DATA FROM SME e n r0 D,

riise---but don! t tellta otem jr]blRIAA. A year after i, k- t r

Suprene Cow t, tb e 7111' i son thI me saeta .ea :.eswp n8 f

while by no Iean dead i fltter 2The RIAA hsbe di:n noew.

te that staisti cs su petr• die a. piayi

~AA~ (i. in r I )M c it B ti

up ) th agne. Ii Iana"wenh rt

1dca tracking it. CambIdge In aa heIoic-A d Co a t es , Q 11111-k,b a e 4-yS, l o tr c i a c y

leves bsed1In networkusg aid yild resu111tssI 1a to B'gCha agn. Ot , reseaxch. firrins traft

fin ii ,i ih( H H

Vill lisi 1Ow_

di �P I DJ i

� 4 1 CS I

J11101. 1 1 III S

Lii I 'I'.

I (5K

1 (1 115

) dat 'eta. ng to piracy--in-P C~om Score Media Metrix,

Resea(hi--. 'thiey do.r iot offerbe1f 121 do10wnloaders or the

(O Eric Garland liketis the dSo2P saist tt pvrabS

P t e elehat Ther :is ' o 1 co1 1

a sh f�ie II e'- OL H

l( N�t rn 10

4 in ibo it �K e

'111 1 (C�S illa�n I Jr

1 11 1 dl ej s.

s�ir H

Ss 1�II ittIJ'� 1 �II1I

(11 V h-� p

V

nba'�i r IIi ( c'O

I 1'? 1

11 1 �,

I I (I

I �) �'

e I

It

IA I

( 1�

5- Ic' I

it I)

8 1 hIi 1 iW� i n

) 111'

>i�)i i. V pii. tiP �1:

oi

� I1'

H

Ke•

I I

h H ItIi coun s TI'io ere is

It at' nt u ea 1)e 0o1,Jt 1',g aoutan !ecior," ays R s

d I I SC it ILI� pOis s

tits I Ni

we've managed to find a way to make it work withphysical piracy, there is a level of Internet piracy wecan survive.

"We're going to continue to drive down digitalpiracy," Bainwol continues. "But my fundamentalpoint has been that over the last couple of years dig-ital piracy has stabilized when it comes to users."

So has digital piracy really stopped growing at asignificant rate?

The mixed signals coming from P2P trackersnot sourced by the RIAA have fired up a debateover which data providers offer the best informa-tion about piracy rates-and how their data shouldbe interpreted.

When it comes to hard numbers that can be re-lated to online piracy, the two most-quoted sourcesfor P2P data are NPD, a Port Washington, N.Y.-basedresearch firm that has long conducted consumer stud-ies on behalf of the music industry, and Beverly Hills,Calif.-based BigChampagne, a specialist in trackingonline buzz that monitors music usage on P2P net-works and other nontraditional outlets.

They are not alone in monitoring P2P, but theyrepresent the leading schools of methodology for

�,t � I S

puter users, in much the same way Nielsen MediaResearch tracks TV viewership. A volunteer groupof 12,000 Internet users allow NPD to keep tabs oneverything they do online. In the process, NPD tracksdownloading behaviors of users in commercial andunauthorized file-sharing environments and thenprojects its findings to create an estimate of Inter-net users at large. The company has the ability totrack usage by region and file type as well as moni-tor post-download playback trends. NPD also doesusage-behavior surveys to relate consumer attitudesto its tracking data.

By contrast, BigChampagne monitors usage onthe network level, using propriety technology to trackthe number of users logged into a file-sharing net-work at any given time. The company tracks the twobasic activities that can be monitored on P2P net-works: "queries" (searches) and "acquisitions" (down-loads). Then they match a computer's IP address toits ZIP code, creating a map of P2P activity.

Garland says BigChampagne's technology can iso-late the use of any form of copyrighted material, frommusic, feature films, software and videogames to in-structional manuals or TV episodes.

soil

aYS

I

FILE TRADING TRENDSAre the number of P2P households growing or stabilizing? NPD and BigChampagne differ on the pace of adoption during thelast two years, in part because they use different methodology, NPD (below left) measures active downloaders based on asample of 12,000 Internet users, while BigChampagne (below right) monitors average simultaneous users of P2P networks,

The company, which has been in business since 2000, sellsthe data to record labels, marketers and other entertainmentcompanies. But Garland maintains BigChampagne's missionis to help determine the popularity of songs online-not totrack piracy.

From these very different measurement approaches comevery different snapshots of the pirate market. Both firms agreethat P2P usage continues to climb. But they differ on the paceof adoption.

Direct comparison of NPD and BigChampagne data isproblematic.

NPD reports that the number of U.S. households activelydownloading music from P2P sites has grown 14.8% over thelast two years, rising from an estimated 6 million householdsin April 2004 to 6.9 million in April 2006.

But as a percentage of the overall Internet population, P2Pgrowth has slowed dramatically, according to NPD. It esti-mates that active P2P households represent 10.2% of the cur-rent U.S. Internet population, an increase of just onepercentage point over the 9.1% of active P2P households in2004. Over the same period, NPD notes, the percentage ofhomes connecting to the Internet has more than doubled, theamount of storage space on PCs has grown markedly and thenumber of iPods and other MP3 players has exploded.

In terms of music use, what is growing significantly, NPDsays, is the number of files being traded. More than 333 millionfiles were swapped in April 2006, up 54% from an estimated 216million files in April 2004. That stat suggests an entrenchedgroup of hardcore users increasing their consumption.

BigChampagne claims that growth rates for P2P users aremuch higher. It estimates that the average number of simul-taneous P2P users in March 2006 was 7 million, up 49% fromthe 4.6 million users in April 2004. It also says that about 1.5billion songs, 70% of the files offered through P2P networks,are available for download at any given time-although it doesnot project the actual number of downloads.

