advanced innovation group - improve average handling time
DESCRIPTION
There have been various types of project posted on Slideshare by Advance Innovation Group to help the audience understand the way of initiating and closing a Six Sigma project. This presentation posted on Slideshare by Advance Innovation Group is also intended to solve the same purpose. AHT for ABC is too high AHT for Banglore Location is too high AHT for New Agents is too high Our project then becomes in measurable terms: Project Level Improving the AHT for Mortgage Process from 2.69 min. to 2min.The process has been budgeted to achieve AHT target of 2 min for Supernova process. Historically the team has not been able to achieve that. Through this project we can explore opportunities in the process which if corrected may result in achieving the desired results. The goal of the project was to Improve the AHT from 2.63 to 2 min for Supernova process to meet the client SLA and to avoid any penalty up to 20% for the organization. Additionally, it is advisable that you also visit and subscribe Advance Innovation Group Blog (http://advanceinnovationgroup.com/blog) for more Lean Six Sigma Projects, Case Studies on Lean Six Sigma, Lean Six Sigma Videos, Lean Six Sigma Discussions, Lean Six Sigma Jobs etcTRANSCRIPT
Advance Innovation Group
Project: To Improve AHT
6 Project
The Six Sigma Process
The DMAIC Steps
1. Define
2. Measure
3. Analyze
4. Improve
5. Control
3
DEFINE
Why?
Why?Why?
Level A
Level B
Level C
Level D
???
Why?
AHT too high AHT for ABC is
too highAHT for Banglore Location is too high
AHT for New Agents is too
high
Our project then becomes in measurable terms: Improving the AHT for Mortgage Process from 2.69 min. to 2min.
Project Level
Project Scoping
Voice of the Customer-VOC
Customer Comments Critical to Quality-CTQ’s
Dipti Nayak – (Champion)Achieve the AHT target to 2 min for the process by end of
Achieve the AHT up to 2 min , avoid the any revenue loss
Sunil Kumar-(Stakeholder)
The process has been budgeted to achieve AHT target of 2 min for Supernova process. Historically the team has not been able to achieve that. Through this project we can explore opportunities in the process which if corrected may result in achieving the desired results.
Achieve the desired target of 2 min
Smita-(Sr. Manager-Operations)
Improvement in AHT for Supernova process by removing NVA’s from the process and improving the bottom performer.
Improve AHT, meet the target
Project Charter
Business case:We provide technical support to one of our clients XYZ for GPRS. As per XYZ SLA, our focus is to wrap up the call without putting on hold. Any call received in call centre should be resolved in the short span of time with accurate and complete information. However this has been observed that during the call, time taken to resolve the query by a Agent is more than what client has asked for .AHT as a critical metrics has an impact on forecasting of calls, and also lead to client dis-satisfaction for not meeting the target. Improvement in AHT is an important metrics to forecast the business, meet the client targets avoid any revenue loss for the organization which can lead to a revenue loss up to 20%
Team:Project Leader BB:-Ravi RanjanProject Mentor:- Mithlesh NautiyalProcess Owner:- PushpinderProject Champion:- Dipti NayakSponsor:- Neha SharmaTeam Members:- Naveen, Jyotirmay
Problem Statement: For not meeting the defined target as per SLA there is penalty of 20% and additional 1% for incorrect solution, which is a revenue loss to the organization. Over the last 6 months the average AHT of Supernova is 2.63 against the target of 2.
Goal Statement: Improve the AHT from 2.63 to 2 min for Supernova process to meet the client SLA and to avoid any penalty up to 20% for the organization by 30th Nov’2010
In Scope : All Calls received by the agents .Out Scope : Any Calls not part of the Avaya. Milestones Start Date End date
D 7th May ’13 18th May’13
M 14th June’13 6th July’13
A 20th July’13 3rdAug’13
I 14th Sep’13 28th Sep’13
C 19th Oct’13 23rd Nov’13
ARMI
Key Stakeholders ARMI Worksheet
Define Measure Analyze Improve Control
Sponsor I I I I I
Champion I I I I I
Black Belt Leader I&A I&A I&A I&A I&A
Sr. Manager A&I A&I A&I A&I A&I
Black Belt I&R I&R I&R I&R I&R
Team Members I&M I&M I&M I&M I&M
A – Approval of team decisions I.e., sponsor, business leader, MBB.R – Resource to the team, one whose expertise, skills, may be needed on an ad-hoc basis.M – Member of team – whose expertise will be needed on a regular basis.I – Interested party, one who will need to be kept informed on direction, findings.
