Shaking Up the Norm: Journal Pricing and Valua9on ARL- CNI 21 st Century Collec9ons Forum October 13-14, 2011 Ivy Anderson California Digital Library ivy.anderson@ucop.edu
Bergstrom: Hell s Grocery Aspesi: A ShiB of Power Big Deals reinforce publisher monopolies Strategy: license more selec;vely, bargain hard, or walk away Library budget crises and usage- driven value assessment will undermine big deals This will adversely affect the fortunes (and profits) of large commercial publishers
Common theme: Analyze journal value and act on this informa9on UC Approaches: UC Value- Based Pricing CDL Weighted Value Algorithm
UC Value- Based Pricing Strategy: 2007-1) Alterna9ve base price that accounts for scholarly value and impact 2) Value- added contribu9ons from the purchasing ins9tu9on 3) Consor9al transac9on efficiencies 4) Annual price increase transparency How can we establish, validate and communicate an explicit method for aligning the purchase or license costs of scholarly journals with the value they contribute to the academy and the costs to create and deliver them? h"p://libraries.universityofcalifornia.edu/cdc/valuebasedprices.pdf
Value Based Pricing Strategy: Bergstrom- McAfee Rela9ve Cost Index Data from journalprices.com Composite Price Index (CPI): List price per ar9cle * list price per cita9on Rela9ve Cost Index (RCI): Compares CPI to non- profit journals in same discipline UC methodology Journals with poor RCI are progressively discounted Addi9onal discounts applied for UC authorship and consor9al efficiency Value Categories Poor Value Medium Value Good Value Elsevier Journals 594 ;tles (54%) 363 ;tles (32%) 153 ;tles (14%)
Problems we encountered Relies on impact factor not available for all journals Data were not being consistently updated (more current now) Value assignments based on list prices don t reflect actual UC costs Wanted a tool to help with evalua9on of journals and packages Need to facilitate consor9al decision- making for journal cancella9on and reten9on more informed decisions less subjec9vity eliminate labor- intensive 9tle vo9ng
New Approach: CDL Weighted Value Algorithm Assesses the value of individual journals according to 3 vectors of value Ideas we borrowed from Bergstrom- McAfee Numerical score with simple value designa9ons High, medium, low, lowest No9on of rela9ve value Recognize disciplinary differences * Source Normalized Impact per Paper (SNIP) - developed at University of Leiden and maintained by Scopus. hcp://info.scopus.com/journalmetrics/snip.html * Cita9on as well as impact data
Disciplinary differences macer Using Australian and New Zealand Standard Research ClassificaEon (ANZSRC), 2008
Weighted Value Methodology: Publisher A Publisher A Title Value: High, Medium, Low, Lowest
Value distribu9on: Sage Psychology Journals SAGE Psychology Titles Value Distribu;on Lowest 9 20% high 14 31% Low 7 16% Medium 15 33%
All Psychology Journals 100% 90% 80% TF Psychology Titles Value Comparison by Publisher Wiley Wiley Wiley Wiley UC Press 70% 60% 50% Springer OUP LWW Karger TF TF TF 40% 30% 20% 10% 0% Springer Springer Elsevier OUP Karger Project Muse Elsevier MAL Elsevier Cambridge Cambridge SAGE SAGE SAGE Springer MAL Elsevier Cambridge SAGE high Medium Low Lowest
Next steps Weighted algorithm currently being re- factored with new subject data Launching major bibliographer review across UC to validate rankings and trust the metrics Bibliographer input may result in further modifica9ons to the algorithm Current trial run is encouraging U9lize results for major renewals in 2013 For 9tle reduc9ons For nego9a9on purposes
Challenges Will publishers work with us to restructure and re- base our agreements? Recent UC example: License fee reduc9on: 8.5% Title cancella9ons: 32% Publisher methodology would have cancelled 43% Will this change the scholarly communicaeon system??