Using Dashboard and Automated Scripts to Improve End-to-End Processes, From Ordering to Cataloguing

Purpose: The National University of Singapore (NUS) Libraries comprises eight libraries. The acquisitions and cataloguing functions are centralised in the Technical Services Unit. The Ordering and Cataloguing Teams... [ view full abstract ]

Purpose:

The National University of Singapore (NUS) Libraries comprises eight libraries. The acquisitions and cataloguing functions are centralised in the Technical Services Unit. The Ordering and Cataloguing Teams processes more than 10,000 orders and titles every year. Most of these publications are published in the United Kingdom and the United States.

Over the years, the teams had attempted to measure the performance of the acquisition operations. However, the procedures to generate the performance reports were tedious and data extracted was not current. Nonetheless, these reports showed that the ordering process for books from the United Kingdom and the United States took 6-8 weeks and the cataloguing process, another one week. In total, the end-to-end process cycle time took 7-9 weeks (49-63 days). Only 75% of the books ordered were received and catalogued within 50 days, against the Key Performance Indicator (KPI) requirement of 80%.

It was therefore necessary to improve both processes to reduce the cycle time.

Design, methodology or approach:

The Ordering and Cataloguing Teams utilised database query and visualisation tools to measure the performance of the operations and improve the processes.

Two developments aided the teams. Firstly, NUS Libraries upgraded its Integrated Library System (ILS) to Innovative Interfaces Sierra in 2013. It features Sierra Database Navigator (SierraDNA), a facility for querying the ILS externally using Structured Query Language (SQL). Secondly, NUS Libraries had purchased a visualisation and dashboarding tool providing for multiple database connectors and scheduled query, through an earlier project. The combination enabled the teams to construct dynamic queries and export ordering and cataloguing data from the ILS on an hourly basis. With data visualisation, the teams are able to identify bottlenecks, make improvements and monitor the situation. For example, we can now drill down using the orders turnaround report and retrieve the list of orders that did not meet the KPI, so that we can arrange for alternative.

In addition to visualising data for potential delays, the teams reviewed processes to eliminate unnecessary steps and or improve processes further to reduce the processing time. One such improvement was the automatic assignment of Library of Congress (LC) call numbers. The cataloguers identified a pattern for certain copy-catalogued titles where automatic assignment of LC call numbers is possible.

Findings:

Before July 2016, only half the titles ordered were received and catalogued within 50 days. With the dashboard, the Ordering Team can now easily generate the list of outstanding orders for more than 21 days and check the order status. The team can better decide whether the vendors are able to fulfil the titles in time, and take remedial actions where necessary. In addition, other reports allow the Ordering Team to check created orders before these were sent to vendors, ensuring that the orders were sent right the first time. The Cataloguing Team automated the construction of Library of Congress (LC) call numbers using Microsoft Excel. A scripting tool, AutoHotKey, was used to identify such titles, populate the list in Excel, generate the call numbers and insert the call numbers into the records.

The improvements resulted in 89.4% of orders received and catalogued within 50 days. 99.1% of received titles were catalogued within 1 working day, well beyond the KPI of 90% within 2 working days. Productivity and output quality improved; 74 man-hours was saved every month and error rates was reduced. The processes were well controlled and the results are sustainable.

Research or practical limitations or implications:

There were challenges encountered during the improvement efforts. Firstly, the SQL queries are complex and executions take a few minutes to complete. In addition, the dashboard uses a custom SQL-like language, which imposes a substantial learning curve. Secondly, the Ordering Team had to learn the new claims techniques. The team struggled to clear the initial long outstanding orders to bring the ordering process under control. Thirdly, the Cataloguing Team had to be trained to operate the AutoHotKey scripts. The scripts need to be run in stages and in a particular sequence; successful execution of one stage depends on the previous stage. The Cataloguing staff needs to be alert and correct the errors so that they would not be compounded as the scripts progress.

Conclusions:

NUS Libraries utilised a combination of ILS features and the dashboarding tool to produce a dashboard to better manage the book acquisition process and take action to meet the KPI. AutoHotkey Scripts and Microsoft Excel were used to automate the call number assignment for a segment of copy-catalogued titles and manage to exceed the target. These improvements to the end-to-end process enhances the user experience, shortens the wait time and provides for a more consistent service delivery experience to the patrons.

Originality and value:

This is a useful case study for the Technical Services team of other libraries if they are to consider implementing visual monitoring of their acquisition process, and for libraries which need to assign call numbers according to in-house practices.