Communication Plan
Information Or Activity Target Audience Information Channel Who WhenProject Status Leadership E-mails Ravi Ranjan BI-Weekly
Tollgate Review BB,LBB,MBB & Champion
E-mails or Meetings Ravi Ranjan As per Project Plan
Project Deliverables or Activities Members Emails, Meetings Ravi Ranjan Weekly
CTQ Tree
Improvement in Handling TimeImprovement in Handling Time
CTQs
Handling Time per callHandling Time per call
2 Minutes per call2 Minutes per call
Any day when handling time is more than 2 minutes per call is defective day
Any day when handling time is more than 2 minutes per call is defective day
USL 2 Minutes per callUSL 2 Minutes per call
SIPOC
S I P O C
Call Routing to
Avaya
Agent pick the
call,solve the query
of customer
Agent finishes the
task and pick up the
next call
Agent tag the call in
system
Call Routing
Monitoring
Client And End
Customer
Client And End
Customer
Avaya and other
reporting tools
Various Audits
reports from team
leaders and Quality
Checkers
Client Call
Audits(Master
Sheet) Sheet has
been updated
from all the
auditions
10
MEASURE
Data Collection Plan
KPI Operational Definition Defect Def Performance Std
Specification Limit
Opportunity
LSL USL
Handling Time The difference between end time and start time of a call
Handling time greater than 2 minutes will be
defect
Handling Time Should be 2 minutes N/A 2 min Daily
KPI Data Type Data Items Needed Formula to be used Unit
Plan to collect Data Plan to sample
What Database or Container will be used to record
this data?
Is this an existing
database or new?
If new, When will the
database be ready for use?
When is the planned start date for data
collection?
Handling Time Continuous ATT + Hold=AHT (Call End time-Call start time) Minutes Excel Sheet Existing N/A 15th April 2012 January 2012
to march 2012
Validate Measurement System Gage R&R Annova Method
1. % Contribution of Gage R&R is less than part-part variation
2. % Tolerance Gage R&R against total variation is less than10 %
3. Number of Distinct Categories = 841
Gauge is successful, hence the data is good for further analysis
%Contribution Source VarComp (of VarComp)Total Gage R&R 0.0000 0.00 Repeatability 0.0000 0.00 Reproducibility 0.0000 0.00 Operator 0.0000 0.00 Part-To-Part 17.1968 100.00 Total Variation 17.1969 100.00
Study Var %Study Var %ToleranceSource StdDev (SD) (5.15 * SD) (%SV) (SV/Toler)Total Gage R&R 0.00695 0.0358 0.17 1.79 Repeatability 0.00687 0.0354 0.17 1.77 Reproducibility 0.00107 0.0055 0.03 0.27 Operator 0.00107 0.0055 0.03 0.27Part-To-Part 4.14690 21.3566 100.00 1067.83Total Variation 4.14691 21.3566 100.00 1067.83
Number of Distinct Categories = 841
ANALYSE
Organize Potential Cause – CE or Fish Bone Diagram
Handling
Time
ManpowerMaterials
MeasurementMethod
Training Score
Routing
CallHandling
Long Transactions
Marketing
Tenure
Process Knowledge
Age
Gender
Mode Of Communication
Trainers
Assistant Manager
Shift
Qualification
Marital Status
Typing Speed
Normality Test
Normality: P value = 0.005
Shape: Non-Normal
Measure of central tendency :data is non-normal measure of central tendency will be Median = 2.693
InterpretationThe p-value for the normality test (bottom right) is 0.005, which is
less than 0.05, indicating that data doesnot follow normal
distrubution.
