3. 3
Introduction - Opportunity
• Limitations of Software Engineering Quality Assurance Processes
o Focused on “requirements elicitation” and “hunting of bugs”.
o Skewed towards artifacts.
o Error Prevention is more like defect-solving, instead of focusing on how the
defects get introduced in the first place.
• Root Cause Analysis (RCA) can uncover human performance issues
but is insufficient for indepth analysis of human error causal
factors
• B.W. (Ben) Marguglio’s Human Error Causal Factor Taxonomy is a
technique that can be used to:
o Understand human performance and human error prevention terminology
o Classify human error from different perspectives
o Recognize error-inducing conditions / error-likely situations and understand and
use behavioral techniques to counteract these conditions / situations
o Understand and recognize human error causal factors
o Understand how to measure human error
o Design and implement a Human Error Prevention Program
4. 4
Introduction – Scope and Audience
• Scope
o This paper discusses the coupling of RCA and Marguglio’s Human Error Causal
Factor Taxonomy to reduce human error based quality issues; a case study and
derived benefits
• Target Audience
o Those involved in Project Management Functions
especially those engaged in long term maintenance projects
o Root Cause Analysts assessing Human Performance Issues
especially on Business Process Outsourcing (BPO) side
5. 5
Problem Definition / The Challenge
A Maintenance engagement wherein the artifacts are templatized and processes well-tested
over time, but yet the errors were many.
Analysis of the monthly performance metrics –Work Request (WR) Rating, Resource Utilization
and Defect Density; highlighted the soaring defects resulting in a slump in customer ratings.
RCA highlighted the human errors, but without an in-depth analysis it would not be possible to
prevent them.
The following case-study details the engagement, key challenges, results of RCA, usage of
Marguglio’s Human Error Causal Factor Taxonomy and the derived benefits.
6. 6
Case Study: Engagement Overview
• Customer
o Usability and Corporate Websites Management Team within a large global Financial
Services Company
• Project
o The customer company had merged with another financial services company. The merger
called for an update in all corporate websites.
o The Usability and Website Management Team’s responsibility was to ensure that the
changes do not hamper the customer experience.
o But, their focus had shifted to creation Documents, Wire frames; and activities like
Content Updates and User Interface Testing.
o To get back the focus, offshoring artifact creation and testing to MphasiS was a valuable
proposition for the customer
7. 7
Case Study: Key Challenges
• Engagement Firsts
o Build a dedicated offshore assembly-line model, run by a pool of specialists; to support the
ongoing user experience enhancement engagements run by the customer.
o Mixed bag of skill-sets comprising of Team Leads (onsite and offshore), Technical Writers,
Testers, Wireframe Developers. The Usability practice did not have Technical Writers and
Testers onboard. Besides, the team leads required specialized project management skills
coupled with usability consulting experience.
o 24*5 support -
--16 hours offshore support (two shifts) with 3 hours overlap
-- 8 hours onsite support for WR management and task allocations
-- Service Level Agreement (SLA) based usability delivery wherein each WR is rated by the customer
8. 8
Case Study: Key Challenges
• Recruitment and Training
o Hire a team of specialists
o Training – Induction, Customer Processes, Templates, Quality Monitoring and Reporting,
SLAs etc.
o Deliverables and Documentation
• WR Processing
o Interpretation of WRs
o Managing emergency requests / spikes
o Less turnaround time for deliverables
o Reduced Review Time
o Lack of environment support during offshore hours
• SLA Management
o Quality of Deliverable – Error Free
o WR Response Time
o Task Time Effort
o Resource Utilization
9. 9
Solution – 4 Step RCA
Step one—Data collection
• A dataset comprising of WRs rated “Below Expectation” or “Needs
Improvement” was created.
• Pareto Analysis was conducted on the dataset, steps followed in this analysis:
o Listed the problem areas/ issues on each of the WR
o Grouped options where they are facets of the same larger problem
o Applied appropriate score to each group
o Worked on the group with the highest score
• Pareto Chart
• 44% of the “Below Expectation” ratings were due to Human Performance issues
• 75% of the low ratings were due to Human Performance and Communication issues
10. 10
Solution – 4 Step RCA
Step two—
Analysis
• “Causal Analysis –
Fishbone Diagram”
was used to
organize and
analyze the
information
gathered during
the investigation.
• The Marguglio’s
Human Error Causal
Factor Taxonomy
was implanted on
the Fishbone
Diagram to identify
the underlying
reason or reasons
for each causal
factor.
12. 12
Solution – 4 Step RCA
Step three —Decision Phase
• The objective of the Decision Phase was to chalk out an achievable action
plan with short-term and long-term objectives.
• Proposed Action Plan
13. 13
Solution – 4 Step RCA
Step four —Implementation
• Offshore Lead (who undertook the RCA) assigned each action item to relevant team members and
tracked them till completion.
• Poka-Yoke (fool proofing) basis of the Zero Quality Control (ZQC) approach, a technique for
avoiding and eliminating mistakes was implemented.
• The tools are applied within a simple performance-improvement framework known as DMAIC, or
quite simply state the ‘define-measure-analyze-improve-control’ principle.
• The entire root cause analysis will be undertaken quarterly.
• Significant quality enhancements were seen in the following monthly metrics with conspicuous
benefits (see Benefits for details).
14. 14
Benefits
Qualitative Benefits
• Customer satisfaction and higher WR rating
• Improved quality of deliverables / Reduction in Errors
• Productivity Improvement
• Team satisfaction and ownership of quality monitoring
Quantitative Benefits
Graph1: Customer Satisfaction Index (CSI) has improved from 3.1 to 3.35 (8% Improvement in
CSI). Trend shows increase in CSI with fewer 'unclear' WRs
15. 15
Benefits
Quantitative Benefits (continued)
Graph2: The number of WRs with a rating of Below 03 decreased as shown from 5 in July 2006
to 2 in August 2006.
*Rating explained - 01 Needs Improvement, 02 Below Expectation, 03 Meets Expectation, 04
Exceeds Expectation, 05 Excellent
16. 16
Conclusion
• The coupling of RCA with Marguglio’s Human Error Causal Factor
Taxonomy helped
o to uncover human performance issues and communication errors
o formulate strategies to reduce them
• Team involvement in the RCA process led to
o a greater involvement
o participation and ownership of quality enhancement.