Conferences
Noank Media: What I see wrong with the idea
While I was a the Future of Music policy summit in Montreal I listened to Dr. Terry Fisher's presentation on the Digital Media Exchange or its first for-profit incarnation of this project called Noank Media. The idea is that students and/or ISPs would pay a fee that would allow participants in the system to freely download music and other media through their P2P system. The system would track usage of this media and then funnel 85% of the fees to the artists, based on usage statistics collected. It seems that the Noank team hasn't published many details of this project, so I can't give you an official link for more information. Instead, let me point you to other write-ups of this talk: NowToronto and Clicknoise.
I've been thinking about this type of system for quite some time now -- I've even blogged about these systems in the past. I've come to the conclusion that these systems are never going to work. Given that, you can imagine that I had a long list of questions to ask of Dr. Fisher after his talk. However, my critique of the Noank Media project is nothing personal -- I have a great amount of respect for its participants and their motivation for going out and trying something completely new. Cheers!
[ Read the whole post. ]
Technorati Tags: conferences, file-sharing, future of music, innovation
Now on to the heart of this post: I feel that the project is so ambitious and that the team has chosen to bite off much more than is possible to accomplish for the first go around. Let me lay out my main points:
The cost of running this project:
The basic business plan and how much money will be spent on what is probably the must crucial issue in this proposal. The income is meager to start with -- the proposed charter system in China will collect $10 per student for this system and money will be collected from advertising revenues. Now keep in mind that of the collected fees (I think thats the $10) 85% will go to the artist. That leaves Noank with $1.50 per student plus advertising revenues. Not much at all, really.
Where will this money be spent? Employees and hardware -- many engineers and business development folks will be needed to realize this project. Someone will need to go out and get the content. Someone will need to convert the content into whatever format will be used. The software for this entire system will need to be written. Advertisers need to be lined up. Noank plans to provide arbitration services. They plan to track usage of their media, which requires a massive amount of work on the client. More on that later.
I'd say this requires a team of a few dozen people. And they will need computers, paychecks, travel budgets and everything that an enterprise needs. Not cheap at all. I won't do the math of how much money will come in or will go out. I personally do not see how these two will line up though.
Let's compare this enterprise to two ventures I am familiar with: EMusic (that is the old skool EMusic prior to 2001) and Magnatune. Both companies split their revenue with the artists 50/50. EMusic spent a lot of money acquiring its content in order to provide a download service and in the end their revenue model didn't work out and the company was sold to Vivendi. Magnatune doesn't spend much money on acquiring its content and roughly offers the same services as EMusic did. Magnatune is a viable and tightly run small business -- the CEO John Buckman, is the main tech head for the company, keeping costs down. With such a shoestring budget Magnatune can exist, whereas EMusic had to drastically change its model in order to survive.
Noank media proposes to offer vastly more services and software on a much more drastic revenue split favoring the artists. I love that concept, but I don't see how this could possibly be bootstrapped into existence. There is much too large a gap between the EMusic/Magnatune model and the Noank model.
Fraud & artist authentication:
Dr. Fisher never mentioned fraud at all. An audience member asked about it, but Dr. Fisher dismissed fraud as being a serious issue for Noank. This project is going to be handing out money to artists. And wherever there is money to be had, fraudsters abound. What is to prevent crooks from creating phony bands and then gaming the usage counting system to ring up big bucks for their phony band? How will they know that a person coming to collect money for Britney Spears is in fact her and not a crook? SoundExchange is having to deal with this problem and given the recent press it doesn't appear to be easy. This is one of the main things that SoundExchange does and Noank proposes that this will be just one part in their massive system.
I know that PayPal had to process an ungodly number of transactions before they were able to offset the cost of dealing with fraud. They spent tons of money on fraud prevention and on dealing with fraud after it happened. We all wonder why PayPal takes so much money for moving some bits from one account to another account -- well, fraud is the answer. Every honest user has to subsidize PayPal's efforts to counteract fraud.