Randomness And Stability Test
16
The two tests for non-random behavior detect trends, oscillation, mixtures, and clustering in your data. Such patterns suggest that the variation observed is due to "special causes" - causes arising from outside the system that can be corrected.Clusters may indicate variation due to special causes, Clusters are groups of points in one area of the chart.Here, the p-value for clustering is less than .05, so you might reject the null hypothesis of randomness - suggesting clustering.
spec
ial c
ause
s
Graphical Summary
Normality: P value: 0.005
Shape: Non-Normal
Measure of central tendency :data is non-normal - measure of central tendency will be median = 2.7159
Interpretation
The Mean of the data is 2.69 and the Median is 2.66 against the Target AHT of 2, which means both are on higher side from the targeted value.
Target
Process Capability
The capability of the process is very poor as represented by the Z value
Control Chart Of Present AHT
InterpretationThese test results indicate that the process average is unstable and the process is out of control, possibly due to the presence of special causes .
Test Results for I Chart of AHT TEST 1. One point more than 3.00 standard deviations from center line.Test Failed at points: 32, 49, 94, 160, 162, 187, 201, 229, 293, 359, 417,
Test Results for MR Chart of AHT TEST 1. One point more than 3.00 standard deviations from center line.Test Failed at points: 32, 49, 50, 94, 96, 188, 219, 229, 293, 359, 417, 418
Special Causes
1101009080706050403020
100
50
0
C1
Fre
que
ncy
Comparison of Distributions.
Sketch in the Means and Medians on each Distribution.Sketch in the Means and Medians on each Distribution.
Negative Skew Positive Skew
Symmetric Distribution
80706050403020100
300
200
100
0
C3
Fre
que
ncy
Comparison of Distributions.
Tail
13012011010090807060
300
200
100
0
C2
Fre
que
ncy
Comparison of Distributions.
Tail
Different Distributions
105
15
10
5
0
Index
Hrs
Var
Time Series Plot
PaperMiningUtilWater
15
10
5
0
Cus tom er
HrsV
ar
Box Plot
JulJunMay
15
10
5
0
Date
Hrs
Var
Scatter Plot
SalesmanCustomerDate
13.0
10.5
8.0
5.5
3.0
HrsVar
Main Effects Plot - Means for HrsVar
ANOVA (Analysis of Variance)
12013229343112.313.530.044.2
100.0 87.7 74.2 44.2
1000
900
800
700
600
500
400
300
200
100
0
100
80
60
40
20
0
Defect
CountPercentCum %
Per
cent
Cou
nt
Pareto Chart for : Defects
Pareto Chart
Tools Used……..
Histogram Of AHT
InterpretationThis histogram shows the distribution is centered around 2 to 3 AHT and its shape is symmetrical. Values range from 1 to 7. Further, comparing the histogram to the fitted normal distribution suggests that it is a good fit, which suggests that these data satisfy the assumption of normality for tests that require it. A normality test would more rigorously test this hypothesis.
The Histogram with Fit Shown above shows that data is not Normally Distributed , data is in Symmetric Distribution. For further Proof I check with Skew and Kurtosis Quotient.
Data Collection - X’s
S.No. Potential X's Description Data TypeTests To be Performed
1 Process Knowledge Subject Knowledge of the Agent gauged through Knowledge Test. Discrete Mann Whitney
2 Age Agent Age as per records available with HR. Continuous Regression Test
3 Gender The Male/Female Gender of the individual Agent logged in to take calls. Discrete Mann Whitney
4 Typing Speed Typing Speed of the Agent gauged through Typing Tutor test. Continuous Regression Test
5Mode Of Communication Language Used by the Agent to Communicate. Discrete Moods Median
6 Trainers The Agent were trained by three different trainers. Discrete Moods Median
7 Assistant ManagerThe Process Owner, who is responsible for managing the performance of a Team. Discrete Moods Median
8 Training Score Training Scores of the Agent,taken from the process training records. Discrete Mann Whitney
