performance evaluation through control charts

Upload: dilleh

Post on 02-Jun-2018

218 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/10/2019 Performance Evaluation Through Control Charts

    1/21

    1

  • 8/10/2019 Performance Evaluation Through Control Charts

    2/21

    2

    A ............................................................................................................................................. 3

    ............................................................................................................................. 4

    B ....................................................................................................................................... 4

    A ......................................................................................................................... 4

    B : .......................................................................................... 6

    A ................................................................................................................... 7

    ? ................................................................. 7

    ? ......... 7

    : ............................................................................................................................... 8

    A ..................................................................................10

    C C A ................................................................. 11

    C C : .......................................................................................................... 12

    ................................................................................................................................ 13

    C C ................................................................ 15

    C C : .......................................................................................................... 16

    ............................................................................................................................................. 19

    C : ........................................................ 19

    C : ..................................................... 19

    ................................................................................. 20

    ...................................................................................................................................... 21

  • 8/10/2019 Performance Evaluation Through Control Charts

    3/21

    3

    Organizations need to evaluate the performance of their employees at all levels to maintain a

    competitive workforce and provide their employees growth opportunities.

    Organizations further segregate the employees into different level of performance standards like

    excellent, average or poor performers.

    This article focus on use of control charts, a tool used for statistical process control, to create

    these performance levels objectively, with the help of data gathered out of a project for a specific

    duration.

    This article also elaborates how the methodology can be extended for different scenarios and

    organization requirements.

  • 8/10/2019 Performance Evaluation Through Control Charts

    4/21

    4

    The main objectives of this study are:

    To objectively evaluate the performance of a group of employees, relatively against

    each other using control charts.

    To conclude statistically, if a group the employees can be segregated in the groups

    like excellent, average and poor (or similar), based on their performances, using

    control charts.

    To review the reason of the employees lower than average and better than average

    performance using control charts

    [Mean: Simply stated, it is the average of all observations in a given sample. It is one of

    the measures of central tendency. Measures of central tendency provide an idea about the

    locations of the observations in the sample and the value about which they cluster.

    Variance: The variance measures the scatter of the observations from the mean, the

    larger the value, the greater the scatter. This is one of the measures of dispersion. In a given

    sample, a variance provides an idea on the variability, or scatter of observations around a

    given value, usually the mean.

    Standard Deviation: The standard deviation also measures the variability of the

    observations around the mean. It is the positive square root of variance. It has the same units

    as the observations in the sample and thus is easier to interpret.] REF 2

  • 8/10/2019 Performance Evaluation Through Control Charts

    5/21

    5

    Control Charts:[Control charts, also known as Shewhart charts (after Walter A.

    Shewhart) or process-behavior charts, are among one of the tools for Statistical process

    control used to make sure if a manufacturing or business process is in a state of statistical

    control.] REF 4

    [A control chart is a statistical (graphical) tool for monitoring the movement of a

    continuing process. The values of quality characteristic are plotted along the vertical axis and

    the horizontal axis represents the samples or subgroups from which this characteristic is

    derived.] REF 2

    [A control chart consists of:

    Points representing a statistic (e.g., a mean, range, proportion) of measurements of a

    quality characteristic in samples taken from the process at different times [the data]

    The mean of this statistic using all the samples is calculated (e.g., the mean of the

    means, mean of the ranges, mean of the proportions). A centre line is drawn at the

    value of the mean of the statistic

    Upper and lower control limits indicate the threshold at which the process output is

    considered statistically 'unlikely' and are drawn typically at 3 standard deviation from

    the centre line] REF4

    Sample Control Chart:

    0

    2

    4

    6

    8

    10

    12

    1 2 3 4 5 6

    C

    C (C)

    C (C)

    C

  • 8/10/2019 Performance Evaluation Through Control Charts

    6/21

    6

    [The values of the statistic are plotted on control chart, based on the assumption that the

    sample has an approximately normal distribution.

    The most common basis for analyzing whether a process is out of control is to check for any

    sample statistic falling outside the control limits. The reason of such outliers is analyzed and

    corrective actions are planned to bring back the process under control.

    The causes of variation can be subdivided in to two groups common causes and special

    causes.

    Variability caused by special cause is something not inbuilt in the process. That is, it is not a

    part of the process and does not affect all quality characteristic of all items.

    Variability due to common causes is something inbuilt in the process. It is also referred to as

    the accepted variation in a process.] REF 2

  • 8/10/2019 Performance Evaluation Through Control Charts

    7/21

    7

    [Performance management is the process through which the performance of the human

    resources in an organization is identified, measured, managed and developed. Basically

    organization tries to conclude how well employees perform and then to ultimately improve

    that performance level.

