measurement and metrics for test managers

52
© 2013 SQE Training V3.1 1 Introduction MEASUREMENT AND METRICS FOR TEST MANAGERS Hidden Slide for Logos 2 © 2013 SQE Training V3.1

Upload: techwellpresentations

Post on 06-May-2015

205 views

Category:

Technology


0 download

DESCRIPTION

To be most effective, test managers must develop and use metrics to help direct the testing effort and make informed recommendations about the software’s release readiness and associated risks. Because one important testing activity is to “measure” the quality of the software, test managers must measure the results of both the development and testing processes. Collecting, analyzing, and using metrics is complicated because many developers and testers are concerned that the metrics will be used against them. Join Rick Craig as he addresses common metrics—measures of product quality, defect removal efficiency, defect density, defect arrival rate, and testing status. Learn the guidelines for developing a test measurement program, rules of thumb for collecting data, and ways to avoid “metrics dysfunction.” Rick identifies several metrics paradigms—including Goal-Question-Metric—and discusses the pros and cons of each. Delegates are urged to bring their metrics problems and issues for use as discussion points.

TRANSCRIPT

Page 1: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 1

Introduction MEASUREMENT AND METRICS FOR TEST MANAGERS

Hidden Slide for Logos

2 © 2013 SQE Training V3.1

Page 2: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 2

Hidden Slide for SQE Training Copyright

3 © 2013 SQE Training V3.1

Administrivia

Course timing

Electronic devices

Smoking

Meals

Facilities

Breaks

4 © 2013 SQE Training V3.1

Page 3: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 3

Course Agenda

1.  Introduc+on  to  So.ware  Measurement  2.  Metrics—Rules  of  Thumb  3.  A  Tester’s  Dashboard  4.  Es+ma+on  (Op+onal)  

5 © 2013 SQE Training V3.1

1 INTRODUCTION TO SOFTWARE MEASUREMENT

Page 4: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 4

What is software measurement?

   

— Bill  Hetzel  

“It’s  easy  to  get  numbers,  what  is  hard  is  to  know  they  are  right  and  understand  what  they  mean”  

 

7 © 2013 SQE Training V3.1

What is software measurement?

“Quan+fied  observa+ons”  about  any  aspect  of  so.ware  (product,  process,  or  project)    

8 © 2013 SQE Training V3.1

Page 5: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 5

Lord Kelvin “To  measure  is  to  know”    “If  you  cannot  measure  it,  you  cannot  improve  it”    

 “The  more  you  understand  what  is  wrong  with  a  figure,  the  more  valuable  that  figure  becomes”  

9 © 2013 SQE Training V3.1

There Are Lots and Lots of Measures Primi+ve:    

–  Aspirins  consumed  this  week  –  Number  of  staff  assigned  to  project  A  –  Pages  of  requirements  specifica+ons  –  Hours  worked  to  accomplish  change  request  X  –  Number  of  opera+onal  failures  in  system  Y  this  year  –  Lines  of  code  in  program  Z  

 

Computed:    –  Defects  per  1000  lines  of  code  in  program  A  –  Produc+vity  in  func+on  points  delivered  by  person  B  

–  Quality  Score  for  project  C  –  Average  coffee  consump+on  per  line  of  code  –  Accuracy  of  hours  worked  per  week  is  ±  20%  

10 © 2013 SQE Training V3.1

Page 6: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 6

Common Metrics •  Test  defects  •  Defects  a.er  release  •  Open  problems  •  Open  issues  •  Schedule  performance  •  Process  compliance  (e.g.,  ISO)  •  Test  results  •  Reliability  •  Time  fixing  problems  •  Defects  from  fixes  •  Lines  of  code  •  Plan  and  schedule  changes  

11 © 2013 SQE Training V3.1

Uncommon Metrics

• Code  coverage  • Complexity  • Cost  of  rework  • Cost  of  quality  

Defect age

12 © 2013 SQE Training V3.1

Page 7: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 7

Basic Definitions The  four  Ms:  

•  Measure  •  Metric  •  Meter  •  Meta-­‐measure  

Primi+ve  (raw  data)  

Computed  (informa+on)  

13,  34,  17,  74  42,  34,  56,  77  94,  34,  45,  63  45,  67,  12,  31  61,  06,  91,  42  

13 © 2013 SQE Training V3.1

What Makes a Good Measure?

