Professional Documents
Culture Documents
Data-Driven Design: Making Software & Websites Better Through Use of Statistics Patrick Mckenzie April 15Th, 2010
Data-Driven Design: Making Software & Websites Better Through Use of Statistics Patrick Mckenzie April 15Th, 2010
Patrick McKenzie
April 15th, 2010
Creative Commons Licensed CC-BY. Please feel free to share.
Outline
Introduction
自己紹介
Art & Science
芸術と技術
Rationale
目的
Instrumentation
動作監視
Funnel Analysis
ファネル分析
A/B Testing
A/B テスト
Conclusion
まとめ
The Plan
What Happened
What Happened Next
Started Business
Art into Science 芸術から技術へ
デザイナーの芸術を技術の倫理、数学、体
験方式で拡張する。
Rationale 目的
Decrease Costs
開発工数ダウン
Decrease
不確実性ダウン
Uncertainty
ユーザ足度アップ
満足 度ア ップ
Increase Satisfaction
成功率アップ
Increase Success
利益アップ
Increase Profits
Designing With Data
Get Facts to Answer The Questions:
What are users doing? 何をしているか?
Which users are doing things? 誰がしている
か?
Where are they doing things? 何処にしている
か?
(Maybe) Why are they doing it? 何故している
か?
Types of Instrumentation
Analytics software
Heatmapping
Surveys
Ad hoc analysis (Excel, etc)
Custom software
CrazyEgg
Landing Page
Photo of problem solved goes here
Landing Page (Fixed)
Paving The Trails
Picture of paving a path
Instrumenting Software
Instrumenting Software #2
Funnel Analysis ファネル分析
Many applications have several steps followed
by a goal.
Only a percentage of users clear each step.
Funnel analysis finds where folks are falling
out, to concentrate efforts on improvement
Viewed Signed
Landing Did $
Up For
Page Work $
Trial
Software ソフト
Google Analytics
Free but limited
Good for
starter
Mixpanel
Cheap
API
Harder to use
Define Funnel
1.Signs In
2.Opens Word List
3.Saves Word List
4.Customizes Cards
5.Creates Cards
6.Downloads Card
Collect Data
Open Word List needlessly complicated: make
easier.
Save Button was unclear: tweak microcopy,
button design.
Customize Cards needlessly complicated:
show less options.
Before
After
Collect Metrics
Divide similar population into two groups
randomly.
Treat the groups differently.
Measure difference in outcomes.
Scientific Method applied to design.
Eliminates bias, randomness, poor test
design.
Actual A/B Test
Half of Users See... Half of Users See...
Software Tools
Google Website Optimizer
Terrible. Do not use.
Visual Website Optimizer
visualwebsiteoptimizer.com
WYSIWYG alternative creation. Very nice.
Free for now.
A/Bingo or similar
Rails A/B testing library.
Programmer friendly, powerful, easy to use.
Core A/B Testing Loop
Identify conversion and instrument.
Create alternatives.
Assign users randomly to alternatives
Watch behavior.
Identify winning alternative.
Iterate, iterate, iterate...
Landing Pages
Source: whichtestwon.com
Signup Forms
Signup Forms
Signup Forms
New Features
60 ~ 80% of features do not change user behavior
Eric Ries
Old Version
Spot The New Feature
Core A/B Testing Discipline
Accept that you do not know what is right.
(Hardest lesson.)
Test frequently, as a matter of routine. (Helps
to have one-line testing.)
Test changes in isolation.
Use evidence from tests to inform your
marketing and development plans.
Conclusion
What we learned today:
Importance of tracking user interaction.
Funnel analysis – measuring user success and
goal completion.
A/B tests – improving marketing message and
product design.
Thanks for listening. ご清聴、ありがとうござ
いました。
patrick@kalzumeus.com ← My email.
http://www.kalzumeus.com ← My blog.
http://www.abingo.org ← A/Bingo software
(OSS)
http://www.startuplessonslearned.com ← Eric
Ries, very useful.
Photo Credits
Microsoft logo: http://www.flickr.com/photos/scobleizer/2264763977/
Sleeping Salaryman: http://www.flickr.com/photos/kansai/3452417503/
Edamame (Green Soybeans): http://www.flickr.com/photos/kankan/33346213/
Salty Peanuts: http://www.flickr.com/photos/euromagic/2351628831/
Facepalm Statue: http://www.flickr.com/photos/cesarastudillo/3981364314/