    Performance appraisal, on the other hand, is the continuous process of evaluating employees

    performance. Performance appraisals are reviews of employee performance for a specified

    duration. Appraisal is just one component of performance management.] REF [1]

    In software development based IT project, the following attributes can be considered for

    evaluating the performance of development team.

    1) Number of defects injected

    2) Schedule adherence

    3) Process adherence

    4) Total Efforts (may be in hours)

    5) Any customized parameters like points/score given against a set of tasks, in a project

    can also be used as a measurable parameter. I have used customized parameter in my

    study.

    Different roles, like testers, consultants, project managers etc. do have some different and

    some common parameters, that are considered for their performance evaluation. Hence

  • 8/10/2019 Performance Evaluation Through Control Charts

    8/21

  • 8/10/2019 Performance Evaluation Through Control Charts

    9/21

    9

    If we plot the control chart with the available data of employees as mentioned above, it will

    give us an idea about how scattered, the data is and if it is possible to form different groups of

    employees based on their performance. The main benefit is that the outcome of this exercise

    will not be based on a personal hunch but on the objectivity of data.

    The groups based on performance levels on the control chart are identified as the area falling

    between two straight lines parallel to X axis. One of these lines is generally drawn at mean of

    sample and rest may be drawn at multiples of standard deviation from the mean.

    Please refer to Figure 3 below. Digits 1, 2, 3, 4 and 5 depict the various regions/areas which

    may be considered as different groups (levels) of performance.

    Generally, organization utilizes this grouping based on performance (any derived score) for

    deciding the yearly increment or bonus which is different across the groups (may be none for

    lowest performance group) but same within the groups. The employees in different

    performance groups are given different ratings/rankings but same in one group.

    E.g. Employees in top group may be given a rating 1 (Excellent), the second group may get 2

    (good), 3 (average), 4 (below average-poor) and the bottom group may be 5 (unacceptable).

    Above image shows 5 different performance zones.

  • 8/10/2019 Performance Evaluation Through Control Charts

    10/21

  • 8/10/2019 Performance Evaluation Through Control Charts

    11/21

    11

    Developer Points Achieved MR (Moving Range)

    Rahul 140

    Nisha 65 75.00

    Vinod 65 0.00

    Mohit 120 55.00

    Madhu 55 65.00

    Raman 55 0.00

    Shivraj 90 35.00

    Puneet 130 40.00

    Jyoti 130 0.00Pooja 150 20.00

    Rohit 85 65.00

    Ismail 25 60.00

    Lalit 100 75.00

    Param 85 15.00

    Sachin 220 135.00

    Amitabh 100 120.00

    Reetika 85 15.00

    Yash 55 30.00

    Kapil 65 10.00

    Kunal 35 30.00

    Vikram 200 165.00

    Tej 70 130.00

    Anurag 95 25.00

    Pankaj 150 55.00

    Manu 25 125.00

    Abhishek 85 60.00

    Ajay 85 0.00

    Sanjay 6520.00

    Umesh 95 30.00

    Shruti 85 10.00

  • 8/10/2019 Performance Evaluation Through Control Charts

    12/21

    12

    The quality characteristic for above sample is the individual points achieved and

    plotted against the control limits.

    The variability of the process is estimated from Moving Range (MR) found as the

    positive difference of two successive observations.

    The quality characteristic/statistic that is plotted against the control limits in a control

    chart is Points achieved (first control chart in this article) and Defects Injected

    (second control chart in this article) as mentioned above.

    The control limits are set by finding the mean of sample and then by determining the

    3 sigma values which are then plotted as upper (+3 sigma) and lower (-3 sigma)

    limits.

    Centre line is the mean of the sample taken.

    UCL X bar + 3*MR bar/ d2 130.246

    LCL X bar - 3*MR bar/ d2 52.657

    MR bar MR/ no. of MR values 50.833

    Mean ( X bar)Points Achieved/ No.

    of Developers91.452

    We are inspecting each team member. Hence the sample size is 1.

    The score is a variable type of data and hence calls for variable control chat.

    The variable chart used with sample size one is X chart i.e. the values for each

    developer (sample) is plotted against the mean and 3 standard deviations values.

  • 8/10/2019 Performance Evaluation Through Control Charts

    13/21

    13

    The time sequence aspect of sampling is not valid in this case. That means, the samples are

    not drawn as in manufacturing industry, one by one. All the data for each developer was

    maintained and taken together from the records. Hence Moving Range (MR) chart may not

    be completely valid in this case and hence not created here.