•  Simple  •  Objec+ve  •  Easily  collected  •  Robust  •  Valid  

14 © 2013 SQE Training V3.1

Page 8: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 8

What Can Measures Do for You? •  Facilitate  es+ma+on  •  Iden+fy  risky  areas  •  Measure  tes+ng  status  •  Measure/predict  product  quality  •  Measure  test  effec+veness  •  Iden+fy  training  opportuni+es  •  Iden+fy  process  improvement  opportuni+es  •  Provide  “meters”  to  flag  ac+ons  

15 © 2013 SQE Training V3.1

2 METRICS—RULES OF THUMB

Page 9: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 9

Metrics--Rules of Thumb

•  The  Human  Element  •  The  Basics  •  KISS  •  And  a  Myth  or  Two  

17 © 2013 SQE Training V3.1

The Human Element

• Without  buy-­‐in,  metrics  may  be  falsified  

• Without  buy-­‐in,  metrics  may  be  ignored  

Buy-­‐in  is  key  

18 © 2013 SQE Training V3.1

Page 10: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 10

Class Discussion

How  do  you  obtain  buy-­‐in?  

19 © 2013 SQE Training V3.1

Ways to Obtain Buy-in

•  Training  •  Metrics  •  Feedback  loops  •  Reviews  •  Par+cipa+on  

20 © 2013 SQE Training V3.1

Page 11: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 11

The Human Element

•  Measure  processes  and  products  instead  of  people  if  possible  

•  Beware  of  the  dark  side  of  the  Hawthorne  Effect  

21 © 2013 SQE Training V3.1

Two Sides of Measurement

…the  informa-on  may  be  used  against  me.  

…the  informa-on  will  help  me  

understand  what  is  going  on  and  do  

a  be8er  job.  

22 © 2013 SQE Training V3.1

Page 12: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 12

The Hawthorne Effect

Measuring  people  improves  their  produc+vity  

23 © 2013 SQE Training V3.1

The Human Element

Tailor  metrics  to  the  audience  

Users,  managers,  prac++oners  all  have  different  languages  

Set  the  appropriate  level  of  detail  

How  you  present  the  material  maqers  

24 © 2013 SQE Training V3.1

Page 13: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 13

Who is your audience?

Developers

Testers

Users

25 © 2013 SQE Training V3.1

% of Red Cars Soars

2008 2009 2010 25.1

25.5

26

25.4

25.2 25.3 25.4 25.5 25.6 25.7 25.8 25.9 26.0 26.1

26 © 2013 SQE Training V3.1

Page 14: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 14

% of Red Cars Soars?

2008 2009 2010

25

50

75

100

25.5 26 25.4

27 © 2013 SQE Training V3.1

The Human Factor

Training  is  required  

Metrics  are  not  second  nature  

Your  metrics  are  affected  by  how  they  are  collected  

Establish  range  of  expected  values  

Publish  historical  values  

28 © 2013 SQE Training V3.1

Page 15: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 15

The Basics

•  Use  a  metric  to  validate  a  metric  •  Use  meta-­‐measures  •  Use  meters  when  possible  •  Consistency  some+mes  trumps  accuracy  •  Subjec+ve  is  good;  objec+ve  is  beqer  

29 © 2013 SQE Training V3.1

KISS ― Keep It Simple Sir

•  More  is  not  always  beqer  •  All  metrics  are  not  forever      –  Consider  temporary  metrics      –  Consider  sampling  

•  Automate  collec+on  when  possible  

30 © 2013 SQE Training V3.1

Page 16: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 16

This Slide Is Hidden

31 © 2013 SQE Training V3.1

3 A TESTER’S DASHBOARD

Page 17: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 17

A Dashboard

33 © 2013 SQE Training V3.1

Establish a Dashboard

•  Easy  to  use/understand  at  a  glance  

*  Remember  you  need  at  least  two  metrics  per  “instrument”  

Quality  of  product  

Status  

Test  effec+veness  

Resources  

Issues  

34 © 2013 SQE Training V3.1

Page 18: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 18

Measures of Quality

•  It  is  difficult  to  develop  prac+cal  measures  of  quality  

•  The  cost  to  achieve  various  quality  levels  must  be  taken  into  account  

•  Many  quality  metrics  are  rela+vely  subjec+ve  

•  Quality  goals  will  be  affected  by  the  industry  and  corporate  culture  

35 © 2013 SQE Training V3.1

What Is Quality?

