Just as I began to see messages about the publication of Marshall Breeding's report on his survey of library perceptions of their system vendor I was reading The new economics of the BI market by Jerry Held on The Database Column blog.
He talks about consolidation within the BI (Business Intelligence) market: "After more than a dozen acquisitions made by Business Objects, Cognos, and Hyperion over the past few years, these BI tools/analytics industry leaders were themselves snapped up in a matter of months by SAP, IBM, and Oracle respectively." And he notes the earlier consolidation of the underlying database industry around Oracle, IBM and Microsoft.
Held argues that consolidation has improved the overall BI marketplace. It delivers - he suggests - economies of scale and economies of innovation (and, although he does not mention it by name, economies of scope). These 'mega-vendors' offer a range of products. For some customers, the ability to concentrate interaction with a single vendor, a single helpdesk, and a single contract, and to benefit from discounts, are important benefits. For vendors, it should be possible to remove redundant costs in administration and distribution. Competition between a small number of dominant players is good for the market.
He suggests, however, that the mega-vendors find it difficult to innovate or meet new needs; they have a very full array of products spread over a large customer base. This means that there will always be investment available to new entrants who innovate around technology or business models to meet evolving needs.
He points to open source and SaaS (software as a service) as two important business model innovations. He also provides some technology innovation examples, emphasizing performance and price improvements.
Does this map onto the process automation providers within the library community? Here are some thoughts, focusing on the US environment. (And, full disclosure, OCLC has some offerings in some of the areas I discuss below.)
There has definitely been consolidation within the classic ILS environment. This is good in principle, as the library market - not very big to begin with - has been overpopulated with vendors trying to provide a full range of products. In practice, of course, much depends on how the remaining vendors work through integration issues. We can see some potential economies of scope (as diversifying library needs can be met from a single source) and scale (as development, support and R&D are consolidated).
However, none of these vendors is very large, they operate in a small community, and they have limited organic growth opportunities in their historic core. They have moved to meet diversifying library needs with additional products. Accordingly, we have seen that process automation for the 'bought/physical collection' (the ILS) has been joined by process automation for the 'licensed collection' (metasearch, resolution,knowledge base,ERM), and the 'digital collection' (repositories). Other products have also appeared to meet more specific needs (self-service, e-reserves, ...). Recently, a new category of discovery system has emerged which pulls together institutional data (from the ILS and from repositories), and several products have appeared. Now, each vendor has a significant development challenge in creating this full array of products, and we have seen some licensing of other components (support for metasearch or knowledge base, for example). Interestingly, we have not seen these companies acquire new entrants who are also developing these newer products (more of these below).
And, although we have seem some libraries acquire pieces from different vendors this is not as widespread as one might expect for some of the reasons suggested above. There are economies in dealing with as few vendors as possible. In addition, the library community has quite a personalized relationship with its ILS vendor community which adds to the incentives to acquire various components from the same vendor.
Marshall suggests that 'dissatisfaction and concern prevail' in this marketplace. I think we can expect further consolidation, as the number of vendors here reduces to two or three, maybe with particular specialties.
What about innovation? There is some concern that there has been little innovation in the classic ILS space, which matches Held's observation. That said, we can point to Ex Libris's collaboration with Herbert Van Der Sompel around the deployment of resolution as a service as a notable instance, or experimentation with ERM. It is not surprising that as new areas have been identified we have seen a range of new entrants, sometimes emerging from within the library or academic community. See for example Serials Solutions, which aims to provide a complete approach to licensed collections. The metasearch and resolution arena has seen several companies emerge, some of whom syndicate services to other players. See for example Muse Global, Openly Informatics (now part of OCLC), WebFeat or TDnet. And more recently, as we have seen attention to better discovery environments, Aquabrowser is being deployed by some libraries.
One area where innovation has been slow is in how the library systems apparatus engages with the tools that people are increasingly using to organize their own information spaces, at the browser level, or in social bookmarking, social networking, and other network-level sites.
Business model innovation? Held mentions Open Source and SaaS (Software as a Service). We have seen two major areas of open source development. The first is in the area of repositories, where we see Fedora, Dspace, and Eprints. The effort involved in deployment here may be high. Each initiative has gone through some organizational development, looking for ways to sustain itself, and the role of grant/foundation money has been important. The second is in the ILS arena, where Koha and Evergreen are receiving a lot of attention. Koha is more widely deployed; there have been some recent high-profile commitments to Evergreen. There are also some other areas where open source solutions are in use: metasearch (e.g. Index Data, LibraryFind), text searching (e.g. Lucene, Index Data), and a recent interest in 'next generation catalog' solutions (e.g. Solr, Vufind). Index Data has been active for a while, with a strong niche presence in Z39.50 applications and text searching and metasearch offerings. One interesting development is the emerging support industry here, where Care Affiliates, Index Data, Equinox and LibLime will offer support and consultancy. It will be interesting to see how this range of activity develops in coming years. In part it will probably depend on the ability of this nascent support industry to meet mainstream library requirements for support and reliability; and in part of course on the ability to continue to develop the software.
And what about SaaS? SaaS tends to be used quite loosely. Think simply of three levels. The first is where individual instances of an application are hosted. This may save the library some costs (hardware, sysadmin) but does not really alter the service model in other ways. A second is a 'multi-tenancy' model where multiple customers may be served from the same instance, but each with their own virtual application, potentially with configuration options. This may deliver savings but there may also be service improvements. Enhancements, fixes, etc, are available to all at the same time. Serials Solutions' services might be an example here. The third level becomes more interesting where shared use of a service generates network effects. Take a hypothetical example: a supplier could more easily develop recommender systems across multiple circulation systems. An actual example appears to be provided by Aquabrowser's announcement of its MyDiscoveries feature which aims to share user contributions to the catalog across customer instances. The SaaS model has been rapidly adopted in wider contexts, and while there has been some library adoption, it is interesting that there is not a high level of discussion of the approach.
The year 2007 saw considerable upheaval in the library automation industry. To get some sense of the aftermath of the recent rounds of mergers, acquisitions, product consolidations, and to gauge interest in open source automation systems, I created and executed a survey that aims to measure the prevailing perceptions in libraries. [Perceptions 2007: an International Survey of Library Automation]
What is interesting to me is the extent to which the ecology of library process automation is richer than it was a few years ago. If we think of managing three materials workflows (bought/print, licensed/electronic, digitized/digital), and the progressive movement of libraries into the latter two, then we see that library needs are now potentially met by a wide number of players. The classic ILS vendors remain central players, but they have been joined by others.
The ILS vendors have products in all three areas, and are developing new discovery products. We have seen new entrants in the repository space (including ContentDM, now owned by OCLC) and in the licensed materials space (resolover, knowledgebase, metasearch, ERM) where a variety of products are available from a range of vendors. In this context, the collection of services within the Cambridge Information Group is interesting (Serials Solutions, Refworks, Illumina, Aquabrowser as well as other bibliographic products). And, of course, OCLC provides services also. Open source offerings have emerged to meet needs across the board.
We will definitely see more convergence alongside further new entrants. It will be interesting to see how the Open Source offerings develop, and I think that we will see some game-changing offerings in the SaaS space.
I hope Marshall repeats the survey. It would be interesting to extend its scope - if that can be done without too much loss of focus - to consider more of the wider process automation landscape.
- The network reconfigures the library systems environment
- The integrated library system that isn't
- New model library system vendors
- The user interface that isn't
Pointer to MyDiscoveries via Meredith Farkas.