![[INET'99]](/web/20080604122229im_/http://www.isoc.org/isoc/conferences/inet/99/proceedings/inet99sm.gif)
![[ Up ]](/web/20080604122229im_/http://www.isoc.org/isoc/conferences/inet/99/proceedings/up.gif)
![[Prev]](/web/20080604122229im_/http://www.isoc.org/isoc/conferences/inet/99/proceedings/prev.gif)
![[Next]](/web/20080604122229im_/http://www.isoc.org/isoc/conferences/inet/99/proceedings/next.gif)
Paul PIERLOT <pierlot.paul@ic.gc.ca>
Industry Canada
Canada
The paper aims to increase the understanding among industry, government, and the public of the challenges that Internet service providers (ISPs) face in supplying and using electronic commerce (e-commerce), and the myriad of ways they are addressing these challenges and seizing the opportunities of the digital economy.
The results provide a rigorous understanding of the nature of ISP activities in Canada, with a focus on e-commerce and other value-added services, making a significant contribution to knowledge on many relevant fronts. Key areas examined include e-commerce and other value-added services, economic performance, revenue sources, customer distribution by business and consumers, privacy and security issues, self-regulatory initiatives, and practices in combating offensive content and offensive conduct such as spam and hacking.
Building on the widespread success of the 1997 study, this second, comprehensive, national survey of ISPs comprises a very important body of data on a leading-edge industry with an integral role to play in the rollout of electronic commerce. To our knowledge, this is the only statistical work of this magnitude in the world, and it covers an industry about which relatively little is known.
This paper follows up on the Industry Canada paper "Internet Service Providers in Canada: An Economic Analysis" presented at INET'98. The INET'98 presentation was very well attended and well received, and conference participants expressed great interest in seeing the results updated. This year's study updates those results and greatly increases the focus on ISPs' role in enabling e-commerce and connecting Canadians to the digital economy.
The data upon which this analysis is based come from the Survey of ISPs and Computer Services 1997, conducted by Statistics Canada (Canada's official bureau of statistics) on behalf of Industry Canada (the Canadian Federal Department of Industry). The survey was based on a 675-firm sample, and an overwhelming 400 responses were received. This response represents over four times that of the previous survey and covers more ISPs than were thought to exist in Canada just a few years ago. In fact, the Industry Canada research revealed evidence contrary to the popular notion that industry consolidation is reducing the number of ISPs in Canada. Market entry appears to be more than offsetting the effects of consolidation.
The data profile one of Canada's newest industries, a subset of the online information services industry, and include information on e-commerce activities, barriers to growth, content provision, privacy and security of information, and practices to deal with offensive content and conduct. The survey covers a wide range of information, including sources of revenue, market segmentation, product and service offerings, employment, research and development activities, and business practices related to ISP industry self-regulation.
The data will be used to provide much-needed information about this dynamic industry. The survey was conducted in order to provide the ISP industry with intelligence for market analysis, performance assessment, and trend identification. The survey also aims to provide government with intelligence for informed policy making, education, and awareness building.
The paper will have broad-based appeal. It builds upon, and provides an update to, the popular 1997 paper on Canadian ISPs. Based on input from conference participants as well as recent industry consultations, the current paper makes a number of improvements and is strengthened in areas highlighted as being of greatest importance. It greatly increases the focus on e-commerce and other value-added services and looks in detail at ISPs self-regulatory efforts in addressing public concerns surrounding offensive content and offensive behavior such as spam. It will be relevant to ISPs, policy makers, regulators, consumers, and citizens interested in Internet issues.
Internet service providers, or ISPs, are playing an increasingly important role in Canada's transition to a knowledge-based, networked economy. These service providers are connecting Canadian businesses, consumers, and citizens to the global, and increasingly digital, economy and are helping to provide the means by which electronic commerce can be carried out across the country and around the world.
Based on the findings of a recently conducted, national survey of the ISP industry in Canada, this study is part of an ongoing effort to better understand this rapidly emerging sector. Building on the widespread success of the 1995 national survey, Industry Canada (the federal Ministry of Industry) and Statistics Canada (the official statistics agency) again collaborated on a comprehensive survey of the industry, this time increasing the focus on the provision of electronic commerce services and matters of public concern, such as those pertaining to offensive Internet content. The survey has provided industry, government, and other stakeholders with a rich statistical profile of the industry that can be used for business planning, economic performance benchmarking, and public policy development.
