The service industry is rife with decision dilemmas involving choosing between usage-based pricing and access-based pricing, or how best to combine the two. Consumer-oriented services such as Netflix and Office 365 and enterprise variations of “software as a service” offer access- only pricing, i.e., unlimited use for a flat fee, with a subscription model tied to a particular time duration such as a month. For example, ShareLaTeX Collaborator offers cloud-based shareable document typesetting for an annual flat fee of X dollars, with unlimited projects.
The other extreme of simple pricing is the per-unit plan, also often employed by firms that sell digital goods to consumers as well as businesses such as YouTube movies and Amazon IoT. For example, Amazon Simple Email service is priced on usage, with Y dollars for every 1,000 emails. Firms may also combine the two into a two-part tariff model, where consumers are charged a fixed membership or access fee and then a per-unit price based on actual consumption. This strategy had been traditionally employed for club memberships and was adopted by telecom firms that combine a monthly rental fee with a per-minute call rate for post-paid plans. More recently, this model has been adopted by OpenTable, which charges restaurants a monthly subscription fee and then a fee per diner, as well as by Google’s Project Fi — a service with monthly access fees and per-Gigabyte data fees. Alternatively, some firms allow consumers to select from a menu; they can buy unlimited use for a flat fee or pay for usage under a per-unit pricing plan. This model is employed by LexisNexis, a vendor of business information solutions, and AirSupport, a flight planning solutions provider. Despite these use cases, there is no clear and rigorous guidance in the academic literature about when each type of pricing scheme will be more profitable to providers of information services.
Simple pricing plans show glaring weaknesses in the face of heterogeneity or differences among consumers. Consider a flat-fee plan, which charges the same rate to all users regardless of consumption. What happens when users have very different satiation levels? In some industries, consumption levels between light vs. heavy users vary in a 1:1,000 or even 1:1,000,000 ratio. Here, heavy users enjoy the benefits of the flat rate and the firm loses the opportunity to monetise their higher consumption, whereas light users experience a high per-unit cost and may even be priced out of the market.
Now, consider per-unit pricing. Here, light users have easy entry, but heavy users might find the plan unattractive and may even turn to a “do it yourself” (or make vs. buy) approach, and the firm faces a sharp tradeoff between margin and market coverage. Conversely, what happens when users are primarily heterogeneous on their value from consumption? The implications are quite different in this case. It becomes difficult to find a sweet spot under per-unit pricing, whereas the flat-fee pricing problem embeds a margin-volume tradeoff similar to that for durable single-unit goods. Overall, flat-fee pricing becomes more untenable as the level of consumption heterogeneity increases, while per-unit pricing suffers more under high heterogeneity in per-unit valuations, and there can be substantial untapped profit under both simple pricing schemes.
While deciding the extent to which usage can be monetised, firms must consider factors on the supply side as well as the demand side. On the supply side, it is imperative to consider the influence of marginal costs as well as the costs the firm would incur if it was to monitor usage across consumers. Earlier theoretical work has suggested that higher marginal cost tilts the decision towards usage-based pricing and higher monitoring cost does the opposite. On the demand side, the firm should identify consumer heterogeneity on usage because consumption levels may vary across customers and the nature of heterogeneity varies across business applications. Unlike single-unit durables, there are few theoretical guidelines for multi-unit goods due to an absence of tractable models for examining the effect of consumption heterogeneity along with valuation heterogeneity. In this paper, the authors develop a new way to jointly capture and vary the two forms of heterogeneity (valuation and consumption), obtaining richness of representation and maintaining computational tractability. They define a new metric, Relative Variation in Satiation Levels (RVSL), which measures variation in satiation (maximum consumption at zero marginal cost) levels in comparison to variation in first-unit valuations.
The authors find that between the two simple pricing plans, per-unit and flat-fee, per-unit pricing is more profitable when RVSL is high; alternatively, flat-fee pricing is a better choice when variation in satiation levels is low compared to the variation in first-unit valuations. The careful analysis reveals that RVSL is far more consequential in selecting a pricing scheme than traditional valuation heterogeneity. The authors find that if a firm is looking to adopt a simple pricing mechanism (i.e., a per-unit or a flat-fee) under unknown consumption heterogeneity, the per-unit pricing is a safer option, owing to its robust profit performance and higher market coverage across various market scenarios. The paper extends its comparative evaluation to various performance metrics beyond profits and also examines market coverage and consumer surplus under various scenarios. Per-unit pricing yields higher market coverage and flat-fee pricing yields more consumer surplus. Therefore, when market coverage is important to a company’s goals such as with network goods, per-unit pricing is the sensible choice. If engagement is more crucial than market coverage, flat-fee pricing is the better option.
The authors also examine how a company can make the most of the polarised advantages of flat-fee pricing and per-unit pricing, whether by combining the two as a two-part tariff (2PT) charging every buyer for both access and usage, or simply giving the customer a menu of options from which they can self-select between per-unit and flat-fee pricing. The menu allows customers to choose a pricing plan in line with their preference, so intuition might suggest that the menu would hurt the firm’s profitability because consumers would pick the plan that gives them better value. This intuition does hold under most conditions, but the authors make the notable discovery that the profit shortfall of the menu (relative to 2PT profit) is very small, and, in fact, the menu performs better when RVSL is very high and first-unit valuations are also quite diverse, whereas the menu dominates the 2PT substantially on measures such as market growth and consumer surplus. The results suggest that it would be wiser for firms to offer a menu of options and let the buyer choose between per-unit or flat-fee pricing, rather than imposing both usage and access fees under a two-part tariff.
However, authors caution against an unlimited plan when some consumers have very large satiation levels because higher consumption could lead to congestion effects or impose strong negative externalities. In such cases, firms should either limit extreme consumption or use a three-part tariff, which imposes zero marginal price up to some allowance level in return for an access fee, but also imposes a per-unit fee for higher consumption beyond the allowance. This three-part tariff may be better for profitability and consumer satisfaction than other “hidden” restrictions (e.g. quality degradation after a certain level of usage) that are often imposed to curb extreme consumption under unlimited plans.
The pricing challenges discussed in this paper have been present and relevant for a long time, but they have become more pertinent in recent years. Services are a bigger part of the economy. Moreover, the Internet has allowed firms to monitor usage, converting even products previously sold as “durables” into “usables” or services whose usage could be monetised. For example, a company that sells design software doesn’t care about the extent of usage across clients and charges every client the same fee, but if it offers a cloud-based platform to run the software, then this is an opportunity to monetise usage. Along the same lines, GE sold its wind turbines to E.ON as a service instead of as a durable, while charging based on how much electricity was produced. However, in order to be able to profitably monetise “usables”, firms must develop a solid understanding of demand heterogeneity and its impact on choice of pricing scheme. Failure to do so in an effective manner can be an expensive mistake or a lost opportunity. For example, upon on introducing unlimited service plans, telecom service providers ultimately had to set a cap on usage after learning that a small number of heavy users impacted the system in a disproportionate manner. The insights presented in this paper can help a variety of firms grappling with such pricing challenges.
In a nutshell, the authors recommend that although many consumers might find unlimited consumption attractive, firms need to be cautious about this strategy when marginal costs are high or when satiation levels are very diverse among consumers. Alternatively, usage-based per-unit pricing is more robust under various market conditions including high marginal cost, except when monitoring usage is difficult. Firms willing to try more sophisticated pricing strategies may be better off offering both (and letting customers choose between per-unit and flat-fee plans) with the caveat that excessive usage by a few consumers would make the flat-fee option undesirable and steer the ideal plan towards a two-part tariff.
Read more: Forbes