The Main Trends in the ISP Market

Related Research Work on ISP Market

In Internet economics literature there is no comprehensive work on analyzing the ISP market. According to the existing research work, initially emphasis was given on the classification of ISPs. This classification was based on various criteria such as network coverage and geographic location; customer types, and core business activities like network service provision versus information service provision. Recent research work has focused on describing demand dynamics and their implications on the ISP business.

Greenstein (2000) stated that the Internet market value chain had not settled yet. ISPs use different strategies for exploiting existing business opportunities. He illustrated this situation by presenting the U.S. ISP market data on their network geographic coverage, and services provided. He classified ISPs into five groups according to the service they offered: access service, networking and maintenance, Web hosting, website development and servicing, and high-speed access services. He argued that ISP business in the U.S. expanded based on provision of services tailored to specific customer needs. Lakelin, Martin, and Sherwood (1999) introduced another classification to capture the development of European ISP business models. They classified ISPs according to their position on the Internet value chain (e.g., packaging, connectivity, infrastructure) and the origin of the company (e.g., IT sector, telecommunication sector, or new entrants). They stated that in the near future the dominant players would be the branded and content-based ISPs.

Huston (1999) provided a more comprehensive analysis of the ISP market, taking into account both technical (i.e., IP networks, network equipment) and business aspects. He argued that the ISP market was characterized by uncertainty with respect to demand patterns and positioning of key players. He emphasized cost factors and provided guidelines for a successful ISP business plan. He suggested that future research should focus on revenue sharing and cost-compensation mechanisms.

These approaches do not provide an integrated framework for analyzing the ISP business relationships in the Internet market. They are mainly limited to analyzing the relationship between the ISP and the end user, neglecting to discuss relationships between ISPs like revenue sharing and cost compensation.

We analyze the ISP business in a broader perspective in order to identify revenue-sharing mechanisms. To this end we introduce a reference model about ISP business activities, at a high level of abstraction, that provides an integrated framework. This reference model will facilitate the analysis of ISPs' business relationships.

ISP Market Trends and Pricing Strategies

Although the ISP market growth is slowing down, technology advances in IP networks (e.g., broadband services, Ipv6) and in mobile networks (e.g., UMTS, WLANs) create new business opportunities for ISPs (Constantiou, 2002). In addition to this, opportunities for service differentiation and revenue generation emerge from the increasing demand for network service quality needed for real-time applications. Along with this growth, industry consolidation has already taken place. Emerging telecom operators and incumbent telecommunication companies have acquired most of the pioneers ISPs on the Internet market (Lakelin, Martin, & Sherwood, 1999).

In the network services market segment, there is an open debate on two opposite trends: the price competition on the provision of network services, and the ISPs' strife for profitability. This debate comes from the inability of some ISPs to offset the cost of building and running their network, as well as the cost of attracting new customers. In addition, strong pressure is being exerted on ISPs to devise new service-price mixtures that differentiate them, avoiding direct competition by prices, in a commodity market. A wide variety of pricing schemes could be applied in the Internet market, ranging form hourly charges, fixed-rate charges without usage metering, to volume-based charges. However, a limited number of pricing schemes are offered.

To attract customers and generate high traffic, several ISPs try to offer free Internet access to their customers and generate revenue from advertisement. However, this is not a long-term viable model, as the advertising companies want to reach different target groups, namely customers that are willing to spend money (Constantiou, 2002). Currently, most of the ISPs that offered free access service have gone out of business. Other ISPs are migrating their customers to payment schemes for access services (e.g., United Online, NetZero). The free-to-payable conversion represents a basic strategy for many ISPs seeking to diversify their revenue streams and reduce their reliance on advertising revenues. Finally, a few ISPs (e.g., Winfire) use a free access model for a short time period, to introduce new access technologies like DSL and attract new customers, that is then upgraded to payable premium service (TheList, 2002).

Many ISPs use flat-rate access pricing schemes that involve a low fee for a fixed time period and allow end users to have unlimited access to the Internet. Flat-rate pricing is the most prevalent form of pricing in the U.S. (ACCC, 2000). It is used for introductory offers to new access technologies or as an attempt by start-up ISPs to rapidly build a customer base. However, this pricing scheme attracts a disproportionately large number of high bandwidth consuming users and rapidly erodes any economy of scale that new customers may bring by deteriorating the quality of service of all users.

The flat-rate price is based either on the peak data flow rates or on a ceiling on the total data volumes passed to and from the ISPs' network. This allows service differentiation. Its main advantages for the ISP are the low administrative and billing costs. Invoicing and payment may also take place in advance of service delivery. From the end user perspective, the main advantage is that incremental costs of service provision are not charged. However, with flat-rate access pricing, lower volume customers may increase their traffic without generating corresponding increase in the ISPs' revenues. Some possible reactions from the ISP include the increase of the flat fee to recover higher costs, the adaptation of marketing strategies to promote premium services, or the attempt to access larger capacity transit networks that offer the ISP lower unit cost of transport service (Huston, 1999).

A variation of flat-rate pricing is tiered usage-based pricing (Altmann & Rhodes, 2002), a combination of access for a number of hours per month and a number of megabytes downloaded per month up to a certain threshold, which is offered by many ISPs (e.g., AOL, AT&T WorldNet, EarthLink, FreeServe). For example, a typical offer includes thresholds of 50 hours access and 250 MB data volume or limited services session time (TheList, 2002). The threshold placement enables ISPs to differentiate themselves from their competitors. ISPs will benefit from end users, who spend more time online and download greater data volumes, paying a proportionately greater amount. Additionally, better resource allocation may be achieved as tiered usage-based pricing allows segmentation and profiling of customers (Altmann & Rhodes, 2002). Moreover, it prevents customers from monopolizing ISP modems or inbound lines, and gives all the customers a "more reasonable" chance of logging on.

Some ISPs, in particular backbone service providers, have recently introduced usage-based pricing, mainly volume or time based, in their transit agreements. With this pricing scheme every MB delivered incurs a charge to the customer. Rates may vary depending on time of day to encourage smooth utilization of the available bandwidth resources. There are also some more sophisticated usage-based pricing schemes like those of UUNET, which offers dedicated lines to business customers, who are charged according to their "mean" transmission rates (TheList, 2002). Such pricing schemes expose customers to the incremental costs of service provision and give them an incentive to monitor their usage of network resources. It is worth mentioning that similar pricing policies are indirectly offered to customers in many European countries, where local telephone calls are still charged based on the duration.

Finally, ISPs (e.g., SPRINT, Covad, UUNET) also offer Service Level Agreements (SLAs) to their customers. The SLA is an agreement that creates a common understanding of the service and the performance level required from the ISP by a customer. An SLA typically includes the service types, service quality parameters to be provided, the reporting and troubleshooting processes, the time period for resolving the problems, and the process for monitoring and reporting the performance and quality delivered. SLAs mainly set usage-based pricing schemes (Jensen, Grgic, Espvik, & Rohne, 2000).