Statistics Canada collected the data upon which this study is based between June and July 1998. The sample frame of 675 ISPs -- where ISPs were defined as organizations that, as a minimum requirement, provide Internet access to the public -- was compiled by Industry Canada using existing company data as well as online ISP directories. By the end of the collection period, almost 400 surveys were returned, representing a response rate of 60 percent. This response was considered to be a major success, especially given that it was nearly triple the response of the last Industry Canada ISP survey. As well, the research conducted to compile the sample frame provided much evidence that the well-publicized mergers and acquisitions in the Canadian ISP industry have been accompanied by high rates of market entry.
It is important to note that the survey does not represent an industry census and is not necessarily statistically representative of the industry. Nevertheless, it provides us with a wealth of information on the distribution of ISPs across Canada by size and by business activity, the types of services ISPs provide, how successful ISPs are, and how ISPs are approaching issues related to consumer confidence and public policy.
ISPs from every Canadian province and territory were represented in the survey. In order to conduct an analysis of regional variation, and to align the results with other Industry Canada studies, respondents were grouped along regional lines. This grouping explicitly assumes that similar market conditions are faced by individual ISPs that are in the same region. As illustrated in exhibit 3.1 (from west to east), 25 percent of respondents are located in British Columbia (BC) and the Territories; 14 percent in the Prairie provinces; 41 percent in Ontario; 14 percent in Quebec; and 6 percent in the Atlantic provinces. Although the actual distribution of ISPs across Canada is not known with certainty, the distribution of respondents is believed to be reasonably representative of the true distribution.
As noted in the methodology section, to be considered an ISP for this survey, a firm had to provide Internet access to the public as a minimum requirement. However, many firms that offer ISP access services do not derive the majority of their operating revenue from ISP activities. In fact, an ISP's primary line of business may be the provision of cable television services, software development, retailing, consulting, or even transportation. The evolving nature of ISPs makes it important to consider both overall firm size and level of ISP-based activity when analyzing the economic characteristics and performance of the industry.
The revenue figures presented in exhibit 4.1 reflect the total operating revenue that firms generate from all of their corporate activities, including their ISP business units. As such, they serve as a measure of overall firm size, rather than firm size in the ISP market specifically.
| Under $50K | 17% |
|---|---|
| $50K-$99K | 15% |
| $100K-$199K | 17% |
| $200K-$499K | 19% |
| $500K-$2.49M | 22% |
| $2.5M or More | 10% |
This grouping represents a relatively even distribution, with about one-third of firms generating less than $100,000 in total operating revenue; just over one-third earning between $100,000 and $500,000; and the remaining third generating over $500,000 (all currency quoted in $CDN).
For analytical purposes, respondents were grouped according to their revenue, as follows: micro firms: total revenue from operations under $50,000; small firms: total revenue from operations between $50,000 and $249,999; medium firms: total revenue from operations between $250,000 and $999,999; large firms: total revenue from operations of $1 million or more.
As shown in exhibit 4.2, the majority (53 percent) of respondents fall into the small and micro firm categories. Specifically, 17 percent are micro firms; 36 percent small firms; 26 percent medium firms; and 21 percent large firms.
The survey results showed that large firms make an extremely large contribution to the national operating revenue figure. Large firms (those with operating revenues of $1 million or more) make up approximately one-fifth of the respondent base, yet they account for 96.5 percent of total operating revenues. Moreover, when the 10 largest respondents are removed, total operating revenues in the large firm category fall from $1.88 billion to $200 million. Combined with the distribution reported in section 4.2, this finding indicates that the Canadian ISP industry is characterized by the coexistence of many small firms and a small number of large, dominant firms.
Average revenue by firm size also varies quite dramatically across the industry: The average revenue of micro firms is $24,000; small firms $130,000; medium firms $500,000; and large firms $23.5 million (note that the average revenue across all respondents is approximately $5.1 million).
