The continuing travel warning notices about the SARS virus notwithstanding, librarians and vendors still turned out in droves to see scheduled exhibits and enjoy the remarkable summer weather in Toronto. Only a few ILS vendors cancelled: noted was the absence of Sagebrush, TLC, Follett, for example. It’s a challenge to stay on top of developments in the changing world of library automation.

Comprehensive reporting is impossible, so again this year I chose to focus on two narrow topics: the overall state of vendors in a flat economic market and their official positions on sharing API software with their library customers. There were no dramatic new announcements of corporate mergers.

Two new interesting technical developments (not covered in detail here) are becoming deployed. RFID technology for Circulation and security control involves replacing barcodes with “chips” embedded in labels. This “radio frequency identification” technology means that data can be modified on the book label without need for replacement. Raw label costs and replacement labor are going to be the issues here. And Dynix initiated a new Acquisitions protocol, VIP or “Vendor Interface Protocol” for expediting e-commerce with booksellers.

Between the ILS vendor’s software maintenance services and the library customer there exists a spectrum of optional support that takes the form of the vendor’s Application Programming Interfaces (commonly referred to in the industry as APIs). Prior appointments were scheduled with select vendor representatives to query them during the Conference regarding the extent of their customer support for the APIs.

Here we present a simple overview of the API issue, as we plan to publish later a more comprehensive tutorial that will incorporate more detailed information and specifications than interviews can yield.

For most libraries there is no need or inclination to modify software provided by the vendor. Libraries are generally considered to lack the technical resources to take advantage of the vendor’s software potential or to incorporate emerging technologies. At the same time, it is certain that there are some library technologists who are fully capable of modifying the vendor’s software, most commonly by extending services outside the core ILS modules. Furthermore, some libraries have staff knowledgeable enough about XML and HTML to participate in modifying and updating web pages as part of the Library’s Portal.

We do not know how many such technologists there are at present. More to the point, we don’t know how rapidly they will develop or become available as open APIs become prevalent. Certainly many service innovations are possible at the local library level. An example would be creating programs to enable a seamless online interface between the library’s Acquisitions and Circulation activity for exchanging purchase order and fines data with the local fiscal office.

Systems that offer well-defined APIs enable interested libraries to make improvements locally, without having to wait for a new software release from the vendor. In theory and practice, the more “open” the API, the more prone is the individual or library to use it. One can speculate that successful API-based implementations at the local level will be transferred broadly within the library community by the owning vendor.

This would even be likely where such APIs are based on “open source” models. More on this later, as we have asked select vendors to provide documentation, white papers, or other policy statements to enable Biblio Tech Review to provide a subsequent, more comprehensive tutorial on APIs.

The automation consultant takes library staff through a process of identifying local service needs and matching them to the “best” or most appropriate vendor(s). The library pays the consultant for experience, competence, and integrity, commanding primary loyalty to the employer. In a much broader context, the consultant ultimately performs a service for BOTH library and vendor by helping them to “make a market”; that is, it is the consultant’s role to help to bring together the library and vendor expectations with respect to functionality, performance, and price.

By bringing the library’s needs forward through a mutually understood dialogue, the consultant helps the library purchase the best service at an appropriate price. To this end, the library needs to understand the functionality offered by the vendors, and to assess the importance of this functionality. Then functional needs can be weighted according to immediate need vs. potential availability. The selection factors will weigh differently for each library, but it is in the matching of these common elements that “make the market”:

One widely-read source of information for those interested in the trends of library automation sales and marketing is Library Journal, which every April publishes a survey of the sales performance for automation vendors summarizing the prior year’s activity. Following years of reading and studying the published data, one could expect that a conventional rule of Marketing 101 would apply; that is, some 80% of the library automation business would be done by 20% of the vendors. In other words, it would take only 5-6 of some 25-30 mainline vendors to achieve $400Million of some estimated $500Million in the reported annual sales of ILS vendors.

By the metrics gathered here, the common perception that consolidation of ILS vendors in the library market does not appear to be happening. Rather, the market appears fractured and lacking in consolidation in spite of the impressions we may have that big fish are eating the little fish -- TLC acquired CARL, Sirsi acquired DRA -- to name two recent acquisitions -- and remembering not so long ago that Ameritech purchased NOTIS along with the original Dynix.

Any attempt to use the metrics is likely to encounter the complaint that the numbers are unreliable. Before dismissing these numbers out of hand, it is worthwhile to do the analysis and then subject it to the test of “sensibility”; that is, do the results conform in the main to people’s sense of what’s actually taking place in the market?

Those of us working in this market recognize

The Editor of Biblio Tech Review will consider publishing updates to the tables and revising editorial comment if there is sufficient feedback to warrant revision. Subscribers will be advised if there are changes. - Peter Evans, Publisher

that published numbers can also be misleading, by mistake or by intention. It would be realistic to consider that sometimes vendors fudge the numbers for their own purposes. LJ makes a valiant effort annually to track data, but their survey numbers are compiled by different individuals over time, under different rules for counting; some vendors do not provide comparable data year-by-year, and some place restrictions on what data can be published. The best example, perhaps, is the lack of clarity over the definition of “site”. One metric for a site is the “installed server”; another is the number of libraries sharing a server. When do you count a site as a “site”, when you sign the contract or when you complete the installation? Does a site get counted twice if it is migrating from a vendor’s older generation product to its next generation software? When is a site no longer a site? As a library moves from one vendor to another, does it get counted once by the current vendor and again by the installing vendor?