Stan Liebowitz, a professor of economics at the Universityof Texas at Dallas' School of Management, has been closelymonitoring the reports of P2P tracking firms. He says thateach data source, regardless of methodology, suffers from oneor more deficiencies.

In an April 2006 study titled "File-Sharing: Creative De-struction or just Plain Destruction?" Liebowitz breaks downthe flaws of the data trackers. Panel-based data sources likeNPD might underrepresent the population of people actively

HOUSEHOLDS USING P2P(In Thousands)

7

5

'17h: 711O51.

SIMULTANEOUS P2P USERS(Average, In Millions, US)

47/04 1/05 7/05 1/06 7/04

iw: 4,435,395

1/05

High: 7,002,830,

7/05 1/06

Below: NPD reports that the percentage of Internet households using P2P has grown minimally between April 2004 andApril 2006. But active users of P2P are downloading more music than ever before.

INTERNET USERSDOWNLOADINGFROM P2P (Percent)

11

10

87/04_ 1/05

engaged in file sharing, Liebowitz argues. He says the mostactive users may be "particularly reluctant" to have their com-puters monitored by third-party software. (Crupnick countersthat the NPD panel includes heavy downloaders.)

As for BigChampagne's method, Liebowitz says metrics basedon the number of users alone might fail to capture increases ordecreases in the number of files exchanged per user. He also ar-gues that by tracking simultaneous users, BigChampagne maybe double-counting some who are logged onto P2P networks forlong stretches of time. (Garland says BigChampagne weeds outrepeat individual users when determining its figures.)

So NPD or BigChampagne? The answer hinges on a secondquestion: What's actually growing, the number of files beingtraded or the number of users trading files?

Therein lies the disparity between NPD data, which indicates

NUMBER OF SONGS P2P DOWNLOADSDOWNLOADED FROM PER HOUSEHOLDP2P (Not Unique, In Thousands) (Average Number of Songs)

file growth, and BigChampagne data, which says P2P house-holds are expanding.

"We're in agreement there is a lot of P2P activity," Crupnicksays. "But BigChampagne suggests the cancer is growing andgrowing. We don't think that's the right diagnosis. We believeours is a more accurate diagnosis of what the disease is."

Regardless of who is painting the more accurate picture,the RIAA says the rise in adoption of legitimate offerings likeiTunes and mobile music is proof that digital piracy is at amore manageable level.

"Our focus is on whether or not digital piracy is so rampantit's taking the oxygen out of the legal marketplace, and what weare seeing is that's not the case," Bainwol says. "We're movingin the right direction and that's a function of the stabilizationof users on the P2P side." .

P2P BY THE BOOKPrivate resea chers are not the only ones t yiing to make sense of P2P piracy. Academia is also exploring the P2P phenomenon, with a particular eyetoward the behavior's effect-if any--on sales, Blilboard offers some quick takes from the leading studies on file sharing.-BG

2004Authors: Felix Oberholzer, HarvardBusiness School; and KolemanStrumpf, University of North Car-olina-Chapel HillStudy: "The Effect of File Sharingon Record Sales: An EmpiricalAnalysis"Bottom Line: "Downloads have aneffect on sales, which is statisticallyindistinguishable from zero despiterather precise estimates. Moreover,these estimates are of moderateeconomic significance and are

inconsistent with claims that filesharing is the primary reason for therecent decline in music sales."

2004Author: David Blackburn, HarvardUniversityStudy: "On-line Piracy and RecordedMusic Sales"Bottom Line: "File sharing reducessales for well-known artists relativeto unknown artists . . Lawsuitsbrought by the RIAA have resulted inan increase in album sales of ap-proximately 2.9% during the 23-week period after the lawsuit strat-egy was publicly announced.Furthermore, if files available onlinewere reduced across the board by30%, industry sales would have beenapproximately 10% higher In 2003."

2005Author: Norbert Michel, NichollsState UniversityStudy: "A Theoretical and EmpiricalAnalysis of the Impact of the DigitalAge on the Music Industry"Bottom Line: "Despite major labels'advantage in large-scale distribu-tion, we argue that digital down-loading has the potential to radicallyalter the current Industry structure,and that artists would be unable tosell their music in such an environ-ment without enforceable copy-rights ... We hypothesize thatInternet file-sharing has been un-dertaken by both consumers whowere previously not in the market,and by those who decided to copy

rather than buy."

2006Authors: Thomas Karagiannis, Uni-versity of California, Riverside;Andre Broido, Nevil Brownlee andKC Claffy, Cooperative Assn, for In-ternet Data Analysis-CAIDA SanDiego Supercomputer Center, Uni-versity of California, San Diego; andMichalis Faloutsos, University of Cal-ifornia, RiversideStudy: "Is P2P Dying or Just Hiding?"Bottom Line: "Recent reports in thepopular media suggest a significantdecrease in peer-to-peer (P2P) file-sharing traffic, attributed to the pub-lic's response to legal threats. Ifmeasured accurately, P2P traffic hasnever declined; indeed we havenever seen the proportion of P2Ptraffic decrease over time (anychange is an increase) in any of ourdata sources."

2006Author: Stan J. Liebowitz, Univer-sity of Texas at Dallas, School ofManagementStudy: "File-Sharing: Creative De-struction or Just Plain Destruction?"Bottom Line: "Although file-sharinghas been imperfectly and inconsis-tently measured, [information] nev-ertheless appears to reveal a fairlyclose linkage between changes Infile-sharing and changes in recordsales, Explanations, other than file-sharing, for the recent decline inrecord sales seem to have little orno support."

26 I JULY 1, 2006

500

7/05 1/06 7/04 1/05 7/05 1/06

Z

0