In an effort to increase the understanding of industry structure and firm characteristics, respondents were asked to report the proportion of operating revenues they derived from ISP activities ("Percent ISP"). The results were as wide-ranging as possible, with some firms stating that ISP services accounted for less than 1 percent of their operating revenue, and others for as much as 100 percent. Specifically, 19 percent of respondents indicated that ISP activities account for 0.1 to 49.9 percent of their total revenue; 13 percent indicated 50 to 74.9 percent; 31 percent indicated 75 to 99.9 percent; and 37 percent indicated 100 percent.
Smaller firms tend to be more dependent than larger firms on the revenues they derive from ISP activities. In fact, smaller firms are more likely to fall into the "75 to 99.9 Percent ISP" or the "100 Percent ISP" categories, while larger firms are more likely to be in the "Less than 50 Percent ISP" category. These findings indicate that smaller firms are more likely to rely on ISP activities as their primary source of operating revenue.
Accordingly, many of the largest respondents in terms of operating revenue fall into the smallest Percent ISP category. In fact, firms in the "Less than 50 Percent" category account for 87 percent of total operating revenues. In other words, on average, large firms rely on their ISP business units for less than 50 percent of their operating revenue (telecommunication carriers, for example, derive the majority of their revenue from telecommunications services).
A similar pattern emerges when average operating revenue is examined: Respondents in the 100 Percent ISP category have average firm revenue of $0.9 million; in the 75 to 99.9 Percent ISP category, an average of $0.7 million; in the 50 to 74.9 Percent ISP category, an average of $0.6 million; and in the 0.1 to 49.9 Percent ISP category, an average of $24.3 million.
The "Less than 50 Percent ISP" category is heavily influenced by the presence of the largest respondents. When the 10 largest respondents are removed, the following averages result: Less than 50 Percent ISP: $1.5 million; 50 to 74.9 Percent ISP: $0.6 million; 75 to 99.9 Percent ISP: $0.65 million; and 100 Percent ISP: $0.5 million.
Exhibit 6.1 presents a detailed breakdown of the distribution of respondents according to the ISP revenue they generate from the provision of ISP services.
| $50K or Less | 22% |
|---|---|
| $50K-$99K | 20% |
| $100K-$249K | 22% |
| $250K-$499K | 14% |
| $500K-$749K | 7% |
| $750K-$999K | 4% |
| $1M-$2.49M | 6% |
| $2.5M-$4.99M | 2% |
| $5M or More | 3% |
Note: K = thousands, M = millions.
Nearly two-thirds of respondents generated less than $250,000 in ISP revenue in 1997, while 11 percent generated $1 million or more. This is consistent with the findings in section 4.1, where approximately two-thirds of respondents generated less than $500,000 in total operating revenue, and 10 percent generated $2.5 million or more.
For analytical purposes, respondents were grouped into three ISP revenue categories, as noted below. The distribution of respondents according to this classification is presented in exhibit 6.2.
This particular grouping fit the distribution of respondents well, but has the disadvantage of grouping several very large firms with arguably medium-sized, or even small firms, in the high ISP revenue category. Future analysis will consider the very large firms in further detail.
Respondents were asked to report the total number of ISP customers they have -- both business and residential. A total of 1,219,136 were reported, representing a large proportion of Canadian Internet users. The distribution of respondents by the size of their customer base is shown in exhibit 7.1. Note that more than one-half of respondents reported having fewer than 1000 customers, while 8.5 percent reported having more than 10,000 customers.
| Less than 250 | 23.0% |
|---|---|
| 250-999 | 31.0% |
| 1000-1499 | 9.5% |
| 1500-2499 | 12.0% |
| 2500-9999 | 16.0% |
| 10,000 or More | 8.5% |
Average annual ISP revenue across all respondents is $167. As indicated by exhibit 7.2, there is an apparent positive relationship between firm size and ISP revenue per customer. It may be the case that large firms (those that generate $1 million or more in corporate operating revenues) sell more value-added services than smaller respondents do. Alternatively, large firms may have more business customers than smaller firms do, and they may charge higher rates or offer more services to these business customers. This relationship also holds when firm size is measured by ISP revenue.