• Mee+ng  requirements  (stated  and/or  implied)  Quality  

36 © 2013 SQE Training V3.1

Page 19: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 19

Sample Quality Factors and Criteria •  Correctness  •  Reliability  •  Testability  •  Flexibility  •  Usability  •  Portability  •  Interoperability  •  Efficiency  •  Integrity  •  Maintainability  •  Revisability  •  Survivability  

Correctability Correctness Correctness Correctness Correctness Correctness Correctness

37 © 2013 SQE Training V3.1

Defect Density/Clustering

Module Name

# of Defects

per 1,000

Lines of Code

D B A C E F

38 © 2013 SQE Training V3.1

Page 20: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 20

Defect Density

Issues Coverage of tests

Weighting of defects

Weighting by relative risk

What to use as the denominator

39 © 2013 SQE Training V3.1

Effect of Complexity on Quality

Complexity

Pro

babi

lity

of P

ost-r

elea

se

Def

ect

40 © 2013 SQE Training V3.1

Page 21: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 21

Other Measures of Product Quality •  Customer  sa+sfac+on  •  Repeat  customers?  •  Referrals?  •  Calls  to  the  help  desk?  •  Timeliness?  •  Defect  age?  •  Complexity?  •  Rework?  •  Reliability?    

41 © 2013 SQE Training V3.1

Quality of Product

•  Record  any  current  measures  of  product  quality  that  you  are  using  here.  Give  them  a  grade  for  effec+veness  (A,  B,  C,  etc.)  

   •  Any  new  metrics  you  would  use?  

* Remember you need at least two metrics per instrument

42 © 2013 SQE Training V3.1

Page 22: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 22

Establish a Dashboard

•  Easy  to  use/understand  at  a  glance:  

 

*  Remember  you  need  at  least  two  metrics  per  instrument  

Quality  of  product  

Status  

Test  effec+veness  

Resources  

Issues  

43 © 2013 SQE Training V3.1

Status Reporting •  The Master Test Plan should

specify

– What to report – How often – To whom

Bugs

44 © 2013 SQE Training V3.1

Page 23: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 23

Common Test Status Metrics % of Test Cases Executed

Issues: •  Weighting of TC by coverage metrics •  Weighting of TC by risk •  Weighting of TC by execution effort •  Weighting of TC by time to execute What do you really want to know?

45 © 2013 SQE Training V3.1

Sample Test Status Report (raw data) Project: Online-Trade Date: 4/23/2009 Feature Total # % # % Tested Tests Complete Complete Success Success Open Acct 46 46 100 41 89 Sell Order 36 25 69 25 69 Buy Order 19 17 89 12 63 ….. ….. ….. ….. Totals 395 320 81 311 79

46 © 2013 SQE Training V3.1

Page 24: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 24

Open and Closed Over Time

2 4 6 8 10 12 14 16 18 20

Weeks

Def

ects

Incoming

Fixed

Released

0

10

20

30

40

24 22 20 18 16 14 12 10

8 6 4 2 0

Days

Cum

ulat

ive

Def

ects

Detected

Open

0 10 20 30 40

47 © 2013 SQE Training V3.1

When Is the Software “Good Enough”?

•  Test  exit  criteria  met  •  Return  On  Investment  (ROI)  not  sufficient  •  Defect  arrival  rate  •  Resources  exhausted  

–  Time  – Money  

•  Profiles  (based  on  failures  encountered  using  profiles  of  real  data)    

•  Project  cancelled!  

When  to  stop  tes-ng  

48 © 2013 SQE Training V3.1

Page 25: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 25

Software Psychology What  is  “good  enough”?  

Time

# of Bugs

49 © 2013 SQE Training V3.1

Economics of Test and Failure

Source: IBM Systems Sciences Institute

50 © 2013 SQE Training V3.1

Page 26: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 26

Stopping Criteria ― Revisited Abnormal

•  Resource exhaustion –  Schedule –  Budget –  System access –  Patience

•  Project redirection

Normal •  Test set exit criteria •  Remaining defects

estimation criteria –  Defect history of past software –  Defect history of current item –  Software complexity –  Combination of these

•  Diminishing return criteria

–  Cost to Detect Next Defect

•  Combined criteria “There is no single, valid, rational criterion for stopping. Furthermore, given any set of applicable criteria, how each is weighed depends very much on the product, the environment, the culture, and the attitude to risk.”

