioned a study by a Dutch law firm called [inaudible], which received only around 40 responses from the publishing community. The only statistical evidence that was submitted was from the Gale Directory of Databases. Apparently, they had great difficulty in finding any measure of what the database industry was. This directory showed an increase in the volume of database entries in the EU and US between 1996 and 1999 and a falloff from 2001. This was interpreted by the Commission as evidence of the Directive not working and of the industry declining the sui generis right.
Nikolaus Futter
We sent a seven-page position paper to the Commission yesterday in which we stressed the Gale data is not reliable. Gale is part of the academic and has a different approach to what a database is. The interpretation of the decline was erroneous. There were economic factors at play. Many databases were founded during the Internet boom, which subsequently went out of business. The reduction is also reflected in the different media used for databases. In addition, portals count as one database even though there may be lots of other databases behind it. This is the wrong approach to measure the industry. We urge the Commission to commission a decent study on how to size the industry. The EADP will also try to measure the sector. For this we will need to get information from you on how many databases you have, your turnover, your background and so on. It could also be the case that the paper is a provocation to the publishing industry. Perhaps they just want to urge us to give them more information on the industry so that they can work on the sui generis right.
Nikolaus Futter
Without risk an initiative, a database cannot be protected. There are many positive effects of the Directive on the industry but there is also some risk that the Commission could look at in a different way.
Laurie KAYE
Irrespective of whether or not it got the balance right, there is the bigger risk that the Commission does not have the right statistics, it might conclude that the sui generis right is not achieving its purpose. There are those who support its abolition.
Nikolaus Futter
The academic world argued that its costs increased because of the protection of databases. This is nonsense because they are exempt. Erroneous arguments are used to weaken the Directive.
Stella KRANKUSOVA, Slovak Telecom
Is the data considered protected in the following example? A company, a telco company for example, creates a branch to which basic customer data is moved, say, on a daily basis. This branch then collects more data on this customer, such as opening hours or logos, publishes it in print, Internet or voice. Is this data obtained or created?
Laurie KAYE
I cannot give you a specific answer to that question in a public forum. Are they two separate entities?
They are daughter companies. The revenue is consolidated at the end.
Laurie KAYE
What was the data collected in the first place?
Customer data including name, telephone number and address. This is the information that we have when the customer registers the line, the information we need for billing.
Laurie KAYE
Are you a telco?
Yes.
Laurie KAYE
The principle is this: if the data that you are making available is a by‑product of your main activity, providing the telco service, it does not qualify for the database right.
Is this true even if I add additional data, which I need to go to the customer to obtain and verify, such as categories, key words, opening hours and so on.
Laurie KAYE
This is the hybrid investment: data that spins off your main activity and created data. The courts would ask if there was substantial investment in the additional data obtained from third parties. If it just sits on top of it, you might still have a problem. This is why you may want to think about separating the activities that are the spin‑off off of the telco from the directory publishing activities. It is sitting on a weak base. This is said without prejudice. Please do not sue me.
Trevor FENWICK
It is unfortunate that the first serious test case has involved a non‑publisher. The main economic activity of the British Horseracing Board is regulating and administering horseracing. It would be much more clear to us as publishers if the case had involved a publisher. Many B2B publishers created content that they used to distribute in print or off‑line. Now they create access to that. It is data that they have created, with a lot of added value and analysis. My belief is that, in this case, the economic activity of publishing, where there is risk and investment, would be supported. This is one of the clarifications we need to obtain from this review of the Directive from the Commission.
Laurie KAYE
Both the ECJ and the UK courts stressed that the British Horseracing Board was the only organisation that could create the fixture list. It was sole source data. William Hill could not have obtained it anywhere else. This made the position of the British Horseracing Board even weaker. For data that anybody could have created, the position is better. It depends on whether you are the only source of that data, for example, if it is the result of your telco activities.
Nikolaus FUTTER
In Austria, we use the company registered data for our database as a basis. We were sued by the Austrian government for breaking the copyright they have on the database based on the sui generis right. In the proceedings, they tried to ban us from our economic activities. Thus it was a big threat to the company. The High Court ruled that the database is protected by the sui generis right. This, in our opinion, was a mistaken ruling. The database cannot be protected in this way. There was no risk. Investment was made for administrative purposes, not for creating the database. However, the court also said that it is sole source data and the only qualified source from which to obtain information on the companies, and thus we are entitled to use the data.
The interesting thing about this case is that the data protection and public sector information issues were both taken into consideration. The protection will be far weaker for databases that were not created as databases but for other purposes than for those created as databases as an economic activity.
API Online Sweden
If you just give the name of the source, can you then copy as much as you want?
Laurie KAYE
No. There are some exceptions for non_commercial research and private use. However, if the data is protected by this right and you take a substantial part of it for commercial purposes while citing the source, you are not protected.
From the floor
How much do you have to take for it to be considered substantial?
Laurie KAYE
You have to look at the investment. You have to look at what you take and imagine that you were the person that made the investment to create it. How much would it affect you? Quantitatively, all sorts of numbers have been used. In qualitative tests, you have to look at where the database producer made the investment. For example, they may have spent a lot of money verifying the data obtained from a third party. If you then take data that has been verified, applying the qualitative test, you would probably infringe. Insubstantial parts can be taken.
From the Floor
Do you know what the US legislation currently is in this area? Does it compare with European legislation?
Laurie KAYE
The US has no equivalent to the sui generis right. In Europe, there are two types of protection: the database right protecting investment and the copyright. These cases did not look at what copyright means. Copyright might cover, for example, the way in which a website has been structured. Taking a data dump out of it, you will not infringe the copyright. In the US, they only have copyright. They have an equivalent bill lying around that would be similar to the database right, but there is a lot of lobbying against it. Other remedies may apply: breach of contract, breach of confidence and so on.
From the Floor
My understanding is that in the US there is no copyright on Federal documents. There may be in state or lower level documents. The quality of public information holders in Europe is much higher. American mapping, for example, is held to be much weaker as a public resource than European mapping or UK mapping. The American national federal copyright does not exist or is not enforced.
Nikolaus FUTTER
It is definitely free. There is no copyright on federal documents. You only have to pay for marginal costs in the US. If it is online, it is free.