| Micro | $20.00 |
|---|---|
| Small | $96.00 |
| Medium | $152.00 |
| Large | $190.00 |
| All | $167.00 |
Respondents were asked to estimate the proportion of ISP revenue that they derived from their business and residential subscribers. Not surprisingly, most respondents cited residential subscribers as being their primary source of revenue, accounting for nearly 60 percent of total ISP revenue reported. On average, revenue from business subscribers constituted almost 35 percent of respondents' revenue. Other sources accounted for just over 5 percent of total ISP revenue (fees for extra use: 2.6 percent; advertising on ISP's website: 0.5 percent; fees for nonsubscribers: 0.1 percent; installation charges: 1.7 percent; unspecified sources: 1.1 percent).
As shown in exhibit 8.1, respondents in the high ISP revenue category derive a greater proportion of their revenue from business subscriptions than firms in the other categories do, probably because the business market is more lucrative than the residential market.
ISPs were also asked to identify the sources of their revenue according to the particular ISP service that their revenue was attributable to. The distribution of total revenues across the entire respondent base is shown in exhibit 8.2. As shown, 76 percent of ISP revenues are derived from the provision of access services; 14 percent from backbone services; 2 percent from electronic commerce (e-commerce) services, 6 percent from value-added services other than e-commerce; and 2 percent from "other" sources (e.g., installation charges). Although the large majority of ISP revenue attributable to access services was expected, the small proportion of revenues attributable to e-commerce services was somewhat of a surprise. It is likely, however, that despite the fact that ISPs have significant e-commerce capacity in place, the relatively slow take-up of e-commerce by consumers and businesses has restricted ISPs revenues in e-commerce supply. As discussed in further detail below, the use of e-commerce is expected to accelerate as issues such as online security are resolved.
The provision of ISP services, focusing especially on e-commerce, is considered in the following section.
Across Canada, 97 percent of respondents reported that they derive revenue from the provision of access services. Regardless of whether the respondent base is segmented by region, by operating revenue, by ISP revenue, or by size of customer base, we find the same result: 95 to 100 percent of respondents offer access services. This is not surprising given that the survey's definition of an ISP is a firm that, at minimum, offers Internet access to the public (the firm may also, of course, offer other Internet services). This high level of supply explains the high proportion of total ISP revenue that is attributable to access services.
A large proportion of ISPs also offer backbone access services. Across Canada, 30 percent of respondents reported that they derive revenue from the provision of these services. The results also show that large ISPs are more likely to offer backbone access services: 42 percent of large firms; 39 percent of medium-sized firms; 24 percent of small firms; and 10 percent of micro-sized firms offer backbone access services. A similar picture is obtained when respondents are grouped by ISP revenue: 39 percent of high ISP revenue firms; 35 percent of medium ISP revenue firms; and 20 percent of low ISP revenue firms provide backbone access services. These findings are not a surprise given the fact that large ISPs (both facilities-based and "Tier 1" ISPs) tend to supply small ISPs with their backbone access.
Although ISPs in Canada rely on the provision of access services for the bulk of their revenue, and a relatively small proportion of their revenue is currently attributable to the provision of e-commerce services, they are increasingly moving into the provision of these and other value-added services.
E-commerce services provided by ISPs tend to be complementary to access services, and are often offered in a bundle, or package, of services. Across Canada, 6 percent of respondents indicated that they include e-commerce services in their monthly subscription packages.
In order to follow up on the industry data collected by Industry Canada in 1995, the 1997 survey increased its focus on the provision of particular e-commerce and other value-added services. As indicated in exhibit 9.1, 44 percent of respondents indicated that they provide credit card services in their e-commerce service offerings. Given that credit cards are one of the most common instruments for online payment, this relatively high level of supply is to be expected. As credit card suppliers develop mechanisms to increase the security of online credit card use, the provision of credit card services can be expected to grow.
| Website Hosting, Vendors | 64% |
|---|---|
| Credit Card Services | 44% |
| Security Services | 43% |
| Pay-per-Use Web Sites | 22% |
| EDI Services | 13% |
About two-thirds of ISPs provide website hosting for merchants, consistent with widespread media coverage of the growing popularity of online store fronts and "shopping malls." This finding is also consistent with the 1995 survey which found that 80 percent of ISPs offered corporate website hosting (noting that corporate website hosting is broader than merchant website hosting -- it is believed that website hosting for vendors has grown significantly over the past few years).
