This paper, published by Public Knowledge on April 23, 2012, is a deep dive into questions about data caps and so-called "usage-based billing." It considers economic and historical perspectives and raises some concerns, both in regards to competition and to national priorities such as broadband adoption, education, employment, and innovation more generally.

The full paper is available here as a PDF. Below is the executive summary and final recommendations.

The paper was written by Andrew Odlyzko, University of Minnesota; Bill St. Arnaud; Erik Stallman, Holch & Erickson, LLP; and Michael Weinberg, Public Knowledge.

Executive Summary

Usage-based pricing, today most commonly encountered in the form of data caps, is rapidly becoming part of the Internet access landscape.  Wired and wireless Internet service providers – most of whom had traditionally operated on an unlimited basis – are evaluating or implementing pricing strategies that limit the amount of data a customer can use, charge customers for using data beyond a predetermined amount, or combine the two.

Although some providers have been quick to embrace these pricing structures, consumers have generally not been enthusiastic, and have often expressed strong protests.  For their part, regulators have largely avoided asking even basic questions about this trend.

This whitepaper is an attempt to begin a serious consideration of usage-based pricing.  It attempts to move beyond rhetoric and recognizes that usage-based pricing is a tool.  As with any tool, usage-based pricing can be used for both productive and destructive ends.  Sometimes these ends are intentional.   Other times, they are a byproduct of other goals or even a lack of careful consideration. 

Regardless of the motivation driving its implementation, usage-based pricing has the potential to significantly impact how networks are designed and used.  This, in turn, impacts the innovation that relies on those networks.  Before deciding if and when usage-based pricing is desirable, it is critical to fully understand the history of usage-based pricing, how it impacts markets, and both the benefits and harms that such a model can bring.

This paper aims to explain the basic issues surrounding usage-based versus flat-rate pricing.  Section I examines the trend towards usage-based pricing in both the wired and wireless markets.  Section II then considers the benefits and justifications for using usage-based pricing.  This is followed in Section III by a review of the history and economics of flat rate pricing.  Since broadband access is central to so many national and societal goals, the penultimate section – Section IV – discusses the problems that might be caused by usage-based pricing.  Finally, we end with a series of conclusions and recommendations for responsible implementation of usage-based pricing.

Conclusions and Recommendations

Flat pricing schemes should be regarded as an ideal goal, supporting innovation and social and economic welfare, and not as irrational aberrations that promote inefficiency and waste.  However, sometimes temporary resource constraints may make flat rates infeasible.  Currently, on the wired Internet, that does not appear to be a real concern, as the rate of progress in technology appears to be comparable to the rate at which traffic demand is rising, so that should be possible to support the growth in traffic without increases in the level of investment.  On the wireless side, traffic is growing faster than carriers are investing in capacity improvements, so the case for UBP appears far stronger.

In any event, it appears that many service providers have already or are moving towards implementing some form of UBP.  In light of this, the following practices will help to minimize (but not eliminate) the negative effects of UBP.

1. Transparency

If users are going to be charged on a per-bit basis, or limited to a certain number of bits, there must be straightforward, up-to-date ways for users to measure their network activity.  A user must be able to check how much data she has used at any given moment in order to make informed decisions about downloading a new app or watching a video.  If these meters are not accurate, users cannot reasonably be held accountable for overage fees.

Transparency should also extend to the underlying justifications for the pricing structures.  Service providers must be forthcoming about how caps are set, tiers are created, and pricing is determined.  Providers must explain what goals UBP is designed to achieve.  AT&T’s recent across-the-board $5 increase for its data plans came with little warning or explanation.  It is easy for service providers to manipulate UBP in anticompetitive or counterproductive ways, but transparency and accountability will help to mitigate that threat.

2. Implement UBP in a Granular Way

Data limits and UBP that does not take time of day into account are ill suited to address service provider concerns about network congestion.  As described above, network congestion is not a cumulative phenomenon.  Instead, it occurs at specific times of day on specific parts of the network.  A UBP scheme that does not recognize that cannot credibly claim to be designed to reduce network congestion.

At a minimum, carriers should restrict UBP to specific times of day that are most likely to be congested.  Data sent or received during off-peak hours have no meaningful impact on network congestion and should be excluded from the scheme.  Data sent or received during peak hours could be charged at rates that reasonably reflected their impact on network congestion.  This practice could encourage users to manage their network usage more efficiently and reduce congestion generally.

3. Government Oversight to Protect Competition

In today’s concentrated landscape, regulators must vigilantly monitor UBP schemes to ensure that service providers do not leverage market power to increase costs and suppress demand for competing services delivered over IP.  There are many market forces pushing service providers in this anticompetitive direction, and precious few guarding against it.  Until there is robust competition among service providers to discourage network pricing manipulation, regulatory oversight is critical to maintaining a competitive landscape for services delivered over IP.

4. Prevent Artificial Scarcity

UBP should not become a substitute for investments and technical solutions to address congestion and increase broadband deployment and capacity.  Network operators have experienced congestion in the past, and always invested and innovated their way out of the problem.  UBP can create incentives to monetize network congestion and scarcity, and therefore to create artificial scarcity.  If such incentives were to become standard market features, it could further depress the United States’ already low broadband deployment and adoption numbers.

5. Documentation

Due to the complex nature of the communications market and the dearth of available information, service providers should collect and report detailed information about their offerings and how they affect consumers.  This should include, but not necessarily be limited to, the caps and pricing plans in place, how many users go over any caps and how much those users pay (in total and on average) as a penalty for exceeding the cap, procedures for notifying consumers once they are at or near their broadband caps, and the rate of continued investment in network infrastructure.  This information will help regulators and the public understand the challenges faced by service providers, as well as the impact that UBP is having on use, adoption, and deployment.

