UKSG SERIALS AUTOMATION SURVEY 1994

Peter Walsh is Sub Librarian at King's College London, Campden Hill Road, Kensington, London W8 7AH; Tel: 071 333 4498 The fourth UK Serials Group Automation Survey was carried out in December 1993 and January 1994. A four page questionnaire was sent to a total of 420 libraries, which included all library members of the UKSG in the UK and Ireland (a total of 248, and also to 172 nonmembers who had expressed an interest in the Group. These additional libraries were included in an attempt to broaden the scope of the survey, and particularly in order to cover more special libraries. Replies were received from 131 members and 47 non-members, giving a response rate of 53% for members and 27% for nonmembers. The overall response rate was 42%. This was slightly lower than the 48% response rate of the previous survey, carried out five years ago1, although the total number of respondents was greater. 1 General


General
Among the respondents there was a wide range in both the size of their periodicals collections and in the amount of money spent on them. As might have been expected, member libraries in general spent more on serials and had a greater number of titles than nonmembers, and 80% of them, as opposed to 60% of non-members, had some form of automation.
Of libraries overall, 46% took fewer than 1000 titles while 14% had more than 4000. In fact seven had more than 11,000 current titles. Expenditure was not as closely linked to number of titles as one might have antiapated. Only 20% of libraries spent less than E30,000, while 24% spent more than E300, OOO.
There were resknses from 85 academic libraries, 27 industrial libraries, 19 research libraries, 14 government libraries, 12 medical libraries, four society libraries and two public libraries. A further 14 libraries were classified as "other". These figures mean that, compared to the previous survey, there were more respondents in every category except government libraries and public libraries.

Extent of automation
Forty-five of the respondents (25%) had no serials automation at all, although 29 of them were considering installing a system. As might have been predicted there have been considerable changes from the situation at the time of the 1989 survey (see Table 11, with a large decrease in the percentage of libraries without any serials automation. There was an even more marked fall in the percentage with an in-house system, with a corresponding increase in those using a commercial package. Most notable is the relative growth in the number of libraries using serials systems from subscription agents. In the analysis of the previous survey there was a division of systems from vendors other than agents into "Co-operatives" and others. This distinction is no longer really valid and all have been grouped here as "Other". One hundred and thirty three libraries (75% of those responding) reported having some form of serials automation. The extent of this varied from the recording of serial titles and locations on an Online Public Access Catalogue (OPAC) as part of an integrated library automation system to the use of a (only a third of the total), four society libraries (100% of respondents) and one of the two public libraries.

Variety of systems
Twenty-three different serials automation systems were in operation in the libraries which responded to the questionnaire. This is excluding systems developed in-house, which were in place in 13% of automated libraries. The most widely used of the systems supplied by subscription agents was Blackwell's Isis which was used by 13% of libraries. This was followed by Dawson's SMS in 6%.
Among systems supplied by organisations other than agents, Libertas from SLS was the most common, being in use in 13% of libraries. The vast majority of these, however, were not using the Libertas serials module, merely displaying their serials holdings on their OPACs. Systems supplied by BLS were present in 10% of automated libraries and the Dymx system was in 6%. CAIRS, GEAC, Soutron and TECHLIB systems were each present in around 5% of libraries. Table 2

Catalogue
Since bibliographic control of the serials collection is basic to other aspects of automation, it is not surprising that the catalogue was again the most commonly automated activity, a computerised version being present in 86% of libraries. This varied by type of library, from 66 out of 69 academic libraries to 11 out of 18 industrial libraries.

Display of serial titles on an OPAC
Given the above, it would be expected that OPACs would be reasonably widespread, and in fact, in relative terms, they have become even more popular than in 1989, being present in 71% of libraries. Academic libraries have obviously placed more importance on this aspect of automation, as shown by the presence of OPACs in 60 of them, compared to six among the industrial libraries and four out of fourteen government libraries.

Union List
67% of libraries made use of some form of automation to add data about their serials holdings to a union list. Research libraries did this in 11 out of 14 cases, while academic, government and industrial libraries did it in 48, ' eight and ten cases respectively. The next most widespread function was automated check-in, which was in operation in 64% of libraries, the biggest relative increase since 1989. Here government and research libraries were to the fore, with about 92% of each having check-in automated, compared to 52% of academic libraries. It is notable, however, that only 5% (or seven out of 85 libraries) used barcodes to check in periodical parts. This is obviously an area where we should expect to see considerable advance in the future.

Claiming
Automated claiming, at 56% has also increased greatly over the last five years, as subscription agents have no doubt noticed. It was taking place in 35 academic libraries (a relatively low proportion), ten industrial, nine government, and 11 research libraries.