Pay-per-use websites, such as those offering online services or entertainment, are also popular with survey respondents. As indicated in table 7.1, almost one-quarter of respondents offered this service. As consumer confidence in online security and privacy is increased, this type of service will most likely increase in popularity. More merchants can also be expected to move to this type of distribution channel for multimedia and other Web-friendly services.
In their effort to enhance consumer confidence, many ISPs are providing security services, such as secure websites and secure messaging. Forty-three percent of respondents indicated that they provide such services to their business and residential customers. Section 10.0, below, discusses the efforts of ISPs to foster trust in the online environment in further detail.
Despite the fact that electronic data interchange (EDI) likely remains the most common type of e-commerce even today, relatively few ISPs (13 percent of respondents across Canada) provide this type of service. Two points, however, are important to note. First, ISPs handle an increasing volume of EDI traffic as more and more commerce is moving from closed networks to the Internet. This trend is partly a result of the increased use of security mechanisms, such as encryption, that allow for more secure Internet data flows and electronic transactions. Due to the fact that the transport and routing of this EDI traffic is not an EDI service per se, it would not show up in the 13 percent figure noted above. Second, it is likely that the relatively low level of EDI service provision is a result of more and more EDI services being transitioned to Web-based e-commerce services that are more efficient and more user-friendly.
The provision of e-commerce services and firm size (as measured by total firm revenue) are positively related: 22 percent of large firms; 16 percent of medium-sized firms; 11 percent of small firms; and 7 percent of micro-sized firms reported that they derive revenue from e-commerce services. This positive relationship may be due to the earlier take-up of e-commerce technologies by large firms, or to the fact that large ISPs tend to have a greater proportion of business customers (who are heavier users of e-commerce services).
A similar pattern emerges when firm size is measured by ISP revenues: 23 percent of high ISP revenue firms; 14 percent of medium ISP revenue firms; and 9 percent of low ISP revenue firms reported that they derive revenue from the provision of e-commerce services. Again, this trend likely results from the greater capacity and earlier take-up of e-commerce by larger firms.
Given that firm size and number of customers are positively related, it is not surprising that the provision of e-commerce services and the number of residential customers are also positively related: 21 percent of ISPs with 2,500 customers or more; 14 percent of ISPs with 500 to 2,499 customers; and 10 percent of ISPs with fewer than 500 residential customers reported that they derive revenue from e-commerce services.
Although the provision of e-commerce services and the number of business customers do not follow as strict a relationship, those ISPs supplying a larger business customer base are more likely to derive revenue from e-commerce services than those supplying smaller markets. Across all respondents, 24 percent of ISPs with 250 or more business customers; 7 percent of ISPs with between 75 and 249 business customers; and 8.5 percent of ISPs with fewer than 75 business customers reported that they derive revenue from e-commerce services. Again, business customers tend to be heavier users of ISP e-commerce services, partly explaining this trend.
As noted above, ISPs are offering a number of services to increase consumer confidence and foster trust in the online environment. More and more, ISPs are realizing the importance of addressing security- and privacy-related concerns of both online buyers and sellers, and of dealing with such challenges as offensive and illegal content. This section of the study considers a variety of these factors.
The Canadian legal study "Cyberspace is Not a No-Law Land" (see <http://strategis.ic.gc.ca/SSG/it03117e.html>) concluded that, among other things, liability for involvement with illegal or offensive content, such as hate literature, is greater for those providing the content than for those involved in the essentially unknowing transport of such content ("unknowing transport" results from the fact that ISPs normally route and transport a large volume of Internet traffic that they do not know the contents of). The legal study also found that ISPs and other intermediaries can reduce their potential liability by employing certain measures or carrying out particular activities. Some of these measures were considered in the 1997 survey, as reported below.
| Business Information | 73% |
|---|---|
| Entertainment or Leisure | 47% |
| Family-Oriented | 44% |
| Educational or Training | 43% |
| Adult-Oriented | 7% |
| Other | 22% |
Over one-third (36 percent) of respondents indicated that they provide Internet content. As indicated in exhibit 10.1, the most popular type of content provision -- by far -- is business information, at 73 percent of content providers. Also very common is the provision of entertainment or leisure content at 47 percent, family-oriented content at 44 percent, and educational/training-related content at 43 percent. Only 7 percent of content providers supply adult-oriented content.
