Why Adopt
Analytics Driven Testing
Ori Bendet | Inbound Product Manager, HPE Software |
@bendet_ori
#SrijanWW | @srijan
↓80%Reduction in regression time
7y in HPE Software
in various managerial positions
Started in testing
Today:
Inbound Product
Manager
StormRunner Functional
ABOUT ME
AGENDA • Applications Development Overview
• Today’s Testing Challenges
• What are the risks?
• Real World Examples
• How can you empower your testing by
using Analytics
The World of Software is changing
Mega-trends create demands on modern applications
MODERN APPLICATION DEVELOPMENT
Reduce costs
Increase customer attraction/retention
Increase the value of your brand
Get to market faster
300,000
tweets
200 million+
emails sent
220,000
new photos posted
50,000 apps
downloaded
$80,000
in online sales
72 hours
of new video content uploaded
2.5 million
pieces of content shared
Agile
Every minute*…
* Source: “The Data Explosion in 2014 Minute-by-Minute”by ACI InformationGroup.
http://aci.info/2014/07/12/the-data-explosion-in-2014-minute-by-minute-infographic/
Mobile
Cloud
Digital
DevOps
8
What we need to
test
TestingTime
=Quality
$59.5BAnnual cost of software defects
How to
mitigate the
risk?
• Have more testers
• Crowdsource
• Don’t test
• Let your users test for you
• Or… use Analytics!
Analytics
Example
Example • My product: 30 features
• Supported Matrix: 10 environments
(OS/Browser)
• My testing team: 10 testers
• Testing a feature on an environment = 1 day
• Target regression cycle: <3 days!
30 features X 10 combinations = 300 days
300 days / 10 testers = 30 days
Google Analytics • A freemium web analytics service offered by Google
that tracks and reports website traffic.
• Google launched the service in November 2005
after acquiring Urchin.
• Google Analytics is now the most widely used web
analytics service on the Internet.
• Note: I’m not affiliated with Google Analytics in
anyway
* Source: https://w3techs.com/technologies/overview/traffic_analysis/all
Environment
Optimization • Focus on the majority of combinations (OS x
Browser)
• Reduce your regression risk to a minimum
• Wait for customer feedback on the other areas
Let’s see it live
30 features X 10 2 combinations* = 60 days
60 days / 10 testers = 6 days
(*) with 89%
confidence
Example
User Behavior • Focus on the majority of the functionality
• Target the most active areas
• Wait for customer feedback on other areas (% of escaped
defects)
• Calculated risk(!)
Let’s see it live - again
0 12 features X 10 2 combinations* = 24 day
24 days / 10 testers = 2.4 days
(*) with 76%
confidence
Example – part
2
Building a risk
calculator • Test everything (100%) = 300 days
• All functionality, top combinations = 60 days
• Main features, top combinations = 24 days
• Add additional levels based on your needs and
time
Building a risk
calculator
Top flows
1 OS +
Browser
Main
Functionalitie
s
top
combinations
All
Functionalitie
s
top
combinations
All
Functionalitie
s
Additional
combinations
Test
Everything
50%
75%
90%
100%
95%
BewareZombie tests!
Don’t have
analytics? • Use market analysis and statistics
• https://netmarketshare.com/
• https://www.w3schools.com/browsers/
• Many more!
• Conduct annual user survey
• CustomersValidations – before & after
A word about Test Automation
Re-use calculator for
Automation
Top flows
1 OS +
Browser
Main
Functionalitie
s
top
combinations
All
Functionalitie
s
top
combinations
All
Functionalitie
s
Additional
combinations
Test
Everything
50%
75%
90%
100%
95%
CI
Nightly
Full
regression
Summary
Summary • Don’t test everything!
• Get to know your users
• Calculate your risk
• Measure your un-tested areas
• % of Escaped defects
• Customer Satisfaction
[Srijan Wednesday Webinars] Why Adopt Analytics Driven Testing