9 Tenure Agent tenure in the current process , as per records available with HR. Continuous Regression Test
10 Qualification Education Qualification of the Agent is available , as per records with HR. Discrete Moods Median
11 Marital Status Marital Status available as per HR records. Discrete Mann Whitney
12 Shift The Agent shift typing available as per Assistant Manager. Discrete Moods Median
AHT Versus Process Knowledge- Mann Whitney Test
Mann-Whitney Test and CI: PK(Process Knowledge) Above 70, PK Below 70
N MedianPK Above 70 688 2.5484PK Below 70 639 2.6294
Point estimate for ETA1-ETA2 is -0.040995.0 Percent CI for ETA1-ETA2 is (-0.0929,0.0114)W = 446098.0
Test of ETA1 = ETA2 vs ETA1 not = ETA2 is significant at 0.1238The test is significant at 0.1238 (adjusted for ties)
Since P -Value is greater than 0.05, Process Knowledge is not impacting AHT
AHT Versus Age-Regression Analysis Test
Regression Analysis: AHT versus Age
The regression equation isAHT = 2.864 - 0.007493 AgeS = 0.581758 R-Sq = 0.8% R-Sq(adj) = 0.7%Analysis of VarianceSource DF SS MS F PRegression 1 3.666 3.66581 10.83 0.001Error 1325 448.437 0.33844Total 1326 452.102
Since P -Value is less than 0.05, Age is impacting AHT
AHT Versus Gender(M/F)-Mann Whitney Test
Mann-Whitney Test and CI: Male, Female (Gender)
N MedianMale 514 2.6954Female 813 2.5193
Point estimate for ETA1-ETA2 is 0.137295.0 Percent CI for ETA1-ETA2 is (0.0856,0.1887)W = 376401.5
Test of ETA1 = ETA2 vs ETA1 not = ETA2 is significant at 0.0000The test is significant at 0.0000 (adjusted for ties)
Since P -Value is less than 0.05, Gender is impacting AHT
This shows that Out of Male-2.69 and Female-2.51 ,Females AHT are better than Males.
AHT Versus Typing Speed-Regression Analysis Test
Regression Analysis: AHT versus Typing Speed
The regression equation isAHT = 2.654 - 0.000721 Typing Speed
S = 0.584106 R-Sq = 0.0% R-Sq(adj) = 0.0%Analysis of VarianceSource DF SS MS F PRegression 1 0.039 0.039033 0.11 0.735Error 1325 452.063 0.341180Total 1326 452.102
6050403020
8
7
6
5
4
3
2
1
Typing Speed
AHT
S 0.584106R-Sq 0.0%R-Sq(adj) 0.0%
Fitted Line PlotAHT = 2.654 - 0.000721 Typing Speed
Since P -Value is greater than 0.05, Typing Speed is not impacting AHT
AHT Versus Mode Of communication-Mood Median Test
Mood Median Test: AHT versus Mode of Communication
Mood median test for AHTChi-Square = 16.32 DF = 1 P = 0.000
Mode of Individual 95.0% CIsCommunication N<= N> Median Q3-Q1 ------+---------+---------+---------+English 284 357 2.654 0.623 (----*-------)Hindi 379 305 2.527 0.608 (------*-------) ------+---------+---------+---------+ 2.520 2.580 2.640 2.700Overall median = 2.595
A 95.0% CI for median(English) - median(Hindi): (0.063,0.192)
Since P- Value is less than 0.05, Mode Of Communication is impacting AHT
AHT Versus Trainer-Mood Median Test
Mood Median Test: AHT versus Trainer Mood median test for AHTChi-Square = 15.12 DF = 2 P = 0.001 Individual 95.0% CIsTrainer N<= N> Median Q3-Q1 -------+---------+---------+---------Kishore 199 245 2.648 0.572 (----*------)Ruhani 252 187 2.482 0.661 (------*---------)Sudesh 213 231 2.616 0.625 (------*------) -------+---------+---------+--------- 2.480 2.560 2.640Overall median = 2.596
SudeshRuhaniKishore
8
7
6
5
4
3
2
1
Trainer
AHT
Boxplot of AHTSince P- Value is less than 0.05, Trainer is impacting AHT
This shows that out of three trainers- (Kishore-2.64,Ruhani-2.48,Sudesh-2.61), Ruhani Performing better than others. With a Mean of 2.48.
AHT Versus Assistant Manager-Mood Median Test
Mood Median Test: AHT versus A.M. (Assistant Manager)Mood median test for AHTChi-Square = 15.12 DF = 2 P = 0.001 Individual 95.0% CIsA.M. N<= N> Median Q3-Q1 -------+---------+---------+---------Ajay 252 187 2.482 0.661 (------*---------)Monica 199 245 2.648 0.572 (----*------)Pardeep 213 231 2.616 0.625 (------*------) -------+---------+---------+--------- 2.480 2.560 2.640Overall median = 2.596
PardeepMonicaAjay
8
7
6
5
4
3
2
1
A.M.