As noted above, it is the content providers themselves who are most likely exposing themselves to a significant level of liability, especially if it is offensive or illegal content they provide. Both content providers and "unknowing transporters" of content are, however, receiving customer complaints about offensive or illegal content and conduct (conduct such as hacking). Interestingly, over twice as many ISPs have received at least one complaint for offensive conduct (65 percent) than those who have received at least one complaint for offensive content (30 percent). This finding suggests that hacking, harassment, and other types of offensive and illegal conduct may be a greater challenge to ISP industry growth and development than the much-publicized offensive and illegal content problem. However, there does appear to be significant concern over illegal and offensive content: Over two-thirds (69 percent) of respondents indicated that they are asked to provide products, services, or policies to restrict access to such material.
In response to these concerns, and to increasing pressure from other stakeholders, Canadian ISPs are taking significant steps to deal with offensive and illegal content and conduct. Over half (55.1 percent) of respondents indicated that they would discontinue the subscriptions of offenders; almost 50 percent adhere to a code of conduct (that normally specifies what action the ISP will take in the event of a customer complaint); and over 40 percent engage in customer education and awareness-building. If it is technically possible, almost 40 percent will remove the offending or illegal material from their servers; over one-third will utilize blocking or filtering software; and over one-third will consult with law enforcement officials.
The use of corporate policies to specify user and ISP responsibilities, and to specify action that will be taken in a given situation, is an increasingly popular mechanism used by ISPs. Almost three-quarters of respondents have corporate policies in place, with over one-third of respondents using them to deal with offensive and illegal content and conduct. As shown in exhibit 10.2, 84 percent of respondents that use corporate policies have acceptable use policies in place, 53 percent denial of service policies; 44 percent content policies; 43 percent complaint policies; 39 percent privacy policies; 38 percent security policies; and 38 percent service level policies.
| Acceptable Use | 84% |
|---|---|
| Denial of Service | 53% |
| Content | 44% |
| Complaint | 43% |
| Privacy | 39% |
| Security | 38% |
| Service Level | 38% |
| Other | 2% |
* This table represents 73 percent of respondents with corporate policies in place.
Though not as prevalent as corporate policies, customer and supplier contracts are used by about 28 percent of respondents to counter offensive and illegal content. Moreover, 27 percent block access to offending sites or newsgroups; 21 percent train their employees; and 10 percent provide "family friendly" viewing areas. Only 12 percent of respondents indicated that they have no practices in place to deal with offensive and illegal content.
Online security and privacy and the protection of personal information are important for fostering trust in the online environment and digital economy. Accordingly, ISPs were asked about their views on these issues. Somewhat surprisingly, only 4 percent of respondents considered the lack of online privacy and security to be a high barrier to growth, whereas 63 percent considered the lack of online privacy as a low barrier to growth and 62 percent the lack of online security as a low barrier to growth. ISPs responded in this way despite the fact that 40 percent of respondents reported that online privacy and security was of high importance to their customers. It appears, however, that this set of findings can be reconciled by the fact that ISPs are taking steps to address their customers' concerns: 48 percent of respondents indicated that they are placing a high degree of emphasis on online privacy and security in their business planning; 55 percent on the protection of personal information; and 60 percent on data and transaction security.
Over the coming months, Industry Canada will continue to conduct analysis on the rich variety of statistics that were collected under the survey. The ongoing analysis will not only consider e-commerce and other value-added ISP services in further detail, but will also focus upon employment, wages and other expenses, barriers to growth, profitability, research and development, and strategic alliances. Ongoing analysis will also consider the potential for future growth and development of the ISP industry in Canada.