— Boris Beizer

51 © 2013 SQE Training V3.1

Test Summary Report •  Report identifier •  References

–  Test items (with revision #s) –  Environments –  References

•  Variances (deviations) –  From test plan or

requirements –  Reasons for deviations

•  Summary of incidents –  Resolved incidents –  Defect patterns –  Unresolved incidents

Adequacy assessment Evaluation of coverage Identify uncovered attributes

Summary of activities System/CPU usage Staff time Elapsed time

Software evaluation Limitations Failure likelihood

Approvals

52 © 2013 SQE Training V3.1

Page 27: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 27

Status •  Record  any  current  test  status  measures  that  you  are  using  here.  Give  them  a  grade  for  effec+veness  (A,  B,  C,  etc.)  

     •  Any  new  metrics  you  would  use?          *  Remember  you  need  at  least  two  metrics  per  instrument  

53 © 2013 SQE Training V3.1

Establish a Dashboard

•  Easy  to  use/understand  at  a  glance:  

 *  Remember  you  need  at  least  two  metrics  per  instrument  

Quality  of  product  

Status  

Test  effec-veness  

Resources  

Issues  

54 © 2013 SQE Training V3.1

Page 28: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 28

How Do You Measure Test Effectiveness?

55 © 2013 SQE Training V3.1

A Common Answer  

– Coverage  – Defect  age  (phase  or  product  version)  – #  of  bugs  – Defect  density  – Defect  removal  efficiency  – Defect  seeding  – Muta+on  analysis  – Customer  complaints  

56 © 2013 SQE Training V3.1

Page 29: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 29

Three Major Categories

57 © 2013 SQE Training V3.1

Customer Satisfaction Measures

Issues

Who  to  ask  

“A.er  the  fact”  

Difficulty  in  measuring  

Doesn’t  differen+ate  between  the  effec+veness  of  development  and  tes+ng  

58 © 2013 SQE Training V3.1

Page 30: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 30

Customer Satisfaction Measures •  Subjec+ve  is  good  •  Objec+ve  is  beqer    

59 © 2013 SQE Training V3.1

Defect Measures

•  Why  is  it  important  to  track  defects?  •  What  are  some  ways  to  analyze  defects?  •  DDP  •  Defect  density  •  Defect  age  

60 © 2013 SQE Training V3.1

Page 31: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 31

Why is it important to track defects? •       Iden+fy  process  improvement  •       Iden+fy  training  needs  •       Iden+fy  problema+c  (high-­‐risk)  areas  •       Determine  test  status  

61 © 2013 SQE Training V3.1

Defect Analysis ― Example

•  Phase  •  Type  •  Severity  •  Priority  •  Author  •  Age  •  Module  

62 © 2013 SQE Training V3.1

Page 32: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 32

Defect Detection Percentage (DDP)

DDP = Defects Discovered

x 100% Defects at Start

85% is the average DRE for US software projects greater than 1,000 function points in size.

— Capers Jones

63 © 2013 SQE Training V3.1

Defect Detection Percentage (DDP)

Issues  

Severity  and  distribu+on  of  defects  

How  to  know  when  all  bugs  are  found  

“A.er  the  fact”  

What  cons+tutes  bug-­‐finding  ac+vi+es?  

Some  bugs  cannot  be  found  in  tes+ng  

64 © 2013 SQE Training V3.1

Page 33: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 33

Defect “Value” (Cost Avoidance)

Requirements 1 High level design 1 Detailed design 1 Code 1 Unit Test 3 – 5 Integration test 5 –10 System/acceptance test 10 – 30 Production 20 – 60+

When discovered Typical hours to rework/fix

65 © 2013 SQE Training V3.1

Defect Age (PhAge)

