Overview
This paper is a follow-up to an article I wrote in 1998 entitled, “What a tangled web we weave…A review of pricing models and the forces that drive them.” I’m going to begin by reviewing the pricing models that are currently in use, go on to discuss why pricing information is such a challenge, and end up by suggesting that there can be no universal pricing model. Instead we need numerous models, each suited to particular kinds of information and particular groups of customers.

Pricing models
Today there are some sixty different models of pricing in active use:

Many of these models can be used in combination with others. The combinations allow for some 20,000 different kinds of models. For certain vendors – for example, Dialog – the models themselves have become so complex that the vendor needs to use databases to calculate prices.

The models fall into some general categories:

Usage based pricing sets up a measure that assumes a certain value for each action and then charges accordingly. Even site licensing is a variant of usage based pricing, where the limit is the description of the site and there is still an assumption of the amount of use that will go on within the site.

Pricing discounts are usually tactical. Their purpose is to reward customers who help the company by buying early or to pass on cost savings for ordering more volume. Notice that often a discount is actually a price. For example, if I buy electronic and paper together, I get a lower price.

Value added forms of pricing differentiate customers according to particular features or portions of the service that a customer might value as desirable. It’s similar to value removed forms of pricing that disable functionality or reduce value so that the product can be priced less expensively.

Sponsored pricing applies when advertising, sponsorship, or other goodwill events dispense with fees completely. In some cases, the customer must undertake an action to get the discount, such as providing the sponsor with a number of click-throughs on banner advertisements.

Developments since 1998

Since 1998, I found only three pricing models that have essentially been discontinued. In all cases, the models had been rendered obsolete by technological advances. Today, almost all modems are now 56K or faster, so pricing by BAUD rates replaced by pricing by bandwidth. Wherever a pricing model has disappeared, a new version of the model appeared to replace the old. New models based on numbers of workstations or geographical definitions, for example, exist while old models based on definitions of LANs and WANs have gone away.

A few completely new models have appeared:

The DialUnit is a particular model from Dialog. To quote the Dialog Website: “A DialUnit is a measure of system resources used to execute search and display commands. Each command - except administrative commands, such as the help commands - generates some portion of a DialUnit. DialUnits are calculated based on the system resources (commands) that are used from LOGON or a BEGIN command to the next BEGIN command or LOGOFF. DialUnits are tracked as partial units as small as 0.001. DialUnits do not accumulate during ‘think time’ or browsing.

netLibrary introduced a model that mimics the physical world. Each virtual book can be checked out by one user, in which case no other user can access the item. A library can buy multiple “copies” of each book to serve the number of potential users.

Other new models simply reflect technological advances. Fees for downloading to particular PDA platforms, fees based on IP addresses, and other new models have emerged.

In summary, the landscape has changed very little. We continue to see ever more complicated models. Why? In 1998, I identified five forces that were causing the complexity. They were:

Customers press for lower prices
Publishers create additional models to maximize revenues
Different products need different models
Technology enables pricing models that reflect value more accurately
Sublicensing makes models more complex still

Four years later, these forces remain, along with a strong new force that has emerged:

Publishers need to maintain revenues

The challenge for many publishers has been to shepherd their businesses through each technological change without seeing revenues decline. In the past few years, a particular publisher may have needed a new model for online delivery through an online aggregator, followed by another new model for CD-ROM, and yet another for networked CD-ROM. Now the same publisher, fresh from creating new pricing models for the Web, is now struggling with models for PDA delivery.

The best way to lessen the effect of such migrations on revenue is to create new pricing models that make such migrations revenue-neutral. Giving customers both print and electronic versions of a journal does just this. The publisher makes the decision to keep things simple and load the cost of the migration across all customers. Instead of rewarding one customer for staying with the old technology, all customers are encouraged to move to the new technology.

This is why some publishers are reticent about giving discounts to customers who drop the print or who don’t use the electronic versions. The cost of supplying both versions to the customer can be cheaper than trying to sell each independently.

A similar need to preserve revenues can be seen operating in the recent use of embargos with respect to aggregated journal collections.

In the past few years, large aggregators have played a major role in pricing models. To the customer’s benefit, the aggregators have pushed content owners towards standard models. They’ve also created very significant price savings. In 2000, many colleges paid approximately $1.70 per FTE to provide access to some 4,000 magazines and journals. This works out to less than $0.0001 per article, or several thousand times less than the cover price of the printed versions.

While librarians were intent on keeping both paper and electronic subscriptions, many publishers felt that aggregators brought them additional revenues that they could not otherwise secure. It meant that existing revenues were protected by the customer’s unwillingness to discontinue the paper.

