UNITED STATES DISTRICT COURT SOUTHERN DISTRICT OF NEW YORK
June 18, 2008
UNITED STATES OF AMERICA, PLAINTIFF,
AMERICAN SOCIETY OF COMPOSERS, AUTHORS AND PUBLISHERS, ET AL., DEFENDANTS.
The opinion of the court was delivered by: Conner, Sr. D. J.
OPINION AND ORDER
In the Matter of the Application of AMERICA ONLINE, INC., Applicant,
: for the Determination of Reasonable License Fees
In the Matter of the Application of REALNETWORKS, INC., Applicant,
: for the Determination of Reasonable License Fees.
In the Matter of the Application of YAHOO! INC., Applicant,
: for the Determination of Reasonable License Fees.
On April 30, 2008, this Court, in its capacity as the "rate court" under Section IX of the Second Amended Final Judgment in United States v. ASCAP, 2001 WL 1589999 (S.D.N.Y. 2001), rendered a 153-page Opinion and Order determining reasonable fees for blanket licenses for the unlimited performance of all the ASCAP-repertory music by three internet proprietors, AOL LLC ("AOL"), Real Networks, Inc. ("Real Networks") and Yahoo! Inc. ("Yahoo!") (collectively "Applicants") on their respective internet websites. The Court set, for each of the three Applicants, a fee determined by multiplying the total revenue of the licensed business unit, less customary deductions for advertising sales commissions and traffic acquisition costs, by a music-use-adjustment fraction whose numerator is the total number of hours music is streamed to users by the licensee (as measured by the Applicant itself and conceded to be accurate) and whose denominator is the total number of hours of use of the licensee's website (as measured by comScore or other means approved by the Court) and applying to the resulting music-use-adjusted revenue a fee rate of 2.5%.
The Court found that, although comScore tracks total site hours by a different method than Applicants use for tracking music streaming hours, the comScore metric is the most accurate available measure of total hours of site use and, because comScore's reported site-hour figures are substantially greater than those of its rival media metrics service, Neilsen, use of the comScore data is conservative and would result in substantially lower license fees (Opinion and Order at 149, n.10). However, in recognition of the possibility that a more accurate measure of total site hours might be, or might later become, available, the Court added: "If Applicants can establish in a supplemental rate proceeding that such a figure is available, it may be used in computing the license fees." (Id. at 146).
At a court conference on May 12, 2008, Applicants informally requested a supplemental rate proceeding to establish that they already had available, although they had not previously mentioned, a more accurate measure of total site hours than that provided by comScore, which should be used in the denominator of the music-use-adjustment fraction. The Court granted Applicants leave to file a motion for a supplemental rate proceeding but stressed that the motion should present a detailed offer of proof: "I want you to tell me what you would expect to prove at the supplemental rate proceeding and how you expect to prove it." (Tr. at 36).
Applicants' motion was filed May 27, 2008 and was fully briefed by June 12, 2008. In their motion, Applicants offer to prove the following facts (condensed to essentials, followed by the Court's comments on the significance of each fact):
1. comScore collects data not from every user of the website but from a "representative" sample of users or "panels" and extrapolates therefrom to the universe of users. This is a commonly accepted statistical practice and, although Applicants were aware that it was being employed by comScore, it did not cause them to cease paying substantial fees for the comScore metrics, praising their accuracy and using them in computing advertising rates.Until Applicants are willing to pay the higher cost of collecting data from every user,thecomScore metrics are the most accurate measure of total site time known to the Court.
2. comScore stops counting the time of user engagement with a webpage when another page or window is opened over that webpage. This assertion is doubly specious. First, when another page or window is opened on top of a webpage, the user's attention is obviously focused on the overlying page; indeed it is likely that only the outer margins of the original page remain visible.
To count page time not only for the original page but also for each of the overlying pages would be a ridiculous example of double or multiple counting. Second, and even more important, the music-use-adjustment fraction employed by the Court is not based on page time but on overall site time.
Even if comScore improperly counts page time, that would not affect the accuracy of its count of site time, which is the aggregate time users are visiting the site, regardless of whether each of them has only one or any number of windows opened at the same time.
3. comScore "caps" the time spent on a webpage at two minutes. This is another utterly irrelevant fact. As previously noted, the Court's music-use-adjustment fraction is not based on page time but on total site time, and is not affected by any error involved in capping time spent on individual pages.
4. comScore stops counting site time when a user does not react with the site for 30 minutes. When a user has had no interaction whatever with a site for half an hour, it almost surely indicates that the user is otherwise occupied. Indeed, a strong argument could be made for capping the measurement of site time after less than 30 minutes of inactivity. Continuing to count site time for such a long period of inactivity probably results in some overcounting of the time the user's attention is actually directed to the website, which reduces the licensing fee.
5. comScore counts time only on websites specifically identified by AOL and Yahoo!
It would be surprising if comScore compiles metrics that are not ordered and paid for by a client. If, as the Court proposes, the Applicants' license fees are based in part on total site time, they will surely instruct comScore to track user time on all their licensed websites and thereby minimize their fees.
6. comScore does not measure streaming activity. This is another utterly irrelevant fact.
The only comScore metric that affects the license fee is total site time. Streaming time is tracked by Applicants themselves and the accuracy of their measurements is stipulated.
7. comScore does not include website activity accessed through mobile devices such as cellphones and Blackberries or through the MAC OS or Linux operating systems. There is no evidence, either in the present record or proffered for a supplemental hearing, that a significant number of users are accessing Applicants' websites through cellphones and Blackberries or through one of the alternative operating systems. Apparently Applicants do not consider such activity sufficiently important to track it and factor it into their advertising rates. The Court has been given no reason to believe that taking it into consideration would significantly affect the license fees.
In sum, Applicants have failed to show that, at a supplemental hearing, they could adduce evidence to establish that the Court's use of comScore's measurement of total site time as the denominator of the music-use adjustment fraction was improper. Applicants' attempt to show that the "internal measurement databases" of Yahoo! and AOL are more reliable than the comScore metrics consists essentially of assertions that Yahoo! and AOL do all the things that comScore improperly fails to do, and do not do any of the things that comScore improperly does, as set forth in paragraphs 1 through 7 above. This argument is puzzling because Applicants do not track total site time at all but rely on outside services such as comScore to do so. Moreover, as explained above, all of Applicants' assertions of error in comScore's methodology are either specious, irrelevant or of indeterminate significance. For example, Applicant's cumulative counting of page time not only on an originally visited page but on all the overlying pages or windows results in obvious double or multiple counting. And, in any event, the count of page time has no effect on the music-use-adjustment fraction, whose denominator is overall site time, which is counted without regard to the number of pages users might have open simultaneously.
For all the foregoing reasons, Applicants' motion for a supplemental rate hearing is denied. The parties are directed to cooperate in expedited preparation of a final judgment effectuating this Court's Opinion and Order of April 30, 2008.
© 1992-2008 VersusLaw Inc.