Application of Lean Six Sigma in IT Industry

Posted in Operations & IT Articles, Total Reads: 1179 , Published on 20 February 2015
Advertisements

If you do just Six Sigma, you're not going to maximize the potential of your organization. You have to do both," - Mike Carnell, President of Six Sigma Applications

As Lean matured and Six Sigma started to receive acceptance across organizations other than Motorola and Toyota, they both became successful and competing methodologies for business improvement. Today, several companies rest upon either Lean or Six Sigma in order to streamline and improve their processes. Smart companies however are finding ways to integrate both methodologies, under the umbrella of Lean Six Sigma (LSS). This paper deals with a case where a large software service provider firm LPS implemented lean Six Sigma methodology to improve upon the existing process outsourced by a large international client X1RT Ltd.




1. Introduction

X1RT Ltd. is an IT organization and is a part of a large conglomerate X1R Group in USA. X1R group has its operations spread across various sectors like manufacturing, real estate, electronics etc. X1RT has a large part of its data processing operations being outsourced to LPS, a large Indian IT firm. Data analysts at LPS found out that a huge amount of time and effort was being wasted in verification of data. To resolve this problem, the analysts tried to implement lean six sigma technique to automate the manual process. The project was named ‘Bot’.


2. Background

Initially a code was created to perform Extraction, Transformation and Loading (ETL) of data from source to server. The code was first run in the developer environment and thereafter sent to Quality Analysis environment in the form of ETL object. An ETL object consisted of different stored procedures used to perform several tasks for example aligning variable names as per convention, session connection, lookup, filter, storing log files etc. After being passed through the Quality Analysis phase, the object was moved to the final Production phase to be implemented. Validation of the code was to be performed at all stages manually. Errors found at any stage were relegated to the original programmer and the cycle was repeated. This led to loss of man-hours every time before and after the code migration was done.


3. Project Bot

Project Bot was aimed at improving the existing process by automating the ETL objects’ standard validation process. The goal was to minimize effort and time. After initial analysis, it was found that there was a clear opportunity of saving approximately 7-8 hrs on weekly basis (i.e. 384 hrs/annually) during the entire code migration process. This could lead to $ 17000 potential saving for the whole financial year. This could be achieved through avoiding the error prone manual process, decrease in dependence on personnel, and curtailing man-hours spent.


A special team of four members was formed at LPS to carry out Project Bot. Two of the team members were Informatica Programmers. A Six Sigma Black Belt Certified LPS associate was roped in from a project other than X1RT. His job was to mentor and supervise the work of the programmers. Apart from these three, there was a final approver at the helm of the project, who was a Six Sigma Master Black Belt Certified associate.


Over the course of next four weeks, the programmers and their mentor engaged in exhaustive brainstorming. They discussed possible ways to tackle the problem while minimizing variation. After much deliberation, the group came down to two possible scenarios. They could either bring about the required change by modifying the existing process or could create a new process altogether that could be integrated with the existing system.


4. DMADV Implementation

DMAIC and DMADV are the two most widely used Six Sigma methodologies. DMAIC stands for Define, Measure, Analyze, Improve, and Control. DMADV stands for Define, Measure, Analyze, Design, and Verify. In both the methodologies, standard statistical tools are used to find solutions for improving process quality in an organization. However, DMAIC typically defines a business process and how applicable it is i.e. current performance of a process. Whereas DMADV defines the needs of the customer as they relate to a service or product i.e. measures customer specifications and needs. DMAIC is applied on a product or process that already exists but is no longer meeting customer need and needs modification. DMADV, on the other hand, is implemented to design a new product or process altogether.


As Project Bot involved creation of a new automated program to be integrated into the current process, DMADV methodology was best suited to the project.


Define Phase

This phase deals with the current status of the project before implementation of the new automated program. The problem to be addressed is defined in this phase. Five factors- Supplier, Input, Process, Output, Customer- also known as SIPOC, are understood in detail. The status of the existing process was factored according to the five parameters and can be explained using the following diagram:


Figure 4.1: Define Phase


Suppliers were the several data sources from where the data required in the existing process was extracted. The data sources were flat files, COBOL format data files, Bloomberg data files, Wall Street Suite format files etc. So, it can be determined that the data sources lacked any standard format. This was one of the limitations faced by the Project Bot team.