Now librarians are canceling paper in favor of the electronic. This directly threatens the original publishers who rely only on aggregators. Not only are the electronic prices many orders of magnitude smaller than their print counterparts, the original publisher also has to pay the aggregator. As a result, original publishers sought to remove their titles – until the danger of cannibalization of their existing revenues was averted through journal embargos. Yet again, a new model was created in response to publisher and customer pressure.

Products are growing in value and becoming still more complex

Much of the complexity in pricing is caused by the complexity in valuing information. In a perfect world, the price of information would reflect perfectly the value of the information. We would reward people who create information of great value and provide incentive for them to create more of the same.

The problem is that there is no effective way to quantify that value. Here’s why:

The quality of content varies. In a database today, where you may have peer-reviewed and non-peer-reviewed content, value changes by how widely something is known – what is the value of an unpublished manuscript?
The form of the content varies. The same article can be displayed in PDF, with images or without images, in ASCII or facsimile, as part of a portal or not, with detailed indexing or not.
The value changes over time. The value of a stock quote delivered to you a second before everyone else has it is enormous. But the same stock quote fifteen minutes later can be had for free on the Internet. How do you balance the value of a preprint that gives you a lead in your research, versus the same article after it’s been reviewed but when it’s old news?
The value changes by reader. What is the value of a scholarly journal article for a Ph.D. student, in comparison to the value for a high school student?
The value changes by the performance of the retrieval engine and software. If you cannot find a key article in a database, does it have any value at all?
The value changes as the technology enables it. A web subscription that can be used by the entire campus has more value than a networked CD-ROM that has to be maintained by the institution.

In the print world, this was much less of a problem. Information stayed neatly within its covers, in conveniently sized units, intended for a particular group of people. There were no software features, multiple formats, interlinking, or search engines to worry about. The pricing of each unit was made easier because the target customer was known, the content could not easily morph into another form, and only one user at a time could read it.

In the list above, most of the pricing models attempt to replicate the comforting world of paper. They deliberately ignore essential elements of electronic value, such as linking, search power, or response time, and focus instead on trying to lock in the value– just as we did with books.

Perhaps because of the size of electronic products, it has become standard to consider information as a commodity. We measure the value of a product by the number of articles it contains, the number of journals it has, or the number of times it is used. This is a natural tendency – an effort to quantify what cannot easily be dealt with qualitatively.

The truth is that information in electronic form is far from a commodity. Not only does the underlying content have myriad kinds of value, the way in which it is delivered can enhance or detract from the value as well. Technologists measure the size of the Internet in terms of bytes and bits, but librarians and publishers measure it in a variety of ways. Each journal or book is special, each electronic implementation different, and each product has unique value.

With so many different kinds of value possible in the electronic world, it is to be expected that the pricing models should also be different from product to product. We cannot expect a universal metric.

Can the models be generalized?

If we accept that the performance of electronic products has countless different kinds of value, how should publishers price and what should customers be prepared to pay?

At Alexander Street, we begin with some general principles: We believe that similar customers and consortia should pay the same amount. We encourage as much use of our resources as possible. We make sure that our licensing partners receive a reward for the inclusion of their materials, and we ensure that our pricing does not result in a loss to the rights owners. Above all, we make sure that our model results in prices that the customer finds reasonable and that will cover our costs. If we don’t satisfy these last conditions, we will not have a product.

The two models that we use the most frequently are the annual subscription site license and the one-time permanent license. Both models allow for unlimited access (unlimited simultaneous users):

The annual subscription site license is tiered by materials budget, to allow smaller schools with lower funds to participate.
The one-time permanent license allows a library to deliver the resource to their patrons for in perpetuity. In the short term, this option is more expensive for an institution, but in the long term it costs much less and ensures permanence.

These models are simple, and they encourage users to use the products to the maximum. The models make no attempt to quantify the value of the various attributes of each product. We do not attempt to work out the value of a single download, a single simultaneous user, or the cost of having large numbers of simultaneous users on the system. Finally – and most importantly – the models are ones that our customers like and that allow us to recover our costs. In short, they result in a fair price to the customer and a fair revenue stream to us.

Can these principles be applied elsewhere? Obviously they can. But they’re not for everyone. A document delivery company would be prevented from using such models by their license agreements, for example.

Summary

The proliferation of pricing models is a natural consequence when publishers and customers seek the best deal for their respective constituencies. The complex nature of electronic information and the richness of value result in still more models. Just as there’s no universal measure of the value of information, so there can be no universal pricing model that fits all kinds of information.

Pricing models have one purpose – to generate prices that are acceptable to both publisher and customer. It is essential to focus not on searching for the perfect model, but to come up with flexible models that result in the best prices. After all – as I said in my last article – it’s about prices, not models.

© Copyright 2008 Alexander Street Press. All rights reserved.                 Last Updated: 05-Mar-2008