The BSA Global Software Piracy Study quantifies the volume and value of unlicensed software installed on personal computers in a given year — in this case, 2011. To compile the report, BSA works closely with two of the world's leading independent research firms — IDC and Ipsos Public Affairs — to measure, understand, and evaluate global software piracy.
The study involves collecting 182 discrete data inputs and assessing PC and software trends in each of 116 markets. A detailed video presentation of the methodology can be found at www.bsa.org/globalstudy.
Measuring the scale and scope of illegal behavior like software piracy clearly has its challenges. While this annual study is considered to be one of the most sophisticated appraisals of piracy produced each year, BSA and its partners continually look for new ways to improve the reliability of the data. This year, in partnership with two prominent IT economic researchers, BSA has made several modifications designed to refine the inputs and ensure the most accurate estimation of software piracy possible.
"A Rigorous and Well-Designed Effort "
In 2010, we were retained by BSA to provide an independent assessment of the methodology used in the BSA Global Software Piracy Study and offer recommendations for improvements.
At the outset, it is worth noting that studies of piracy examine illegal or unreported behavior, making it difficult to obtain perfect estimates. As such, accurate estimation of the software piracy rate requires good data inputs and a rigorous well-designed process to ensure that the results present as accurate a picture as possible of both the existing and the changing nature of software piracy over time. Based on our review, we believe the BSA Global Piracy Study has these elements in place.
In our review process, we examined documents provided by BSA that described the methodology used in the 2010 study. We also interviewed key people at IDC and Ipsos. While we did not review the accuracy of the data provided by IDC or Ipsos, IDC's Worldwide PC Tracker, a key source of data, is widely recognized by the IT industry as the best available on the market. Likewise, Ipsos has a strong reputation for high-quality surveys. It should be noted that in 2010, the end-user survey substantially increased the number of respondents, which has improved the quality of the data.
In addition to data quality, the results of the study are determined by the appropriateness of the computational procedure. The methodology must account for different software categories and the associated prices, the average value of software across different channels, installed software variability by country, and so on.
In this regard, the study utilizes standard approaches of estimating secondary data inputs. In order to reduce the level of uncertainty, a more subjective approach could introduce to year-on-year results.
It is well understood in the academic research community that any study methodology introduces a margin of error. The challenge for researchers is to design a study that keeps the error within an acceptable margin. As part of this undertaking with BSA, we provided several recommendations for improvement, which should enhance the accuracy of the results of this year's study. These recommendations are described in more detail in the methodology of the white paper.
However, in our assessment, the BSA Global Software Piracy Study methodology represents a rigorous and well-designed effort to take into account the key features of the software market and to measure the level of activity occurring in what is essentially a black market. We believe the study provides reliable estimates of piracy rates and the commercial value of pirated software, particularly in larger markets.
— Professors Vijay Gurbaxani & Vidyanand Choudhary University of California, Irvine
Vijay Gurbaxani and Vidyanand Choudhary are Taco Bell Professor of Information Systems and Associate Professor of Information Systems at the Paul Merage School of Business, University of California, Irvine. Their views and opinions expressed in this report are strictly theirs alone and do not necessarily reflect the views and opinions of the University of California.
Global Survey of Software Users
A key component of the BSA Global Software Piracy Study is a global survey of software users, led by the highly regarded research firm Ipsos Public Affairs. The survey was conducted among 14,700 business and consumer PC users, online or in-person, in 33 markets that make up a globally representative sample of geographies, levels of IT sophistication, and geographic and cultural diversity.
The survey is used, in part, to determine the "software load" for each country — that is, a picture of the number of software programs installed per PC, including commercial, open-source and mixedsource programs. Respondents are asked how many software packages, and of what type, were installed on their PC in the previous year; what percentage were new or upgrades; whether they came with the computers or not; and whether they were installed on a new computer or one acquired prior to 2011.
In addition, the surveys are used to assess key social attitudes and behaviors related to intellectual property, software piracy, and other emerging technology issues. This insight provides fresh perspective each year on the dynamics underlying software piracy around the world.
Among the improvements to the study this year is the development of a rotational strategy for the survey to enhance country coverage year upon year. Going forward, 11 countries will be surveyed annually, and 42 countries will be surveyed at least once every two to three years. The remainder will be surveyed on an ad hoc basis, generally one a year. As has been the case, the total survey population in any given year will account for approximately 90 percent of the PC
Calculating Software Piracy Rate
Since 2003, BSA has worked with IDC, the leading provider of market statistics and forecasts to the IT industry, to determine software piracy rates and the commercial value of pirated software.