Inputs were the ETL objects that needed to be retrieved for validation. Following were some of the properties that were validated manually before the inception of Project Bot-naming convention, references, and query overwrite logic etc.

The existing process was to be overhauled whereby the manual process of validating was to be replaced by automated program. As a result, the end user would get a log of errors detailing violations of standard occurred during the process run. This constitutes the expected output. The final error log was to be used by the Informatica programmers to rectify the violations of standard found.

The group performed feasibility analysis keeping two factors in mind. The first of the two was CTQ (Critical to Quality) Factor. In order to align the final result with the quality standard of the client, certain CTQ factors were considered. The second factor was the required output and skill capacity at LPS’s end. It was also known as ‘Voice of Customer’ where the customer or stakeholder was the programmer. The current skillset available with the team was to be checked for its alignment with the final output expected out of the automation project. The project in question needed knowledge of Informatica standard codes and knowledge of SQL to access the database and create error report log. The two programmers in the team had the required skills to start off the project with a mentor to guide them. Keeping the expectations and limitations in mind, the group reached to the conclusion of going ahead with creation of a new automation process and integrating it with the current system. The six sigma methodology that was to be used here was DMADV.


Measure Phase

The goal of the Measure phase is to identify and define appropriate metrics and a measurement system that will capture the performance of CTQs. The metrics defined should be without any measurement ambiguity. The metrics should be SMART- Specific, Measurable, Achievable Targets, Realistic Goals and Time Bound. In this phase, two metrics were determined. The first metric was ‘what to measure’. The metric defined for the same was the total time taken, in hours, to identify the occurrence of violation of standard in the existing process. The second metric was ‘how to measure’ which dealt with information collected from the search activity done by business users.

An X-Y Matrix was used to relate the metrics determined (on Y axis) with the CTQ factor of effort spent (on X axis). A strong relationship between the two was deduced using the matrix.

Furthermore, TAKT time was calculated. In Lean, takt time is the rate at which a finished product needs to be completed in order to meet customer demand. The TAKT time involved in the project was calculated by dividing the time in hand by the number of transactions demanded by the client within the same time frame. Based on data collected the mean cycle of current process was determined. In order to determine the Upper Control Limit (UCL) in the Process Control Indices, improvement cycle was needed to be calculated. This is calculated by subtracting TAKT time from Cycle Time.



Figure 4.2: X-Y Matrix


Analyze Phase

In this phase, the following actions were to be taken: developing design alternatives, identifying the optimal combination of requirements to achieve value within constraints, developing conceptual designs, evaluating and selecting the best components in order to finalize and act upon the best possible design.



Fig 4.3: Functional Analysis


The goal of the Project Bot team was to minimize variations while automating the existing process. In the Analyze phase, the team mulled over the possible variations that could occur during the project flow. Ishikawa diagram (or Fishbone Diagram) was used to determine possible variations and sources of those variations.


Fig 4.4: Ishikawa Diagram



Design Phase

In this stage, Project Bot team designed a solution for the implementation of automation process. Initially a process map was designed as shown in figure and then automation code was developed to simplify the process by elimination of non-value added activities in the existing scenario.


After implementing the newly designed process, the following results were achieved:

1. The process of ETL object standard validation was improved efficiently as error prone manual process was avoided.

2. The efficiency of entire process was improved since this process was automated and personnel dependence was reduced.

3. It could be used as a document to refer in case someone needed to find the invalid objects of a particular ETL folder for further analysis.

4. Issues related to identify the invalid ETL objects were reduced to a great extent. This led to customer satisfaction and cost saving due to bug fixations.


Fig 4.5 Design Phase


After designing a new system, it needed to be tested for its robustness. So test cases for Unit Test Plan and System Test Plan were designed and tested for different test conditions. Results were as per expectations and testing was done successfully.