This paper, published by Public Knowledge on April 23, 2012, is a deep dive into questions about data caps and so-called "usage-based billing." It considers economic and historical perspectives and raises some concerns, both in regards to competition and to national priorities such as broadband adoption, education, employment, and innovation more generally.

The full paper is available here as a PDF. Below is the executive summary and final recommendations.

The paper was written by Andrew Odlyzko, University of Minnesota; Bill St. Arnaud; Erik Stallman, Holch & Erickson, LLP; and Michael Weinberg, Public Knowledge.

Executive Summary

Usage-based pricing, today most commonly encountered in the form of data caps, is rapidly becoming part of the Internet access landscape.  Wired and wireless Internet service providers – most of whom had traditionally operated on an unlimited basis – are evaluating or implementing pricing strategies that limit the amount of data a customer can use, charge customers for using data beyond a predetermined amount, or combine the two.

Although some providers have been quick to embrace these pricing structures, consumers have generally not been enthusiastic, and have often expressed strong protests.  For their part, regulators have largely avoided asking even basic questions about this trend.

This whitepaper is an attempt to begin a serious consideration of usage-based pricing.  It attempts to move beyond rhetoric and recognizes that usage-based pricing is a tool.  As with any tool, usage-based pricing can be used for both productive and destructive ends.  Sometimes these ends are intentional.   Other times, they are a byproduct of other goals or even a lack of careful consideration. 

Regardless of the motivation driving its implementation, usage-based pricing has the potential to significantly impact how networks are designed and used.  This, in turn, impacts the innovation that relies on those networks.  Before deciding if and when usage-based pricing is desirable, it is critical to fully understand the history of usage-based pricing, how it impacts markets, and both the benefits and harms that such a model can bring.

This paper aims to explain the basic issues surrounding usage-based versus flat-rate pricing.  Section I examines the trend towards usage-based pricing in both the wired and wireless markets.  Section II then considers the benefits and justifications for using usage-based pricing.  This is followed in Section III by a review of the history and economics of flat rate pricing.  Since broadband access is central to so many national and societal goals, the penultimate section – Section IV – discusses the problems that might be caused by usage-based pricing.  Finally, we end with a series of conclusions and recommendations for responsible implementation of usage-based pricing.

Conclusions and Recommendations

Flat pricing schemes should be regarded as an ideal goal, supporting innovation and social and economic welfare, and not as irrational aberrations that promote inefficiency and waste.  However, sometimes temporary resource constraints may make flat rates infeasible.  Currently, on the wired Internet, that does not appear to be a real concern, as the rate of progress in technology appears to be comparable to the rate at which traffic demand is rising, so that should be possible to support the growth in traffic without increases in the level of investment.  On the wireless side, traffic is growing faster than carriers are investing in capacity improvements, so the case for UBP appears far stronger.

In any event, it appears that many service providers have already or are moving towards implementing some form of UBP.  In light of this, the following practices will help to minimize (but not eliminate) the negative effects of UBP.

1. Transparency

If users are going to be charged on a per-bit basis, or limited to a certain number of bits, there must be straightforward, up-to-date ways for users to measure their network activity.  A user must be able to check how much data she has used at any given moment in order to make informed decisions about downloading a new app or watching a video.  If these meters are not accurate, users cannot reasonably be held accountable for overage fees.

Transparency should also extend to the underlying justifications for the pricing structures.  Service providers must be forthcoming about how caps are set, tiers are created, and pricing is determined.  Providers must explain what goals UBP is designed to achieve.  AT&T’s recent across-the-board $5 increase for its data plans came with little warning or explanation.  It is easy for service providers to manipulate UBP in anticompetitive or counterproductive ways, but transparency and accountability will help to mitigate that threat.

2. Implement UBP in a Granular Way

Data limits and UBP that does not take time of day into account are ill suited to address service provider concerns about network congestion.  As described above, network congestion is not a cumulative phenomenon.  Instead, it occurs at specific times of day on specific parts of the network.  A UBP scheme that does not recognize that cannot credibly claim to be designed to reduce network congestion.

At a minimum, carriers should restrict UBP to specific times of day that are most likely to be congested.  Data sent or received during off-peak hours have no meaningful impact on network congestion and should be excluded from the scheme.  Data sent or received during peak hours could be charged at rates that reasonably reflected their impact on network congestion.  This practice could encourage users to manage their network usage more efficiently and reduce congestion generally.

3. Government Oversight to Protect Competition

In today’s concentrated landscape, regulators must vigilantly monitor UBP schemes to ensure that service providers do not leverage market power to increase costs and suppress demand for competing services delivered over IP.  There are many market forces pushing service providers in this anticompetitive direction, and precious few guarding against it.  Until there is robust competition among service providers to discourage network pricing manipulation, regulatory oversight is critical to maintaining a competitive landscape for services delivered over IP.

4. Prevent Artificial Scarcity

UBP should not become a substitute for investments and technical solutions to address congestion and increase broadband deployment and capacity.  Network operators have experienced congestion in the past, and always invested and innovated their way out of the problem.  UBP can create incentives to monetize network congestion and scarcity, and therefore to create artificial scarcity.  If such incentives were to become standard market features, it could further depress the United States’ already low broadband deployment and adoption numbers.

5. Documentation

Due to the complex nature of the communications market and the dearth of available information, service providers should collect and report detailed information about their offerings and how they affect consumers.  This should include, but not necessarily be limited to, the caps and pricing plans in place, how many users go over any caps and how much those users pay (in total and on average) as a penalty for exceeding the cap, procedures for notifying consumers once they are at or near their broadband caps, and the rate of continued investment in network infrastructure.  This information will help regulators and the public understand the challenges faced by service providers, as well as the impact that UBP is having on use, adoption, and deployment.