Financial control
This was automated in a slightly smaller percentage of libraries than in 1989, somewhat surprising given the importance of this aspect of serials management. The overall 47% includes variation from 58% of academic libraries to a ' third of industrial and 29% of research libraries.

Management information
One would hope that an effective automated system wodd allow access to information about the performance of agents, prices in different subject areas etc. in order to aid management decisions. Here, however, with automation in 43% of libraries there was no great change from 1989. Industrial libraries had this function automated in I I cases, and the lowest incidence was in govenunent libraries where there were only five with this operation automated.

Ordering
In 41 % of libraries automation included ordering of serials, with just more than half having this activity automated in all categories of library except academic, where it was only 35%.

Routing of periodical parts
Many libraries circulate (or route) new periodical parts to selected members of staff. This process was automated in 36% of libraries, ranging from 67% of industrial libraries, where this is obviously an important activity, to 20% in academic libraries where the practice of routing periodicals is less common.

Lending
30% of libraries had lending of periodicals automated, a smaller proportion than in 1989. The variation was from 50% in research libraries to 20% in academic libraries, although, of course many libraries of all types do not lend periodicals at all. It is notable that in almost all of the academic libraries where lending of periodicals takes place, the loan status is displayed on an OPAC, whereas this is not the case in most of the other categories of library.

Binding records
These were automated in only 8% of libraries, and this proportion was about the same in all library types. In almost all of these cases, however, the binding status was displayed on an OPAC.

Electronic communication
It may seem hard to imagine that in 1989 only four libraries that replied to the questionnaire used e-mail to contact subscription agents. The number has increased to 40 in 1994, but this is still relatively low, representing just 30% of the automated libraries, and presumably e-mail communication is something that will grow considerably in the next few years. On the other hand, 59% of automated libraries have access to agents' databases, with Blackwell's Connect and DataSwets being the most common. The use of electronic means to load invoice data, however, is still in its infancy with only four libraries employing this technique.

System satisfaction and migration
The average length of time that libraries have been using their serials systems is just over three years. When asked whether they would choose their present system again only 29% replied that they would, 37% were not sure and 20% said they would change (some did not answer the question). Those saying that they would choose their present system again varied from 64% of research libraries to only 19% of academic libraries.
Perhaps this explains why migration from one system vendor to another, a phenomenon that was virtually unknown at the time of the previous survey, is becoming more widespread, with 20% of libraries reporting that they have changed their automated system. Certainly, as automated systems advance, upgrading will become necessary in many libraries and it is important that this can be done with the minimum of additional work for serials librarians.
A total of 44 libraries had an integrated library automation system of which they were not using the serials module. This includes 29% of the libraries which reported being without any serials automation at all, as well as 23% of the automated libraries. Among the latter, as well as those who have purchased a separate serials package, are those whose automation is limited to using, for example, the OPAC facilities of their integrated system to display serials holdings data. Why are such a sigxuficant minority not using the serials module? Not everyone answered that question, but by far the largest number (20 out of 44) replied that it was because they did not like the serials module of their integrated system.

Planning and setting up the system
Only nine libraries used a consultant when planning and developing their serials automation, and five of these were industrial libraries. Most consultants used were internal to the organisation rather than external.
The questionnaire asked for an estimate of the amount of data that was entered manually and the amount that was imported. In the libraries that answered this question, an average of 80% of data was entered manually, with only 12% being entered by the supplier and 8% by a third party. Despite the amount of work involved in this, and perhaps reflecting the hard-working nature of serials librarians, only 25% of libraries reported using extra staff to enter this data.

Conclusion
In retrospect it appears that it would have been useful if the questionnaire had enabled a clear cut distinction between libraries with a fully implemented serials package and those where only some aspects are automated, either as part of an integrated system or otherwise. Nevertheless, the results are an interesting reflection of the changes in the serial automation scene over the last five years. Although the proportion o f libraries with some form of serials automation has grown from two thirds to three quarters, there is obviously a long way to go before all libraries with serials collections large enough to profit from automation have a system that serves all their needs to a reasonable level of satisfaction.
One major source of inconvenience appears to be the need to enter data manuallly when many systems are installed, and it is to be hoped that this problem can be overcome before the next survey. Perhaps it would also be helpful if some producers of integrated serials systems paid more attention to the needs of serials librarians in designing or modifying their serials packages.

The database
Any member who would like further information from the complete database, or from sections of the database sorted by type of library, is asked to contact Peter Walsh on 071 333 4498. It should be noted, however, that in certain instances the name of the responding library will remain confidential. A considerable number of the respondants expressed a willingness to advise new and inexperienced users of particular systems, and details of these are also available.