Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.
The Dos and Don’ts of 
Hypothesis Testing 
{ 
@ttorres 
Teresa Torres 
www.ProductTalk.org 
! 
! 
! 
!
Product & UX are going through a 
mindset shift. 
@ttorres
Old way: 
Build 10 things, find out only 3 worked. 
@ttorres
New way: 
Run 10 experiments, build the 3 that worked. 
@ttorres
You can waste thousands of dollars, hundreds of 
hours, and countless sprints running 
experiments that don’t 
matter. 
@t...
How to get the most from your experiments. 
@ttorres
Know what you want to learn. 
@ttorres
Don’t throw spaghetti at the wall and see 
what sticks. 
@ttorres
Understand what level you are testing. 
@ttorres
Start with a testable hypothesis. 
@ttorres
A good hypothesis can be supported or 
refuted by an experiment. 
@ttorres
The 5 components of a good hypothesis. 
@ttorres
Have a clear reason for why you think the 
change will drive the desired impact. 
@ttorres
Choose the right participants. 
@ttorres
Draw a line in the sand. 
@ttorres
Start with your baseline. 
@ttorres
Look at comparable changes. 
@ttorres
Assess how much impact makes the 
investment worth it. 
@ttorres
Don’t stop your test at the wrong time. 
@ttorres
Don’t test too many variations. 
@ttorres
Mix-and-match methods to get the 
answers you need. 
@ttorres
Be mindful of the conclusions you draw. 
@ttorres
Know what you want to learn. 
• Don’t test everything. 
• Understand which level you are testing. 
Start with a testable h...
Let’s keep the 
conversation going. 
ProductTalk.org 
@ttorres 
Teresa Torres 
Product Consultant & Coach 
! 
! 
{
Upcoming SlideShare
Loading in …5
×

The Dos and Don'ts of Hypothesis Testing

5,659 views

Published on

Quick tips on how to improve your Lean experiments. Learn faster. Make better product decisions.

Published in: Internet
  • D0WNL0AD FULL ▶ ▶ ▶ ▶ http://1lite.top/5IZ2d ◀ ◀ ◀ ◀
       Reply 
    Are you sure you want to  Yes  No
    Your message goes here
  • check out the cool video of me training my dog on this page. ●●● http://t.cn/Aie43hbV
       Reply 
    Are you sure you want to  Yes  No
    Your message goes here

The Dos and Don'ts of Hypothesis Testing

  1. 1. The Dos and Don’ts of Hypothesis Testing { @ttorres Teresa Torres www.ProductTalk.org ! ! ! !
  2. 2. Product & UX are going through a mindset shift. @ttorres
  3. 3. Old way: Build 10 things, find out only 3 worked. @ttorres
  4. 4. New way: Run 10 experiments, build the 3 that worked. @ttorres
  5. 5. You can waste thousands of dollars, hundreds of hours, and countless sprints running experiments that don’t matter. @ttorres
  6. 6. How to get the most from your experiments. @ttorres
  7. 7. Know what you want to learn. @ttorres
  8. 8. Don’t throw spaghetti at the wall and see what sticks. @ttorres
  9. 9. Understand what level you are testing. @ttorres
  10. 10. Start with a testable hypothesis. @ttorres
  11. 11. A good hypothesis can be supported or refuted by an experiment. @ttorres
  12. 12. The 5 components of a good hypothesis. @ttorres
  13. 13. Have a clear reason for why you think the change will drive the desired impact. @ttorres
  14. 14. Choose the right participants. @ttorres
  15. 15. Draw a line in the sand. @ttorres
  16. 16. Start with your baseline. @ttorres
  17. 17. Look at comparable changes. @ttorres
  18. 18. Assess how much impact makes the investment worth it. @ttorres
  19. 19. Don’t stop your test at the wrong time. @ttorres
  20. 20. Don’t test too many variations. @ttorres
  21. 21. Mix-and-match methods to get the answers you need. @ttorres
  22. 22. Be mindful of the conclusions you draw. @ttorres
  23. 23. Know what you want to learn. • Don’t test everything. • Understand which level you are testing. Start with a testable hypothesis. • Include the 5 components. Have a clear reason why the change will have the desired impact. Choose the right participants. Draw a line in the sand. • Start with your baseline. • Look at comparable changes. • Assess how much impact makes the investment worth it Don’t stop your test at the wrong time. Don’t test too many variations. Mix-and-match methods to get the answers you need. Be mindful of the conclusions you draw. @ttorres
  24. 24. Let’s keep the conversation going. ProductTalk.org @ttorres Teresa Torres Product Consultant & Coach ! ! {

×