Slide 1

Slide 1 text

@"mminor   An Introduction to A/B Testing Photo  credit:  @lucyjspence  and  h7p://www.flickr.com/photos/futureshape/4530140521/  

Slide 2

Slide 2 text

What we’ll be looking at •  What  is  A/B  Tes"ng?   •  What  to  test?   •  What  is  A/B  Tes"ng  good  for?   •  A/B  tes"ng  should  not  replace   •  Do’s  and  Don’ts   •  Tools  

Slide 3

Slide 3 text

Source:  h7p://www.smashingmagazine.com/2010/06/24/the-­‐ul"mate-­‐guide-­‐to-­‐a-­‐b-­‐tes"ng/   What is A/B Testing?

Slide 4

Slide 4 text

What to test? •  Bu7ons:  colours,  sizes,  loca"ons   •  Headlines   •  Product  descrip"ons   •  Form  fields  (adding  and  removing)   •  PPC  landing  pages   •  Email  content   •  Pricing  and  promo"ons   •  Images  

Slide 5

Slide 5 text

What is A/B Testing good for? •  Data-­‐driven  design   •  Ge\ng  stakeholders  on-­‐board   •  Avoiding  personal  preferences  

Slide 6

Slide 6 text

A/B testing should not replace: •  Talking  to  users   •  Usability  tests   •  Acceptance  tests  

Slide 7

Slide 7 text

Do’s and Don’ts •  Test  simultaneously   •  Test  o^en   •  Include  only  new  visitors   •  Show  repeat  visitors  the  same  version   •  Use  results  to  inform  future  development   •  Require  sta"s"cal  significance   DO:

Slide 8

Slide 8 text

•  Test  too  many  ideas  at  once   •  No  clear  idea   •  Don’t  let  your  preconcep"ons  override  results   •  Peek  at  your  results  and  stop  as  soon  as  a   significant  result  has  been  achieved.     See:  Evan  Miller.   Do’s and Don’ts DON’T:

Slide 9

Slide 9 text

The tools Other  tools  are  available:   A/Bingo,  Vanity,  Unbounce,  Performable,  Vertster,  SiteSpect,     Webtrends  Op"mize  and  Omniture’s  Test&Target  

Slide 10

Slide 10 text

The Quiz Show

Slide 11

Slide 11 text

The Corkscrew A

Slide 12

Slide 12 text

The Corkscrew B

Slide 13

Slide 13 text

The Corkscrew A/B

Slide 14

Slide 14 text

A

Slide 15

Slide 15 text

Great American Days A

Slide 16

Slide 16 text

Great American Days B

Slide 17

Slide 17 text

Great American Days A/B

Slide 18

Slide 18 text

B

Slide 19

Slide 19 text

Real Gap

Slide 20

Slide 20 text

Real Gap A/B

Slide 21

Slide 21 text

B

Slide 22

Slide 22 text

CityCliq A

Slide 23

Slide 23 text

CityCliq B

Slide 24

Slide 24 text

CityCliq A/B

Slide 25

Slide 25 text

A

Slide 26

Slide 26 text

Vendio A

Slide 27

Slide 27 text

Vendio B

Slide 28

Slide 28 text

Vendio B

Slide 29

Slide 29 text

A

Slide 30

Slide 30 text

Quick Fire Round •  What  were  the  two  main  tools  men"oned?   •  What  percentage  increase  did  the  winning   RealGap  test  achieve?   •  Name  one  of  the  three  things  A/B  tes"ng   doesn’t  replace.  

Slide 31

Slide 31 text

What we looked at •  What  is  A/B  Tes"ng?   •  What  to  test?   •  What  is  A/B  Tes"ng  good  for?   •  A/B  tes"ng  should  not  replace   •  Do’s  and  Don’ts   •  Tools  

Slide 32

Slide 32 text

References •  First  slide  photograph:  h7p://www.flickr.com/photos/futureshape/4530140521/     •  www.google.com/websiteop"mizer     •  h7p://visualwebsiteop"mizer.com/   •  h7p://www.abtests.com/   •  h7p://whichtestwon.com/     •  h7p://www.smashingmagazine.com/2010/06/24/the-­‐ul"mate-­‐guide-­‐to-­‐a-­‐b-­‐tes"ng/   •  h7p://www.conversion-­‐rate-­‐experts.com/ar"cles/101-­‐google-­‐website-­‐op"mizer-­‐"ps/   •  h7p://www.evanmiller.org/how-­‐not-­‐to-­‐run-­‐an-­‐ab-­‐test.html   •  h7p://elem.com/~b"lly/effec"ve-­‐ab-­‐tes"ng/   •  h7p://remoteresear.ch/