  • 8/10/2019 Performance Evaluation Through Control Charts

    14/21

    14

    050

    1

    00

    1

    50

    2

    00

    2

    50

    A

    C

    C

  • 8/10/2019 Performance Evaluation Through Control Charts

    15/21

    15

    Developer Number of

    Reports(Sample

    Size)

    Defects NC per Sample

    Size(Defects/Display

    Converted)

    UCL LCL Final LCL Avg

    NC=TotalDefects

    (NC)/ TotalWeb

    Pages

    Tej3 1 0.33 1.02

    -0.13

    0 0.45

    Amit5 1 0.20 0.89

    0.00

    0

    Deepti5 1 0.20 0.89

    0.00

    0

    Preet3 1 0.33 1.02

    -

    0.130

    Nidhi2 4 2.00 1.86

    -0.97

    0

    Neha4 1 0.25 0.95

    -0.05

    0

    Tushar2 0 0.00 0.45

    0.45

    0

    Aman3 2 0.67 1.26

    -0.37

    0

    Naman6 1 0.17 0.85

    0.04

    0.04

    Mohit5 1 0.20 0.89

    0.00

    0

    Suresh4 1 0.25 0.95

    -0.05

    0

    Sanjay4 1 0.25 0.95

    -0.05

    0

    Ajay3 2 0.67 1.26

    -0.37

    0

    Kapil4 1 0.25 0.95

    -0.05

    0

    Puneet5 2 0.40 1.08

    -0.19

    0

    Sunil6 3 0.50 1.15

    -0.26

    0

    Nitin 2 3 1.50 1.67 -0.78 0

    Yash7 1 0.14 0.82

    0.07

    0.07

    Shruti3 2 0.67 1.26

    -0.37

    0

    UCJaiswal3 2 0.67 1.26

    -0.37

    0

    Vikas6 4 0.67 1.26

    -0.37

    0

    Pankaj2 1 0.50 1.15

    -0.26

    0

    Anurag4 4 1.00 1.45

    -

    0.550

    Saksham 4 1 0.25 0.95 - 0

  • 8/10/2019 Performance Evaluation Through Control Charts

    16/21

  • 8/10/2019 Performance Evaluation Through Control Charts

    17/21

  • 8/10/2019 Performance Evaluation Through Control Charts

    18/21

    18

    0

    .00

    0

    .50

    1

    .00

    1

    .50

    2

    .00

    2

    .50

    C

    (/C)

    C

    C

    AC=(C)/

  • 8/10/2019 Performance Evaluation Through Control Charts

    19/21

    19

    The performance of developers has been objectively divided in to 3 performance groups

    using control charts.

    In both charts, some points (developers) are falling outside (or on) the control limits derived

    based on sample data. The causes for the same were analyzed. In case of Points chart (X

    chart), points falling below LCL, while in case of Defect chart (U Chart) points falling above

    UCL, are of major concern (the below average performances in both cases).

    One of the reasons for fewer points is that some of the developers may have taken

    more leave than other and hence achieved fewer points.

    Another reason can be that some of them just joined the organization and hence

    needed more time to adjust to company culture and technology.

    Some can really be the low performers.

    The developers who work hard and are really good may come on the top. It should

    also be analyzed if these developers follow any best practices which can be shared

    with the team.

  • 8/10/2019 Performance Evaluation Through Control Charts

    20/21

    20

    The developers may be new and required further training to enhance his technical

    skills.

    Developer may not be using the standard checklists.

    Can multiple parameters/goals be accommodated in the same chart?

    o Yes. Assign different weights to different parameters and derive a weighted

    score.

    o X chart can be created as the weighted score will be a variable data.

    o Calculate the mean, define the UCL/LCL and plot the chart.

    Can UCL or LCL be set up by organization explicitly (MBO)? What if all developers

    in the sample fall below organization standards?

    o Yes. As described in OBJECTIVE section, organization can set UCL/LCL,

    preferably (not mandatory) at any multiple of Standard deviation from mean

    of data, e.g. 1 sigma, 2 sigma, 3 sigma etc and plot the graph. If the employees

    in sample fall below the organization set specification limits, they may all be

    grouped in to lower than average/poor performance level.

    Can we create more than 3 performance groups?

    o Yes. As described in OBJECTIVE section, each straight line drawn at a

    multiple of standard deviation, parallel to X axis creates two performance

    regions, one above and one below. Hence desired number of such lines should

    be plotted to get as many performance groups, as are needed.

    Can we apply this methodology to industry other than IT also?

    o Yes. Only the parameters or attributes for evaluating the performance should

    differ in that case.

  • 8/10/2019 Performance Evaluation Through Control Charts

    21/21

    21

    1. http://www.sagepub.com/upm-data/45674_8.pdf

    2. Fundamentals of Quality Control and Improvement (Second Edition) by AmitavaMitra

    3. http://en.wikipedia.org/wiki/Control_chart