HOLLIS Portal Fiscal Year 2003 Usage Statistics
Scope
These statistics report the number of connections to Harvard Libraries portal resources for the twelve-month period July 2002 - June 2003. The total number of connections reported for the year was 3,041,935. This represents a 40% increase over the previous year's figure of just over 2 million.
Note that usage statistics are only recorded for resources whose access is routed through the Harvard University Library proxy server. These resources are identified by the letter I (Valid Harvard ID required) in the Harvard Libraries portal. Reports are not available for unrestricted resources and resources listed as C (Campus-only access).
Please see the sections below for additional information about these statistics and report any problems or omissions to Ivy Anderson.
Go to statisticsTypes of Connections Counted: Direct Logins and Deep Links
Beginning in FY2003, the connections that are counted are of two types: direct logins and deep link accesses. Below is a snapshot of totals in each category, followed by more detailed explanations.
| Type of Connection | Number | Percent |
|---|---|---|
| Direct Logins | 2,709,790 | 89% |
| Deep Links - Unattributed | 278,945 | 9% |
| Deep Links - Attributed (via SFX) | 53,200 | 2% |
| Deep Links - All Types | 332,145 | 11% |
| Totals | 3,041,935 | 100% |
A direct login is counted each time a successful connection is made to the entry page of a given resource – i.e. the page that is linked to by the Harvard-assigned URN. These logins are counted regardless of whether the connection is made via the Harvard Libraries portal, from a Harvard library (or other) web page, or through an individual user's bookmark for the resource.
A deep link access is counted whenever a connection is made to an individual article or other component item of the resource, either directly using a proxy-aware url, or via Harvard's SFX service.
Deep links that do not use SFX generally cannot be attributed to the resources to which they belong. In all, these unattributed deep links (identified in the tables as 'deep links to Portal resources') accounted for 278,945 connections, or 9.2% of all use. This represents a higher level of use than any single resource.
When SFX is used to facilitate the connection, an attribution to the correct resource can usually be made. Attributable SFX-based links comprise another 53,200 connections. SFX links that cannot be attributed to a specific resource are counted in the total statistics for unattributed deep links above. SFX was made available to the Harvard community in January 2003; these statistics therefore represent approximately 6 months of usage data.
Taken together, attributed and unattributed deep links accounted for nearly 11% of all FY2003 use. This figure is evidence of a continued rise in item-level linking, generally in the form of direct access to journal articles. These links are generated by electronic reserves and article linking from course web pages, by full text links from one resource to another via SFX, and by end-user bookmarks to individual articles or other entries within a journal or database.
The effect of the deep links generated by course reserves and web pages is probably not uniform across either resources or faculties. Deep links can be created for any resource that supports persistent URLs, but some resources, such as LexisNexis, do not currently provide this feature. Differential adoption among the Harvard faculties of electronic reserves and similar services may also affect the percentage of faculty usage that can be accurately attributed to a given resource.
Resources Added During FY2003
A total of 1,210 new resources were added to the portal in FY 2003, including some 1,150 electronic journals. For these resources, the statistics that have been gathered represent less than a full year of usage data. To assist libraries in interpreting the results, the number of months that each resource with partial year data has been available online is displayed directly beneath its title in the detailed usage table for that resource (see example at Academic Search Premier). When comparing usage among resources, it may be useful to divide the totals for each resource by the months online to arrive at comparable monthly figures. If the number of months online is not displayed, the resource was available for the entire twelve month period.
Users with More Than One Faculty Assignment
A small percentage of all logins during FY2003 were by individuals who are recorded under more than one Harvard faculty in the University's administrative systems. This may represent, for example, persons who are employees of one faculty and students in another, faculty who teach in more than one school, students pursuing multiple degrees, or similar circumstances (up to three such affiliations can be recorded). In consultation with the libraries, these sessions have been divided equally among the relevant faculties as fractional numbers. Thus, if a particular Medline login is associated with both the Medical School and the Faculty of Arts and Sciences, it will be recorded as .5 for each faculty.
Undetermined Faculty Assignment
Less than 1% of all user logins during FY2003 lack a faculty assignment in the portal logs. This problem may result from incomplete data in the University's LDAP database (e.g. due to a reporting time lag for new ID card holders), or from some other unidentified factor. In order to provide as complete a picture of portal resource activity as possible, sessions which lack a faculty assignment are included in the session totals for each resource and vendor, with the caption "Undetermined." Note however that the percentage of overall use reported for a given faculty in the detailed tables by resource and vendor has been calculated against the total use by known faculties only (excluding undetermined logins), since a chief goal of the faculty breakdown is to assess relative usage among Harvard faculties for cost allocation purposes.