Requirements High level design Detailed design Coding

Req

uire

men

ts

Hig

h le

vel d

esig

n

Det

aile

d de

sign

Cod

ing

Uni

t tes

ting

Inte

grat

ion

test

ing

Syst

em te

stin

g

Acc

epta

nce

test

ing

Pilo

t

Prod

uctio

n

0

Phase discovered

Phase created

1 4 3 2 9 8 7 6 5

0 3 2 1 8 7 6 5 4

2 1 0 7 6 5 4 3

1 0 6 5 4 3 2

66 © 2013 SQE Training V3.1

Page 34: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 34

Defect Age

Issues

Difficult to do root cause

Requires weighting of defects

How to handle latent/masked defects

67 © 2013 SQE Training V3.1

Coverage Measures

Discussion Requirements vs. design vs. code coverage

Completeness/accuracy of test basis

Coverage of test set vs. coverage of tests executed (e.g., we don’t always run every test)

Coverage vs. actual results (DDP)

68 © 2013 SQE Training V3.1

Page 35: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 35

Mapping Test Cases to Requirements

Requirements spec.

3.5.1.3.2

…..

3.5.1.4.7

…..

3.6.4.2.1

…..

3.8.2.7.1

Test plan

Test Case #3

…..

Test Case #5

…..

Test Case #12

…..

Test Case #19

69 © 2013 SQE Training V3.1

Requirements/Design Coverage

Test Case 1 2 3 Covered?

Requirement A X X Y B N C X X Y

Feature A X Y B X X Y

Design A X X Y B X Y C N D X X Y

Conceptual model of requirements/ design coverage:

70 © 2013 SQE Training V3.1

Page 36: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 36

Requirements/Design Coverage

Issues Only as good as test basis

Relatively low coverage of code

Code  coverage  achieved  with  requirements  tests    Major  bank  (20  apps)    20%    Major  DBMS  vendor    47%    Major  h/w  s/w  vendor  60%  

                                                           —  Source:    Bender  and  Associates  

71 © 2013 SQE Training V3.1

Code Coverage

Test run 1 2 3 Covered?

Statement A X X X Y B X X Y C X Y D N E X Y 60% 20% 60% 80%

Conceptual model of code coverage

72 © 2013 SQE Training V3.1

Page 37: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 37

Code Coverage

Issues Requires a tool

Doesn’t prove the code actually “works” correctly

Did we test the “right code”?

Statement vs. branch vs. path

73 © 2013 SQE Training V3.1

Test Effectiveness •  Record  any  current  test  effec+veness  measures  that  you  are  using  here.  Give  them  a  grade  for  effec+veness  (A,  B,  C,  etc.)  

•  Any  new  metrics  you  would  use?  

   *  Remember  you  need  at  least  two  metrics  per  instrument  

74 © 2013 SQE Training V3.1

Page 38: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 38

Establish a Dashboard

•  Easy  to  use/understand  at  a  glance:  

*  Remember  you  need  at  least  two  metrics  per  instrument  

Quality  of  product  

Status  

Test  effec+veness  

Resources  

Issues  

75 © 2013 SQE Training V3.1

Resources •  Resource  es+mates/consump+on  are  necessary  in  order  to  do  test  planning,  es+ma+on,  budge+ng,  and  staffing  

•  You  must  consider  the  level  of  granularity  in  the  collec+on  of  these  metrics  based  on  the  accuracy  of  the  required  metrics  and  your  ability  to  validate  them  

•  Some  people  choose  to  exclude  the  resources  instrument  from  the  dashboard  because  they  feel  it  is  not  a  “day  to  day”  metric  

76 © 2013 SQE Training V3.1

Page 39: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 39

Resources Resource  metrics  are  normally  collected  in  terms  of  

Actual/expected  budget  

Actual/expected  engineering  hours  

Test  environment  u+liza+on/availability  

Staffing  levels  

Contractor  availability  

Other  hardware/so.ware  resources  

77 © 2013 SQE Training V3.1

Resources •  Record  any  current  resource  measures  that  you  are  using  here.  Give  them  a  grade  for  effec+veness  (A,  B,  C,  etc.)  

•  Any  new  metrics  you  would  use?  

