Usage statistics: taking e-metrics to the next level

Download Report

Transcript Usage statistics: taking e-metrics to the next level

Project COUNTER
Trends in Statistical Standards for EResource Management
March 2005
Oliver Pesch
Chief Strategist, E-Resources
EBSCO Information Services
[email protected]
Overview

COUNTER now and future





Beyond COUNTER



The COUNTER basics
Release 2
Auditing vendor reports
E-Reference
Where A&I usage fits
Other important metrics
Things to ponder



Activity captured – is it meaningful?
Effect of metasearch on usage
Capturing the data
COUNTER


Member supported
Members include:




Libraries
Publishers
Aggregators
Goal:

Allow credible and consistent usage
measurement between vendors
http://www.projectcounter.org
Code of Practice
Compliant Vendors
Rules for compliance
Upcoming releases
COUNTER Code of Practice

Addresses:





Terminology
Layout and format of reports
Processing of usage data
Categories
Delivery of reports
Terminology



Ensure consistent use of terminology
Rely on existing standards for definitions
NISO Z39.7 (Usage Statistics)
COUNTER Usage Reports

Journal Report 1:


Journal Report 2:


Total Searches and Sessions by Month and Database
Database Report 2:


Turnaways by Month and Journal
Database Report 1:


Full Text Article Requests by Month and Journal
Turnaways by Month and Database
Database Report 3:

Searches and Sessions by Month and Service
Journal Report 1:
Full Text Article Requests by Journal
Journal Report 2:
Turnaways by Journal
Database report 1:
Searches and Sessions by Database
Database report 2:
Turnaways by Database
Database report 3:
Searches and Sessions by Service
COUNTER: Data Processing



Only count valid requests
For progressive PDF retrieval, only count the
first request
Filter out multiple successive clicks (“double
clicks”) of same link by same user


10 seconds for HTML
30 seconds for PDF
COUNTER basics

Q: Who gets to count full text requests?





Publisher?
Aggregator?
E-Journal gateway?
Link Resolver?
A: The party that delivers the actual text to
the user
User finds citation of
interest in an A&I
database. An
OpenURL link
provided
Linking
Citation
A&I
Database
Link
Resolver
Link
menu
Abstract
or TOC
E-Journal
Gateway
Full Text
Publisher
site
Link-Out
Link-Out
Link-Out
Full text
User finds citation of
interest in an A&I
database. An
OpenURL link
provided
Linking
Citation
A&I
Database
Link
Resolver
Link
menu
Full Text
Abstract
or TOC
E-Journal
Gateway
Publisher
site
Link-Out
Link-Out
Full text
Delivery







CSV, Excel or file that can be imported into
Excel
XML version being tested
On password controlled web site
Scheduled alert or delivery of report
Provide monthly
Available within 2 weeks of end of month
Current and previous year’s data
Compliant vendors/products
Compliant vendors/products
COUNTER basics
Any questions on current COUNTER
Code Of Practice?
COUNTER Release 2



Minor fixes and improvements
New report (Journal report 1a)
Level 2 report eliminated in favor of “optional”
reports
Release 2: minor fixes
Added header rows to
identify, report, criteria and
run date
Release 2: minor fixes
Include Publisher field
Release 2: minor fixes
More forgiving date
representation for Excel.
Jan-2004 instead of
Jan-04
Release 2: minor fixes
“Calendar YTD” now
“Total”
Release 2: Journal Report 1a
“Page Type”
column added
to breakout
activity by full
text format.
Release 2: Journal Report 1a
Two rows per
title
COUNTER: Audit



Passing audit will be required in 2005
Conducted by auditor certified by COUNTER
At vendor expense
COUNTER: Audit

Audit will check each product and report for:




Layout
 Right rows and columns
 Header rows
 Column headings
Formats
 CSV or Excel
Delivery of report
 Email notification of availability
 Access from password controlled web site
Accuracy of data
 Conduct series of tests for each report
 Current tolerance is -8% to +2%
COUNTER: E-Reference




The next initiative?
Focus to-date have been on journals
E-Reference includes:
 e-books
 Dictionaries
 Encyclopedias
 almanacs, etc.
E-Reference materials becoming increasingly
more important.
COUNTER: E-Reference

Unit of access may include:






Entire book
Chapter
Entry (dictionary or encyclopedia)
Page
Paragraph
Challenge is what constitutes a countable
“access”
COUNTER: E-Reference

Access depends on interface and
organization of content.




Entire reference work may be in one PDF (1
access and user reads all of the book)
Each chapter may have own PDF (1 access per
chapter read)
Reference may organized by topic or section with
user linking from topic to topic within a chapter
(many access within one chapter)
Discussions continue…
Looking ahead
(future releases?)
Where A&I Statistics fit



Metrics for product purchase and usage
patterns
Journal-level use metrics help with collection
development
COUNTER represents only in optional report
COUNTER optional report
Other important metrics

Link activity




Where the users came from
Where the users go
By target
By type of target
Example of link activity
User finds citation of
interest in an A&I
database. An
OpenURL link
provided
Linking
Citation
A&I
Database
Link
Resolver
Link
menu
Abstract
or TOC
E-Journal
Gateway
Full Text
Publisher
site
Link-Out
Link-Out
Link-Out
Full text
Link-out activity


Record each time user links from an item to
full text or other service
For each, track:

Source information




Database
Journal/Book
Year published
Target information


Category (full text, ILL, etc.)
Target service (domain)
Year of publication





Record with each full text retrieval
Summarize retrieval activity by journal by
year of publication
Analyze importance of back files
Analyze who using back files
Make archival decisions
Type of material





Journal article, book, book chapter, video,
sound track, encyclopedia, dictionary…
COUNTER moving towards E-Resource code
of practice
Record with each TOC, citation, abstract or
full text request
Summarize retrieval activity by type of
material
Make decisions on value of non-journal
information in hybrid collections
Article level data?



Volume versus value
Minimum data to store?
Is roll-up possible to achieve goals?




Section
Subject (journal level?)
Author
What transactions need this level of detail?



Full text requests
Abstract views?
TOC views?
Other things to ponder
Metasearch

Also known as:







Federated search
Broadcast search
User presented with a single search interface
Searches multiple information sources at
same time
Eliminates (or assists with) resource selection
Retrieve, consolidates and ranks results
A new challenge for statistics
Metasearch: a new challenge

What can be affected



Session counts
Search counts
Why?



“Search all” option or automatic selection/search
of many resources
Perform simultaneous activities
Optimization techniques
Searching without Metasearch
Resources
use
r
EBSCO
-Product 1
-Product 2
-Product 3
…
OCLC
ProQuest
Visits
=1
Sessions = 2
Searches = 2
OVID
With Metasearch engine
Metasearch
use
r
EBSCO
-Search All
-Business
-Medicine
…
OCLC
ProQuest
Visits
=1
Sessions = 20
Searches = 20
OVID
With Metasearch engine
Metasearch
use
r
EBSCO
-Search All
-Business
-Medicine
…
OCLC
ProQuest
Visits
=1
Sessions = 28
20
Searches = 28
20
OVID
Metasearch lessons

Libraries should want to isolate metasearch
sessions and searches




Capture source of activity
Isolate metasearch IP address
Have metasearch access through separate
account
Support metasearch standardization activities
through NISO

We need to be able to recognize a metasearch
session
Summary of COUNTER







Collaboration of vendors and libraries
More than just a standards document
A code of practice
Enforcement and registration compliance
Setting reasonable expectations
Benefits extend to all parties
COUNTER is very active and forward-looking
Thank you
Oliver Pesch
[email protected]