AHT
Boxplot of AHT Since P-Value is less than 0.05, Assistant Manager is impacting AHT
This shows that out of three Assistant Manager- (Ajay-2.48,Monica-2.64,Pardeep-2.61), Ajay Performing better than others. With a mean of 2.48.
AHT Versus Training Score-Mann Whitney Test
Mann-Whitney Test and CI: (Training Score)TS Above 65, TS below 65
N MedianTS Above 65 870 2.6038TS below 65 457 2.5611
Point estimate for ETA1-ETA2 is 0.025095.0 Percent CI for ETA1-ETA2 is (-0.0295,0.0782)W = 583731.0
Test of ETA1 = ETA2 vs ETA1 not = ETA2 is significant at 0.3617The test is significant at 0.3617 (adjusted for ties)
Since P- Value is greater than 0.05, Training Score is not impacting AHT
AHT Versus Tenure-Regression Analysis Test
Regression Analysis: AHT versus Tenure The regression equation isAHT = 2.572 + 0.01865 TenureS = 0.583594 R-Sq = 0.2% R-Sq(adj) = 0.1%Analysis of VarianceSource DF SS MS F PRegression 1 0.831 0.831168 2.44 0.118Error 1325 451.271 0.340582Total 1326 452.102
54321
8
7
6
5
4
3
2
1
Tenure
AHT
S 0.583594R-Sq 0.2%R-Sq(adj) 0.1%
Fitted Line PlotAHT = 2.572 + 0.01865 Tenure
Since P- Value is greater than 0.05, Tenure is not impacting AHT
AHT Versus Qualification-Mood Median Test
Mood Median Test: AHT versus Qualification Mood median test for AHTChi-Square = 99.24 DF = 10 P = 0.000 Individual 95.0% CIsQualification N<= N> Median Q3-Q1 ----+---------+---------+---------+--B.A 157 116 2.49 0.55 (*-)B.B.A 32 56 2.75 0.53 (--*-)B.C.A 61 58 2.57 0.60 (-*-)B.Com 174 107 2.42 0.72 (-*-)B.Sc 106 76 2.48 0.62 (--*-)B.Tech 30 55 2.71 0.48 (--*--)M.A 21 18 2.50 0.83 (-----*------)M.C.A 61 84 2.71 0.63 (--*-)M.Com 2 13 2.92 0.48 (----*----)M.Sc 18 32 2.75 0.52 (---*-)MBA 2 48 2.98 1.53 (*-----------------) ----+---------+---------+---------+-- 2.40 2.80 3.20 3.60Overall median = 2.60
Since P- Value is less than 0.05, Qualification is impacting AHT
AHT Versus Marital Status(M/U)-Mann Whitney Test
Mann-Whitney Test and CI: Married, Unmarried (Marital Status)
N MedianMarried 788 2.5477Unmarried 539 2.6750
Point estimate for ETA1-ETA2 is -0.093395.0 Percent CI for ETA1-ETA2 is (-0.1470,-0.0396)W = 499918.5
Test of ETA1 = ETA2 vs ETA1 not = ETA2 is significant at 0.0007The test is significant at 0.0007 (adjusted for ties)
Since P- Value is less than 0.05, Marital Status is impacting AHT
This shows that out of Married-2.54, and Unamarried-2.67,Married Performing better. With a mean of 2.54.