Verify Phase

In this phase, Project Bot Team validated whether the newly designed process addressed the issues of customer and business in a better manner. The team ran simulations after the new automated process was deployed. A methodology of process capability index was used to analyze the conduct of the new process attributes to designing details. A capability index refers to the voice of customer (determination cutoff points) to the voice of procedure. It was advantageous on the grounds that it lessened complex data about the procedure to a solitary number. In order to calculate process capability indices, the Project Bot team followed area under the curve method. The calculations are shown in Figure X (Appendix).

After performing the necessary calculations, the team developed Process Management Chart to Define and document process, supervised on a progressing premise to guarantee that measures provided feedback on the value stream/capacity of a procedure.

Failure Mode and Effects Analysis (FMEA) is a valuable system for counteracting future issues and hedging the risk to the solution. It was used to recognize and evaluate defects and errors that could threaten the quality, safety and reliability of the process.


5. Results

With the help of powerful Lean Six Sigma tools, the team for Project Bot was able to set up a systematic and guided framework to discover the pain areas in the existing systems and provide a sustainable feasible solution. The steps taken by the team are described in detail in every phase (DMADV) of the case. The remarkable achievements of this project were as follows:

1. All the operations related to ETL object validation were streamlined.

2. The existing manual process of ETL object validation which was quite susceptible to errors was avoided.

3. Personnel dependence was reduced to greater extent and approximately 384 hours were saved annually.

4. The implementation of Project Bot led to a primary cost saving of $25000 per annum for X1RT Ltd.


Also the implementation of Lean Six Sigma framework led to improved efficiency as shown in the following table.

Attributes

Current

After improvement

Average Time taken (in hrs) to complete the ETL Object standard validation

35

3

Short Term Sigma

2.28

3.05

Long Term Sigma

0.78

1.55


6. Conclusion

The global market is increasingly becoming more quality oriented. In order to be more competitive in such a cut throat competition, organizations must develop and adapt to quality control measures to render defect free deliveries in the least possible time. The above case depicts how Lean methodology and Six Sigma techniques which were initiated in manufacturing processes evolved over time period and can be used in multiple sectors such as IT Industry. Taiichi Ohno developed Toyota Production System (TPS), also known as Lean Manufacturing, which is focused on removing wastes and non-value adding activities in the process to streamline it. Six Sigma methodology is about controlling and reducing variations in the process which might lead to defects and customer dissatisfaction. A combination of both the techniques results in a powerful framework aimed at doing things right (Defect Free) and doing things quickly (Speed). This methodology demonstrated benefits (like streamlining processes, reducing personnel dependency, increased efficiency and profitability etc.) to preponderate the costs associated with its development and implementation.


This article has been authored by Kirti Upreti & Ashutosh Gavali from IIM Calcutta


Appendix

Before Implementation



References

I. George, Michael L. Lean Six Sigma for Services. New York: McGraw-Hill, 2003. Print.

II. Laureani, Alessandro, and Yair Holtzman, eds. Advanced Topics in Applied Operations Management. Shanghai: inTech China, 2012. Print. 

III. “Taiichi Ohno.” Wikipedia: The Free Encyclopedia. Wikimedia Foundation, n.d. Web. 22 Aug. 2014.

IV. Thatte, Dushyant. “DMADV Case Study: Performance Management System Redesign.” iSixSigma. N.p. , 23 Jan. 2012. Web. 20 Aug. 2014.

V. Assarlind, M., Gremyr, I. and Backman, K.” Multi-faceted views on a Lean Six Sigma application.” International Journal of Quality & Reliability Management 29.1 (2012): 21-30. Print.

VI. Hoerl, R. W. and Gardner, M. M. “Lean Six Sigma, creativity, and innovation.” International Journal of Lean Six Sigma 1.1 (2010): 30-38. Print.

VII. Nandakumar, P. M., Anitha H.S., and Bhanu Pratap Singh. “Six Sigma Methodology Utilization in Medical Transcription-A DMAIC Process to Identify Six Sigma Projects.” International Journal of Scientific Engineering and Technology 3.5 (2014): 574-8. Print.



Advertisements


If you are interested in writing articles for us, Submit Here