The data is not separated by type of library or geographically. Consider that some vendors sell to more than one type of library. As one vendor states it:

This is a study of “marketing trends” so we decided to build a set of numbers based on raw counts that could be submitted to the nine selected vendors for critique and commentary. During the Toronto ALA conference this year, all vendors were presented with our summaries of their data. The presentation was given with the understanding that the information provided could be subsequently reviewed by the vendors and would be published. The vendors were subsequently supplied with our immediate revisions and given the opportunity to verify accuracy and to make critical commentary. Specificity is our objective. We decided not to be interested in gross revenue estimates expressed in “ranges”. Here’s why. The range of $5Million is of some consequence between $70-75million in reported sales (comprising 3-4% of a total) but of major consequence when sales are reported in a range of $5-10million (comprising perhaps zero to 100%). Indeed, as a result many vendors/owners have come forward with more precise figures for given years and these are reflected here.

We took 2002 as our base year and worked backwards from experience and industry reports to 1999, showing 3 consecutive rollover years of sales revenues and installed sites. This meant incorporating the data of acquired vendors into the acquiring company as if they had been owned for the entire period. This would potentially show the effect of a merger in terms of overall market gain (loss) for the period. While arguments may persist on the details, the overall trends appear clear:

Clearly a market leader in the public library space. From 1999 to 2002, Dynix lost over 5% of market share, so the new management knows that it is on its mettle to perform for the libraries. 2003 may be a turnaround year because under the epixtech name the old Dynix we knew for some 15 years stumbled. Dynix gave permission to say that this version of their numbers is reasonably accurate.

What we can report with a degree of certainty is that Endeavor's revenue has grown an average of 5% between the years 1999 and 2002. By working backwards from an estimated $30Million in 2002, one can calculate these results. Here’s likely an example of “site counts” at year-end equating with number of libraries even as “new sales” for sales purposes likely equates to “new contracts”. The idea of “contract” equates in our definition more closely with “site”.

No comment. Numbers presented by non-quotable source.

The survey data for GIS prompted some interesting commentary by William Schickling, the new CEO. He pointed out that the GALAXY customer base is comprised largely of small libraries. He admitted that the company failed to move fast enough to consolidate these customers into the new Polaris product when first released in 1996 (an early bet on MS Windows NT before MS SQL was mature and prior to release of MS Windows 2000). GIS paid a price for being at the “bleeding edge” of development, as did VTLS with Virtua and DRA with TAOS. Recent customer surveys point to high satisfaction levels by GIS customers, indicating possibilities that the storm brought on by deterioration of service has been corrected. While the number of GALAXY customers moving to Polaris may be on the increase, there remains no hard evidence to suggest GIS is in a position to radically expand operations; thus they are likely to be confined to remaining a “niche” player for smaller libraries, at least until they prove scalability, which may result from their just-signed contract with Maricopa County (AZ).

The number of installed systems at the end of 2002 does not include Geac’s BookPlus customer base of 56 systems. 1999 revenues include an estimated $6million from a business segment selling access to Silver Platter databases. Geac’s policy does not allow the release of financial results for individual products or divisions; thus these numbers are quite “speculative”. One could expect Geac revenues and sites overall to rise after the launch of Vubis Smart in North America, expected soon, but no release date has yet been set.

At one time in the early days of automation (late 1970s-80s), Geac with proprietary GLIS circulation and PAC programs became a dominant force in both academic and large public library systems. The purchase of ALII’s ADVANCE (1989) and CLSI’s PLUS (1992) products seem to have gone nowhere, but the company has financial resources to make major marketing moves, should it decide to do so. Can such a Phoenix rise?

At the time of Sirsi’s purchase of DRA (circa August 2001), corporate management settled on the definition of site that conforms essentially to what is presented here. For the record, Sirsi counts each central server-based system as a single customer, even if that single server supports dozens or even hundreds of individual libraries. Sirsi’s example is that their PINES consortium uses a single server to support some 200 libraries—PINES is counted as a single site.

This method suits our purposes and we worked with Sirsi to create an amalgam of numbers that represents our interpretation of their market position under an assumption that Sirsi owned DRA back to 1999. To clarify this further, one has to accept that Sirsi has not created this template and would not be expected to re-state their numbers. Some of the DRA data comes from SEC public information.

If there is a clear market share winner it is TLC. While TLC does have a high percentage of its sites as public libraries, they have moved aggressively last year into the K-12 market as well. TLC did review these numbers and indicated no corrections needed to be made. Also TLC points out that potential customers always want to know how to sort out which vendors sell to libraries of their type.