Fraud on the net is a painful reality. When walking in bad neighborhoods in the world you can see bars over windows and all sorts of measures taken to keep out the bad guys. However, on the internet we all live in a bad neighborhood -- the crooks are just as far away as all the good guys. Anyone dealing with money on the net needs to spend serious time and money addressing fraud issues, just like people putting up bars in front of their windows. But it doesn't appear to be a serious issue to Noank.
P2P as a delivery platform for paid media:
P2P has been embraced by the masses to deliver free media -- no doubt about it. But crappy download speeds, having to hunt for content and aborted downloads are the usual and accepted fare for a P2P experience. But if users pay for a service, the quality of the service must be better than the average P2P experience or people will not use it.
Noank wishes to use P2P in order to offset bandwidth costs, but realizes that they need to have central servers that can dole out content when the P2P network cannot deliver. P2P only works well for popular media -- if many people have it and many people want it, there are many places to download it from. But for less popular media, P2P networks fail and the central servers will need to dole out that content.
The overall effect is that yes, the bandwidth costs will be reduced by using a P2P network, but I would be surprised if the bandwidth fees are less than half of an equivalent non P2P system. And half of the bandwidth cost is still quite a large chunk of bandwidth to pay for. And, regardless of what people say about bandwidth being cheap, they're wrong. Ask anyone who operates a service with a large number of downloads. Bandwidth and hosting ain't cheap.
Does the trade-off of quality of service vs reduced cost keep the user in mind? I dont think so -- and if your users don't use the system, then the artists get no money and the whole system falls apart.
Usage tracking:
One of the key points of the Noank system is usage tracking of the media. Whenever a user plays a piece of media a play count needs to be incremented on a server somewhere so that census reports can be established to pay artists. The problem is that users are quite attached to their music players and loathe to switch to something else. Once users have found a player they like they don't want to switch. Trust me, when I worked on FreeAmp (now Zinf) while I was at EMusic, we heard all sorts of things from music player users.
So, you either need to write the world's best media player and get people to switch, or write plug-ins for all the existing media players to support what you want to do. Neither of these options is realistic -- writing the best media player is something left to fanatics at places like Apple. And writing plugins for a number of players on a number of platforms? Even if these players provided the plug-in framework to make this happen you'd have to spend a lot of money coding these plugins and releasing them for various platforms. Its a huge time and money sink!
Last but not least, lets examine the case of one really popular player: iTunes. The iTunes plug-in model is very limited -- its essentially locked up by Apple. Well, you could get Apple to open it up, right? Ha! The iTunes team is probably the best sheltered team at Apple -- its nigh impossible to get to them, since every Tom, Dick and Harry wants to have their button/gadget/service included in iTunes. How many actually are included in iTunes? Unless its an Apple project or partner, none.
So, how will Noank convince users to leave behind their favorite player and start using the new player? I have no clue. I do know that when a free service sucks, people will continue to steal.
Bootstrapping:
In the question and answer session Dr. Fisher mentioned that "This will only work if everyone is on board". This is pretty much true for collective licensing systems. But I guarantee you that not everyone is going to be on board with this system -- when have you seen an entire industry be in agreement on anything?
I think this system could work if everyone was on board and the system had critical mass. The problem is that getting critical mass is mind bogglingly expensive and takes a very long time to achieve. The Internet bubble showed us that gaining critical mass is crucial but a very elusive goal -- vastly more companies failed than succeeded.
Noank will need to raise boat-loads of cash to last them for a number of years while gaining critical mass. 5 years for that? 10? Which investor will sign on to a "everyone must be on board strategy"? I wouldn't.
Those are my main points. I've got a few more quibbles here and there, but they don't warrant making this long blog post even longer. I'll leave you with one more thought -- I don't remember exactly who said this to me, but I want to say it was my EMusic boss Brett Thomas: "You can solve one big problem at a time. You can't solve a bunch of big problems and hope to succeed."
Dr. Fisher answered 6-7 questions from the audience with: "I agree, that's a significant challenge". If Noank decides to pick one hard problem to solve, the other 5-6 hard problems will cause the project to fail. Trying to solve all the problems at the same time will also cause the project to fail.
As you can see, I don't have much faith that this project will succeed. But I wish everyone at the Berkman Center and at Noank good luck!