- The basic method for coming up with the piracy rate and commercial value of unlicensed software in a country is as follows:
- Determine how much PC software was deployed during the year.
- Determine how much was paid for or otherwise legally acquired during the year.
Subtract one from the other to get the amount of unlicensed software. Once the amount of unlicensed software is known, the PC software piracy rate is computed as a percentage of total software installed.
To calculate the total number of software units installed — the denominator — IDC determines how many computers there are in a country and how many of those received software during the year. IDC tracks this information quarterly in 105 countries, either in products called "PC Trackers" or as part of custom assignments. The remaining few countries are researched annually for this study.
Once IDC has determined how many computers there are, and using the software load data collected in the survey, it can determine the total software units installed — licensed and unlicensed — in each country.
To estimate the software load in countries not surveyed, IDC uses a series of correlations between the known software loads from surveyed countries and their scores on an emerging market measure published by the International Telecommunications Union, called the ICT Development Index. These correlations are used to group the countries into cohorts, wherein the average software load of surveyed countries in each cohort is applied as the estimated software load for non-surveyed countries in that cohort. In 2011, IDC expanded the number of cohorts from four to six in order to provide more precision in applying surveyed values to similar countries.
To get the number of unlicensed software units — the numerator of the piracy equation — IDC must determine the value of the legally acquired software market. IDC routinely publishes software market data from about 80 countries and studies roughly 20 more on a custom basis. For the few remaining countries, IDC conducts annual research for the purposes of this study. This research provides the value of the legally acquired software market.
To convert the software market value to number of units, IDC computes an average price per software unit for all of the PC software in the country. This is done by developing a country specific matrix of software prices — such as retail, volume-license, OEM, free, and open-source — across a matrix of products, including security, office automation, operating systems, and more.
IDC's pricing information comes from its pricing trackers and from local analysts' research. The weightings — OEM versus retail, consumer versus business — are taken from IDC surveys. IDC multiplies the two matrices to get a final, blendedaverage software unit price.
To arrive at the total number of legitimate software units, IDC applies this formula:
In 2011, IDC implemented several measures to validate its calculations of average software unit price. Analyst teams in 25 countries have been asked to provide additional information on software price by category and estimates of acquisition type (e.g. retail, volume-license, free/open source) to serve as a cross-check against IDC's computed values. Rotating the countries for which information is collected each year will allow IDC to recalibrate software prices periodically, providing a more accurate estimate of legitimate software units from industry revenues.
Finally, subtracting the number of legitimate software units from the total software units reveals the number of unlicensed software units installed during the year.
This process provides the underlying data for the basic piracy rate equation.
Calculating the Commercial Value of Pirated Software
The commercial value of pirated software is the value of unlicensed software installed in a given year, as if it had been sold in the market. It provides another measure of the scale of software piracy and allows for important year-over-year comparisons of changes in the software piracy landscape.
It is calculated using the same blend of prices by which IDC determines the average software unit price, including: retail, volume license, OEM, free, open-source, etc. The average software unit price is lower than retail prices one would find in stores.
Having calculated the total units of software installed, as well as the number of legitimate and unlicensed software units installed and the average price per software unit, IDC is able to calculate the commercial value of unlicensed software.
Which Software Is Included
The BSA Global Software Piracy Study calculates piracy of all software that runs on personal computers — including desktops, laptops, and ultra-portables, including netbooks.
It includes operating systems, systems software such as databases and security packages, business applications, and consumer applications such as games, personal finance, and reference software. The study also takes into account the availability of legitimate, free software and open-source software, which is software that is licensed in a way that puts it into the public domain for common use. It is typically free but can also be used in commercial products.
The study excludes software that runs on servers or mainframes and routine device drivers, as well as free downloadable utilities, such as screen savers, that would not displace paid-for software or normally be recognized by a user as a software program.
It includes software as a service if it is paid for, but excludes free, Web-based services that might supplant the need for a paid-for package to be installed on a PC. Software sold as part of a legalization program — such as a bulk sale to a government to distribute to schools — is included in the study.
The Impact of Exchange Rates
From 2003 through 2008, US dollar figures in the value tables were in current dollars from the year before. For example, the 2006 value of unlicensed software was stated in 2005 dollars; the value of 2007 unlicensed software was stated in 2006 dollars, and so on. In 2009, BSA decided to publish value figures in the current dollars of the year being studied — 2009 values are in 2009 dollars, 2010 values in 2010 dollars, and 2011 values in 2011 dollars.
This is important when evaluating changes in the values over time. Some of the changes will be based on real market dynamics, some on exchange rate fluctuations from year to year.