   *  Remember  you  need  at  least  two  metrics  per  instrument  

78 © 2013 SQE Training V3.1

Page 40: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 40

Establish a Dashboard

•  Easy  to  use/understand  at  a  glance:  

 *  Remember  you  need  at  least  two  metrics  per  instrument  

Quality  of  Product  

Status  

Test  Effec+veness  

Resources  

Issues  

79 © 2013 SQE Training V3.1

Issues

•  This  is  included  to  address  any  important  items  not  otherwise  included  on  the  dashboard.  These  are  normally  subjec+ve  and  not  necessarily  conducive  to  systema+c  analysis  

•  Issues  could  involve  training,  installa+on  of  new  hardware/so.ware,  poli+cs—even  the  weather  

80 © 2013 SQE Training V3.1

Page 41: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 41

A Sample Tester’s Dashboard

Status • % completion • Defect info

Product Quality • Defect density • Performance, etc.

Test Effectiveness • DDP • Coverage

Resources • Engineering hours • Money

Issues

81 © 2013 SQE Training V3.1

Avoiding Dysfunction

•  Measure  processes  and  products—not  people!  

•  Beware  of  the  dark  side  of  the  Hawthorne  Effect  •  Remember  that  more  is  not  always  beqer  

•  Avoid  the  exclusive  use  of  top-­‐down  metrics  

•  Provide  training—not  all  metrics  are  intui+ve  

•  Consider  temporary  metrics  

 

82 © 2013 SQE Training V3.1

Page 42: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 42

Avoiding Dysfunction

•  Define  each  metric,  its  use,  who  will  see  it,  expected  ranges,  etc.  

•  Remember  your  audience  and  tailor  to  their  needs  

•  Always  seek  mul+ple  interpreta+ons  

•  Ask  your  audience  what  their  interpreta+on  of  a  metric  is  before  you  offer  yours  

•  Sell,  sell,  sell,  sell  

83 © 2013 SQE Training V3.1

One Truth and One Myth in Closing The  Truth  Gilb’s  Law  “Anything  you  need  to  quan+fy  can  be  measured  in  some  way  that  is  superior  to  not  measuring  it  at  all”            —Tom  Gilb  

 The  Myth  “Some  metrics  are  always  beqer  than  no  metrics  …”  

84 © 2013 SQE Training V3.1

Page 43: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 43

This Slide Is Hidden

85 © 2013 SQE Training V3.1

4 ESTIMATION (OPTIONAL)

Page 44: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 44

Estimation

Es+mate:      1.  A  tenta+ve  evalua+on  or  rough  calcula+on      2.  A  preliminary  calcula+on  of  the  cost  of  a  project      3.  A  judgment  based  upon  one’s  impressions;  opinion                                    —The  American  Heritage  DicSonary  

   It  is  very  difficult  to  make  a  vigorous,  plausible,  and  job-­‐risking  es-mate  that  is  derived  by  no  quan-ta-ve  method,  supported  by  li8le  data  and  cer-fied  chiefly  by  hunches  of  the  managers.  

       

— Fred Brooks 87 © 2013 SQE Training V3.1

Test Estimation

The  best  es+mates  

Represent  the  collec+ve  wisdom  of  prac++oners  and  have  their  buy-­‐in  

Provide  specific,  detailed  catalogs  of  the  costs,  resources,  tasks,  and  people  involved  

Present,  for  each  ac+vity  es+mated,  the  most  likely  cost,  effort,  and  dura+on  

Es+ma+on:  the  crea+on  of  an  approximate  target  for  costs  and  comple+on  dates      

88 © 2013 SQE Training V3.1

Page 45: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 45

Test Estimation (cont.)

Factors  that  can  influence  cost,  effort,  and  dura+on  include:  

Required  level  of  quality  of  the  system  

Size  of  the  system  to  be  tested  

Historical  data  

Process  factors  (process  maturity,  etc.)  

Material  factors  (tools,  data,  etc.)  

People  factors  (skills,  experience,  managers,  etc.)  

89 © 2013 SQE Training V3.1

Test Estimation (cont.)