[Srijan Wednesday Webinars] Why Adopt Analytics Driven Testing

  • 1.
    Why Adopt Analytics DrivenTesting Ori Bendet | Inbound Product Manager, HPE Software | @bendet_ori #SrijanWW | @srijan
  • 2.
  • 3.
    7y in HPESoftware in various managerial positions Started in testing Today: Inbound Product Manager StormRunner Functional ABOUT ME
  • 4.
    AGENDA • ApplicationsDevelopment Overview • Today’s Testing Challenges • What are the risks? • Real World Examples • How can you empower your testing by using Analytics
  • 5.
    The World ofSoftware is changing Mega-trends create demands on modern applications MODERN APPLICATION DEVELOPMENT Reduce costs Increase customer attraction/retention Increase the value of your brand Get to market faster 300,000 tweets 200 million+ emails sent 220,000 new photos posted 50,000 apps downloaded $80,000 in online sales 72 hours of new video content uploaded 2.5 million pieces of content shared Agile Every minute*… * Source: “The Data Explosion in 2014 Minute-by-Minute”by ACI InformationGroup. http://aci.info/2014/07/12/the-data-explosion-in-2014-minute-by-minute-infographic/ Mobile Cloud Digital DevOps
  • 6.
  • 7.
    What we needto test TestingTime
  • 8.
  • 9.
    $59.5BAnnual cost ofsoftware defects
  • 10.
    How to mitigate the risk? •Have more testers • Crowdsource • Don’t test • Let your users test for you • Or… use Analytics!
  • 11.
  • 12.
  • 13.
    Example • Myproduct: 30 features • Supported Matrix: 10 environments (OS/Browser) • My testing team: 10 testers • Testing a feature on an environment = 1 day • Target regression cycle: <3 days!
  • 14.
    30 features X10 combinations = 300 days 300 days / 10 testers = 30 days
  • 15.
    Google Analytics •A freemium web analytics service offered by Google that tracks and reports website traffic. • Google launched the service in November 2005 after acquiring Urchin. • Google Analytics is now the most widely used web analytics service on the Internet. • Note: I’m not affiliated with Google Analytics in anyway * Source: https://w3techs.com/technologies/overview/traffic_analysis/all
  • 16.
    Environment Optimization • Focuson the majority of combinations (OS x Browser) • Reduce your regression risk to a minimum • Wait for customer feedback on the other areas
  • 17.
  • 18.
    30 features X10 2 combinations* = 60 days 60 days / 10 testers = 6 days (*) with 89% confidence
  • 19.
  • 20.
    User Behavior •Focus on the majority of the functionality • Target the most active areas • Wait for customer feedback on other areas (% of escaped defects) • Calculated risk(!)
  • 21.
    Let’s see itlive - again
  • 22.
    0 12 featuresX 10 2 combinations* = 24 day 24 days / 10 testers = 2.4 days (*) with 76% confidence
  • 23.
  • 24.
    Building a risk calculator• Test everything (100%) = 300 days • All functionality, top combinations = 60 days • Main features, top combinations = 24 days • Add additional levels based on your needs and time
  • 25.
    Building a risk calculator Topflows 1 OS + Browser Main Functionalitie s top combinations All Functionalitie s top combinations All Functionalitie s Additional combinations Test Everything 50% 75% 90% 100% 95%
  • 26.
  • 28.
    Don’t have analytics? •Use market analysis and statistics • https://netmarketshare.com/ • https://www.w3schools.com/browsers/ • Many more! • Conduct annual user survey • CustomersValidations – before & after
  • 29.
    A word aboutTest Automation
  • 30.
    Re-use calculator for Automation Topflows 1 OS + Browser Main Functionalitie s top combinations All Functionalitie s top combinations All Functionalitie s Additional combinations Test Everything 50% 75% 90% 100% 95% CI Nightly Full regression
  • 31.
  • 32.
    Summary • Don’ttest everything! • Get to know your users • Calculate your risk • Measure your un-tested areas • % of Escaped defects • Customer Satisfaction

Editor's Notes

  • #3 http://blog.celerity.com/the-true-cost-of-a-software-bug
  • #6 Set the stage, have a conversation with the audience about their biggest challenges, changes and painpoints in application delivery
  • #11 http://blog.celerity.com/the-true-cost-of-a-software-bug
  • #12 http://blog.celerity.com/the-true-cost-of-a-software-bug
  • #15 30 features = 30 days 10 OS + Browsers combinations = 300 days
  • #16 30 features = 30 days 10 OS + Browsers combinations = 300 days
  • #17 30 features = 30 days 10 OS + Browsers combinations = 300 days
  • #18 How can Google Analytics help you??
  • #19 How can Google Analytics help you??
  • #20 Advantage Online Shopping – Browser Overview
  • #21 30 features = 30 days 10 OS + Browsers combinations = 300 days
  • #22 30 features = 30 days 10 OS + Browsers combinations = 300 days
  • #24 Google Demo Account – Behavior
  • #25 30 features = 30 days 10 OS + Browsers combinations = 300 days