AHT Versus Shift-Mood Median Test
Mood Median Test: AHT versus Shift Mood median test for AHTChi-Square = 0.64 DF = 2 P = 0.727 Individual 95.0% CIsShift N<= N> Median Q3-Q1 ---------+---------+---------+-------Afternoon 223 216 2.590 0.635 (----------------*--------)Evening 249 242 2.582 0.603 (-----------*-------------)Morning 191 204 2.619 0.635 (--------------*------) ---------+---------+---------+------- 2.550 2.600 2.650Overall median = 2.595
Since P -Value is greater than 0.05, Shift is not impacting AHT
This shows that out of three shifts-(Morning-2.61,Evening-2.58,Afternoon-2.59) Evening Shift Performing better. With a mean of 2.58
Summary of Statistical Analysis
Measure Type Data Type P-Value Impact
AHT Y Continuous NA NA
Process Knowledge x Discrete 0.1238 the test is not significant
Age x Continuous 0.001 the test is significant
Gender x Discrete 0 the test is significant
Typing Speed x Continuous 0.735 the test is not significant
Mode Of Communication x Discrete 0 the test is significant
Trainers x Discrete 0.001 the test is significant
Assistant Manager x Discrete 0.001 the test is significant
Training Score x Discrete 0.3617 the test is not significant
Tenure x Continuous 0.118 the test is not significant
Qualification x Discrete 0 the test is significant
Marital Status x Discrete 0.0007 the test is significant
Shift x Discrete 0.727 the test is not significant
IMPROVE
Summarizing Findings on Root Causes
Vital X's Mode of Analysis Findings
Age Regression Analysis Associates age is impacting the AHT.
Gender Mann Whitney Associates Gender is impacting the AHT.
Mode Of Communication Moods Median
Mode of communication is impacting associates AHT.
Trainers Moods Median Trainers training procedure is impacting the AHT.
Assistant Manager Moods Median Assistant manager performance is impacting AHT.
Qualification Moods Median Associates Qualification is impacting the AHT.
Marital Status Mann Whitney Associates Marital Status is impacting the AHT.
Implementing Solutions & Timelines
S.No. Root Cause Improvement Action ResponsibilityStart Date
End Date
Duration(Days)
1 AgeNew job description for New Hiring Batch-incude New Age criteria.
Process Manager/Recruitment Manager
14th Sep'13
24th Sep'13 10
2 GenderNew job description for New Hiring Batch-incude New Gender criteria.
Process Manager/Recruitment Manager
14th Sep'13
19th Sep'13 5
3Mode of Communication Communication Skills Training module included.
Process Trainer/HR Manager
14th Sep'13
19th Sep'13 5
Communication Training assesment included.Process Trainer/HR Manager
14th sep'13
16th Sep'13 2
Mode of communication has been added as one of the KPI. Top Management
14th sep'13
28th Sep'13 14
4 TrainersBest Practices sharing sessions conducted with the trainer-Ruhani.
Process Trainer/HR Manager
18th Sep'13
24th Sep'13 6
Trainer-Ruhani, to conduct sessions with all the associates to improve their AHT.
Process Trainer/HR Manager
20th Sep'13
22nd Sep'13 2
Trainer-Ruhani, to update training modules-to include different ways of improving AHT. Process Trainer
18th Sep'13
20nd Sep'13 2
5Assistant Manager
Best Practices sharing sessions conducted with the AM-Ajay Manager
21st Sep'13
28th Sep'13 7
AM-Ajay, to conduct sessions with all the associates to improve their AHT. Manager
14th Sep'13
19th Sep'13 5
AM-Ajay, to update training modules-to include different ways of improving AHT. Manager
14th Sep'13
19th Sep'13 5
6 Qualification New IJP has been added as per qualifications. Top Management18th Sep'13
28th Sep'13 10
7 Marital statusCounselling Session to conduct for Stressful unmarried Employees. Top Management
17th Sep'13
25th Sep'13 8
CONTROL
Post Project Control Chart
The Control Chart shows that the process Mean has shifted to 2.27, while special cause variations have been reduced.
Mean
Pre Vs post Project Graphical Summary
Pre Post
Interpretation In Pre the Mean of the data is 2.69 and the Median is 2.66 against the Target AHT of 2, which means both are on higher side from the targeted value.
In Post the Mean of the data is 2.27 and Median is 2.23 against the Target AHT of 2, but is reduced from Pre which means the current process AHT is improved .
Difference Between AHT & Improved AHT
Average of AHT Average of Improved AHT2.00
2.10
2.20
2.30
2.40
2.50
2.60
2.70
2.62984842501884
2.27592313489073
AHT In Minutes
Series1
Before
After
Project Sign-off
Reduced AHT from 2.63 to 2.27.
Improved Training Plans.
Improved Communication Skills.
New Hiring done with a Age limitation.
THANK YOU