Posted by Mayhem at October 26, 2006 12:12 PM
You note that the "Noank team hasn't published many details of this project". The following relevant details should answer each of your concerns.
1.Cost of running the project.
You note that "the most crucial issue ...[is that the] income is meager to start with... Not much at all, really."
There are more than ample funds in the Noank busines model to provide our services. The examples you use and the personal history that you have are not relevant to Noank service because of the different scales of the businesses. The Noank business model is based on collecting small fees, working on very small margins and achieving large market penetrations. we have begun to do so. Here are some numbers:
China and Hong Kong - signed agreements for about 15 million student subscribers. The price has not been fixed, but is expected to be in the range of $10-$30 per year, which is about 1% of average student university cost in China. Taking $25/annum X 15 million yields gross revenue of $375 million. 15% for our admin (much lower than 50% that you correctly note as the norm) provides about $56 million to run the Chinese service. Is this adequate? yes. Will there be additional revenues and admin fees? likely yes, from advertising and such.
You incorrectly assume that we would take all the ad revenue, like other services. we don't. Copyright holders get 85% of ALL revenue from Noank, dramtically increasing their payout. They also get to see all the income trails. Our financials are transparent to copyright holders with no hidden "other income".
Will the revenue and admin fees increase dramatically when we add broadband and mobile services in China? yes. Will the global revenue and admin fees increase dramtically when we add other countries to Noank service such as Canada, Brazil, Europe, and the US (next on our implementation aganeda)? yes, dramatically. We are in talks now, for example, with university consortia in Canada and the US, interested in spearheading the Noank North American services.
2. Fraud & artist authentication
The fundamentally different approaches that the Noank service takes result in much smaller incentives and consequences for fraud. For example, if you disable the copy protection on a DRM system, you generally save money - cost of downloading, streaming, etc. In the Noank system, your service provider still pays on your behalf so you don't save any money. The only consequence of disabling the Noank client software is that your content "votes" aren't counted towards paying the artists whose works you consume - which is antithetical to the complaints from most copyright abusers that the existing system does not reward the artists THEY listen to and watch.
The Noank distribution formula takes into account the fees paid on behalf of every individual subscriber and allocates these to the copyright holders whose works are consumed. So... If my fees are $2 month and I only play my own song, once in a month, I would receive 85% of $2 ($1.70). If I game the system and manage to have the Noank system report that I played my song 10, 000 times in the month, my fee is still $2 and I still collect $1.70. While this example is an oversimplification of the distribution formula, which takes into account factors such as novelty (newly released works) and permission to make derivations (paid slightly higher), you get the gist.
Noank is the first service that I know of that deals squarely with the issue of artist fraud. Incubated at the Harvard Law School, Noank has a dispute artibration service as one of our core services that deals with the inevitable contentions of who wrote what (and how much of it). In addition to actual fraud, our system deals with honest misunderstandings and complicated copyright assignment problems amopng different territories, for different terms, etc.
3. Usage tracking
You note that "users are quite attached to their music players and loathe to switch to something else".
Correct. We agree. The Noank counting system is unique. We count usage by ALL players. Players can be time-based, such as iTunes, Windows Media, open source, our own Noank player, or your own favorite. They can be Microsoft Word, Acrobat Reader, Photoshop, or any other application program. The Noank client reports consumption of all content within our catalog on Windows, Mac, Unix, or recent cell phone devices. There is no restriction for users on either content use or players.
4. Requirement to have "Everyone on board"
Professor Fisher corectly notes that the Noank system needs all key stakeholders be "on board". That means content owners, service providers, and end-users. Our experience so far has been that content owners "get" the Noank business model immediately and are keen to participate, particularly since we start with a big pot of royalties to distribute to them. The Noank models does require service providers to enter into agreements with us and pay for all their subscribers - a major difference between our service and most others. So far, our offerings to service providers have proven very attractive to them, which is why we're negotiating and closing deals much more rapidly than expected. lastly, we are committed to treating end-users much better than any other service that we know of. They are at the center of our service plans. The Noank idea began as a cooperative between content users and suppliers and we have retained that spirit throughout the inevitable fine-tunings that real-world commercial ventures inevitably encounter.