A SINGLE PIECE OF DATA
is worth a
THOUSAND OPINIONS

OVERVIEW | This company had several worldwide regions. There had been a huge investment in a major software change. After this change, users were unhappy with the response time to do job required functions. Some users were very unhappy. Transaction response times for various regions and types ranged from an average of 1 second to over 25 seconds. Some transaction times were much much longer. The types of transactions included: creating sales orders; entering production order confirmation; estimating material costs; and, many other essential business functions.

CHALLENGES | The problems had been elevated to business lead team levels. "Corrective" actions were largely based on a LIFO (Last In. First Out) approach. If there was any improvement it would maybe last for a few days, and then issues would return. There were many opinions. There were as many "causes and solutions" as there were people working on the problem. Taking time to analyze what was going on was "not available." A transaction tracking program had been purchased, but data gathered was largely confusing and generally ignored. Also, learning how to use this tracking program required an investment in time that was not available due to the increasingly emotional situation.

THE SOLUTION | Someone essentially said -- "Let's not just do something. Let's stand here." A methodical approach then began. When looking at the transaction tracking program several observations were made. The analytics and reports were weak, at best. However, the program offered the ability to gather raw data for date/time, user identification, location and region, program being utilized, device being called, time for transaction divided by queue time and processing time, and other factors. Moving this raw data into Microsoft Access and Excel provided the opportunity to do effective and efficient root cause analysis. Some key data-based findings and actions quickly emerged.

- There were a few easily identified users who did not know how to use the system. For example, they entered erroneous information. These users were trained and/or retrained.

- Some ad hoc programs had been written locally. When called these programs were largely linked to spikes in the system. These problematic routines were rewritten to be compatible with the new system.

- Running batch programs and rescheduling tasks were creating major system queue and application access times. Batch jobs were rescheduled to only run during non-peak periods.

- Also, storage device utilization was not balanced or managed well. When identified, the loads could be quickly adjusted.

OUTCOME | Queue time excursions were virtually eliminated. Response times became consistent and easily capable of meeting user expectations. The company then found time to develop real-time key performance tracking metrics. Continuous improvement in system usage and cost followed. The return on the large investment could then be realized.