Jared L. Howland

Report of the ALCTS Acquisitions Section (AS) Acquisitions Managers and Vendors Interest Group Meeting - American Library Association Midwinter Conference - New Orleans, June 2018

Howland, Jared L. and Sochay, Lee

Citation Information: Howland, Jared L. and Sochay, Lee (accepted 2019). “Report of the ALCTS Acquisitions Section (AS) Acquisitions Managers and Vendors Interest Group Meeting - American Library Association Midwinter Conference - New Orleans, June 2018” Technical Services Quarterly.

Introduction

The ALCTS Acquisitions Section (AS) Acquisitions Managers and Vendors Interest Group meeting at the 2018 American Library Association (ALA) Annual Conference in New Orleans, Louisiana consisted of a series of presentations and a question and answer session.

Good, the Not So Bad, and the Ugly of Usage Statistics

Rosemary Jarrell: Customer Experience Manager – ProQuest

Presentation file

Rosemary Jarrell reported on the data available to libraries from vendors for assessing collections. The most helpful data for assessing collections depends on what you are attempting to assess. For example, corporate accounts with ProQuest involve transactional pricing. That is, the corporations pay based on the number of searches run in a month. Therefore, search data is more helpful to them than retrieval of useful content. In academic libraries, the opposite is true—retrieval statistics are more meaningful than search statistics. Context should always be kept in mind when looking to assess content.

The Ugly of Usage Statistics

Usage statistics can sometimes be very confusing and complex. Some of the hardest aspects of usage statistics are as follows:

  1. Multiple Platforms : To gain the full perspective of your collections, data must be collected from lots of different places (COUNTER data and vendor-specific data are sometimes stored in different locations, data from other vendors need to be collected to make a meaningful assessment, etc.)
  2. Time Consuming : It is not only time consuming to gather the statistics, it can take lots of time to figure out how to correctly interpret what the data is telling you and how you can use the data to tell the story you need to tell.
  3. Inconsistent : Not all statistics are COUNTER-compliant which can make it confusing or impossible to compare different platforms. Rosemary’s suggestion for ProQuest materials was to use their “DB1” and “DB Summary” statistics for comparing across ProQuest products and COUNTER-compliant statistics for other comparisons.

The Not So Bad of Usage Statistics

ProQuest is working on a consolidated content and back-end platform that should make gathering ProQuest data much easier. ProQuest also has a large data center for all of the raw data across all ProQuest products. If you need a custom report, they can use the raw data to create reports for customers. Other vendors likely have similar capabilities. Finally, most vendors provide the ability to have reports automatically sent to you on a schedule convenient to your needs.

The Good of Usage Statistics

A new version of COUNTER is coming out which will provide even more meaningful reports. ProQuest, and other vendors, are here to help. They have a “Customer Experience Manager” that you can contact to get any help you need to analyze collections (multiple-year, multiple-platform usage for example).

Good, the Bad, and the Better Than Nothing of Usage Statistics

Rebecca Boughan: Continuations, Licensing, and Collection Analysis Librarian – Brigham Young University

Presentation file

Usage statistics are important because they can help prove the impact resources are having on your patron base. At a time when budgets are tight, this is critical. Brigham Young University is currently undertaking a project to compare cost-per-use data from vendor reports to document delivery fees for journal articles. If the cost-per-use is higher than the document delivery fee, they are looking to cancel that subscription and rely on interlibrary loan and document delivery to meet the needs of their campus users.

The Good of Usage Statistics

The following are helpful statistics when evaluating collections:

  1. Full content viewed : Data showing how many full-text views have occurred is one of the most helpful data points.
  2. Abstracts viewed : For abstract resources that contain no full-text, the number of abstracts viewed or number of times users clicked on a link to find full text are the most helpful statistics.
  3. Book Report 1 : Book Report 1 (COUNTER) is usually a better measure of usage than Book Report 2. COUNTER 5 will combine them which will be very helpful.

While COUNTER statistics are critical for comparing resources from different platforms and vendors, vendor reports can provide more granular data that can usefully supplement the COUNTER data. For example, some vendor reports can provide a subject-based breakdown which can be very helpful when sharing data with your subject librarians.

The Bad of Usage Statistics

  1. Misleading data : Sometimes vendors can accidentally provide wrong or misleading data. For example, a vendor once counted page views as section views in COUNTER Book Report 2 greatly overinflating the actual usage. This has since been corrected.
  2. Search statistics : Other ways data can be misleading are if you use a discovery layer. Discovery layers will perform one search but, depending on the vendor, will sometimes count that as a search in each database rather than a single search across the platform thus inflating the data. Even without using a discovery layer, there are problems with search numbers. Are patrons searching more because they are finding useful information or are they searching more because they cannot find what they are looking for?

The Better Than Nothing of Usage Statistics

  1. Anything the vendor can provide : Some data is better than no data. Any data you can get from a vendor is more helpful than not having anything at all.
  2. LibGuides statistics : If you use LibGuides, you can get statistics on which resources patrons are accessing from within LibGuides. Of course, this does not count any access from outside of LibGuides (Google Scholar, your discovery layer, etc.)

Extra Usage Statistics

There are other statistics that can be very helpful when you are assessing your collections. Examples include where your faculty are publishing and what publications are your faculty citing in their research.

Our Story: From Excel to Tableau and SQL

Esra Celtek Coskun: Collection Analysis and Planning Specialist – University of Illinois at Urbana-Champaign

Presentation file

Once you have usage data, you have to analyze and interpret it. Many times, this means cleaning and combining data sets. Fortunately, there are a lot of tools that can help with this aspect of a collection analysis.

If you need to clean up a data set, the first tool many turn to is Microsoft Excel. It has a lot of helpful functions, such as TRIM, CONCATENATE, VLOOKUP, and HLOOKUP, to help you clean data. Limitations include not being able to handle large data sets and limited data visualization options. A great resource to learn how to use Excel for library data is Margaret Hogarth’s book “Data Clean-up and Management: A Practical Guide for Librarians.”

Another helpful tool is Tableau Desktop. It can be expensive, especially if you purchase the training that can come with it. However, it allows you to transform data without editing or changing the underlying data set and live-updating of charts and tables when a data set is updated. You can easily exclude or ignore bad data points.

Esra has also been able to use SQL for analyzing data. It can handle very large data sets that can then be analyzed in Tableau Desktop. They have just begun experimenting with this and are looking into using ETL (extract, transform, load) processes to manipulate data in their SQL Server.

Finally, other tools that can be used include the following:

  1. Tableau PREP (included free for Tableau Desktop users)
  2. OpenRefine
  3. DataLadder
  4. Drake
  5. Trifecta

When deciding which tool to use, there are helpful factors to consider such as the following:

  1. The size of the data set
  2. How often you will be repeating the analysis
  3. How much data cleaning you are willing to undertake
  4. Limitations on your expertise and budget

Conclusion

More information about the interest group, and slides from these presentations, are posted on the interest group’s website on ALA Connect.