•  Delivery  of  es+mates  should  include  jus+fica+on  

•  Nego+a+on  and  re-­‐work  of  es+mates  is  normal  

•  Final  es+mates  represent  a  balance  of  organiza+onal  and  project  goals  in  the  areas  of  quality,  schedule,  budget,  and  features  

90 © 2013 SQE Training V3.1

Page 46: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 46

How Good Is Our Industry (at Estimating)? •  Tata: 62% of projects fail to meet schedule

49% have budget overruns •  Moiokken and Jorgensen: 30-40% overruns              

91 © 2013 SQE Training V3.1

Class Discussion Why  is  es+ma+ng  not  done  well?  Your  top  five  reasons:  

1)  Too  many  variables____________________  

2)  ____________________________________  

3)  ____________________________________  

4)  ____________________________________  

5)      ____________________________________  

 92 © 2013 SQE Training V3.1

Page 47: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 47

Why Estimates Are Inaccurate ― Part I •  Lack  of  es+ma+ng  experience  •  Lack  of  historical  data  on  which  to  base  es+mates  •  Lack  of  systema+c  es+ma+on  process,  sound  techniques,  or  models  suited  to  the  project  

•  Failure  to  include  essen+al  ac+vi+es  and  products  within  the  scope  of  the  es+mates  

•  Unrealis+c  expecta+ons  or  assump+ons  •  Failure  to  recognize  and  address  the  uncertainty  inherent  in  project  es+mates  

       PracScal  SoUware  Measurement          Addison-­‐Wesley,  2001  

           

93 © 2013 SQE Training V3.1

Why Estimates Are Inaccurate ― Part II •  Lack  of  educa+on  and  training  •  Confusing  the  target  with  the  es+mate  •  Hope-­‐based  planning  •  Inability  to  communicate  and  support  es+mates  

•  Incomplete,  changing,  and  creeping  requirements  

•  Quality  surprises  (test  and  re-­‐fix)  

       —adapted  from  Linda  M.  Laird          The  LimitaSons  of  EsSmaSon  

94 © 2013 SQE Training V3.1

Page 48: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 48

Bohem’s Cone of Uncertainty

95 © 2013 SQE Training V3.1

NHC Track Forecast Cone

96 © 2013 SQE Training V3.1

Page 49: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 49

“Testing” Track Forecast Cone

Best case

Expected case

Worst case

T i m e

T as k

s

(or why it is important to constantly re-estimate)

97 © 2013 SQE Training V3.1

The Fantasy Factor

Weeks

Today

0 1 2 3 4 5 6 7 8 9

1st 3rd 2nd

What  would  have  to  happen  to  deliver  this  in  four  weeks?  

What  should  the  es+mate  have  been?  

98 © 2013 SQE Training V3.1

Page 50: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 50

Estimation

1,  2,  3,  or  4  Variables  +  Many  Modifiers:    

If it’s not variable, then it’s fixed.

Time  

Size   Resources  

99 © 2013 SQE Training V3.1

Time vs. Resources

=

100 © 2013 SQE Training V3.1

Page 51: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 51

Test Estimation Techniques ― Examples •  Intui+on  and  guess  •  Work-­‐breakdown-­‐structures  •  Three-­‐point  es+mates  •  Company  standards  and  norms  •  %  of  project  effort  or  staffing  •  Industry  averages  and  predic+ve  models  (e.g.,  FP,  TPA  )  •  Team  es+ma+on  sessions  

– Wideband  Delphi  –  Story  point    sizing  –  Poker  es+ma+on  –  T-­‐shirt  sizing  

101 © 2013 SQE Training V3.1

Karl Wiegers’s Estimation Safety Tips

•  A  goal  is  not  an  es+mate  •  The  es+mate  you  produce  should  be  unrelated  to  what  you  think  the  requester  wants  to  hear  

•  The  correct  answer  to  any  request  for  an  es+mate  is  “Let  me  get  back  to  you  on  that”  

•  Avoid  giving  single  point  es+mates  •  Incorporate  con+ngency  buffers  into  es+mates   102 © 2013 SQE Training V3.1

Page 52: Measurement and Metrics for Test Managers

© 2013 SQE Training V3.1 52

Rick Craig’s Tips for Better Estimates •  Do  it!  •  Collect  metrics  •  Remember  the  “fantasy”  factor  •  Don’t  “pad”  your  es+mates*  •  Don’t  spend  a  ton  of  +me  •  Es+mates  don’t  have  to  be  perfect  

–  Es+mates  are  just  es+mates  –  They  will  change/constantly  as  you  re-­‐es+mate  –  Remember  planning  risks  and  con+ngencies  –  Remember  Brooke’s  Law  

•  If  the  date  is  fixed,  es+mate  something  else  •  Use  tools  •  Use  ranges  of  value  instead  of  discrete  numbers  

103 © 2013 SQE Training V3.1