In the technological olden days – before 2000 – when everyone used 56k modem dialups or the equivalent and their main activities were e-mailing and web browsing, the pricing options for internet service providers (ISPs) were pretty straightforward: they charged by time usage. But all that has changed with the introduction of more applications that require greater amounts of data and speed, such as video streaming. So what is the best pricing option now?

Telecommunications companies have struggled with this question, with even major firms like Verizon at one point releasing conflicting statements about its planned pricing structures. That situation has led Ying-Ju Chen and Ke-Wei Huang to explore the ideal pricing options from a theoretical perspective.

“Practitioners do not have scientific methods for choose pricing metrics for data services, even though this seems to be key to ISPs and cell phone companies being able to capitalise on the explosion of data travelling across the web,” they said.

The authors address that gap by constructing a model involving a monopoly data service provider (the seller) and consumers who have differing needs, and testing which metrics would give ISPs the greater profits.

Given that consumer utility from the service consists of time usage, connection speed (megabytes per second, or Mbps) and data usage (gigabytes, or Gigs), the seller needs to choose which one to base the pricing plan on – a choice that is not a simple matter. The obvious factors of consumer demand and characteristics are not the surest guide, they said. “What could matter more is the flexibility left to consumers to adjust the remaining pricing metrics that are not specified in the pricing plan.”

For example, if pricing is based on time usage, customers of varying characteristics and needs will all download as much data as possible in the shortest period possible, leaving the seller with little option but to charge a uniform price.

“The flexibility left to the consumers is so strong that the monopoly price discrimination power is completely deprived. An implication is that a poorly-chosen metric for price discrimination may not generate more profit than a uniform pricing strategy,” they said.

But if the seller prices by Mbps or Gigs, they are able to suppress the after-sales flexibility of the consumer. This points to optimal pricing strategies being multi-tiered plans that can be tailored for different consumer needs.

Under charging by Mbps, since the consumer has the discretion to decide how long they spend using the service, they will care about connection speed as well as data usage. Homes or small businesses with multiple users may need a faster speed and therefore may be willing to pay more for it.

Pricing by Gigs is more common for cell-phone services, as users in the period studied mainly use 3G services for e-mail or browsing and rarely share bandwidth. This helped to explain why cell-phone services are charged differently from residential services (typically ADSL – asymmetric digital subscriber line), which is usually by Mbps, the authors said.

They also show that price discrimination is beneficial to society because it means consumers with low valuations can be served with low-usage plans.

Overall, “either pricing by Gigs or Mbps can be optimal. Although pricing by Gigs can dominate pricing by Mbps even if the consumer’s utility is more sensitive to changes in connection speed, when incorporating bandwidth costs or congestion costs, pricing by Mbps becomes more attractive as it allows the seller to directly control the congestion effect,” the authors said. They suggested that their findings could help practitioners to develop their own pricing plans and pricing metrics selection.