ISITE Design's Ryan Summers presents "Always Be Testing: How to lift your business by testing with Google Content Experiments" at GAUGE conference, October 2012.
4. $39.5BILLION
IN 2012
Spent on digital advertising and
promotions in the US
$62 BILLION
BY 2016
http://www.go-gulf.com/blog/online-ad-spending
5. 57%
Of companies that
test demonstrate
positive ROI
47% Don’t know or
do not calculate
MarketingSherpa 2011 Landing Page Optimization Benchmark Report
6. • Learn more about your customers
• Deliver better user experiences
• Get better return on your digital
investments
• Build a culture of data-driven and user-
focused design
• Discover something totally new and
unexpected!
Companies that test…
VALUE OF TESTING
7. Set Goals Find Friction
Share & Make
Learn Hypothesis
Test It!
Keep it simple
5 STEPS TO GET STARTED
11. • Multivariate Testing
– No Redirects
– Requires more time and more traffic
– Requires less design resources
– Element Contribution
Multivariate testing
KEY DIFFERENCES
12. 1.Be clear and specific
2.Increase trust
3.Customer first
Eliminate Friction
3 SIMPLE RULES
13. Form Logic
Form Layout
Location of CTA
Headline Copy
Navigation Logic/
Navigation Location
Complexity
Number of CTAs
Button Copy
Body Copy
Content of Images
Page Header Image
Balance of Text and Hero Shot
Graphics
Page Length
Modal Box Design
Number of Columns
* Impact analysis provided by Marketing Sherpa 2011 Optimization Benchmark Report
Prioritize
WHAT TO TEST
14. First focus on the elements on a page and which pieces are important to the experience
and which ones are not. These tests tend to run more quickly because they are more
drastic changes. They are also easier to code because they are usually simple style
changes and do not require high-cost creative assets.
• Real-Estate
• Location
• Size
• Existence
• Position
Triage
WHAT TO TEST
15. After you have identified pieces of the page that are high contributors to
conversion, you can start testing them in combinations and with more granularities.
• Presentation
• What it looks like
• Function
• What it does
• How it is programmed
• How it reacts to user input
• Copy
• What it says
• Audience
• Who it is targeted to
Refine
WHAT TO TEST
16. Significant Impact by Element
60%
50%
50%
41%
40%
36%
30%
30%
20% 17% 18%
10%
0%
Copy CTA Images Layout Navigation Form
Marketing Sherpa 2011 LPO Benchmark Report
Top elements
WHAT TO TEST
18. 2OO% MORE
LEADS
Using Google Website Optimizer we
were able to identify three themes that
changed the way SolarWorld USA
approached lead generation.
http://www.isitedesign.com/insight-blog/12_05/solarworld-americas
23. GCE
Google has a FREE testing tool, so
you have no excuse not to try testing.
24. • Only the control script will be necessary to run tests
• Google Analytics advanced segments and page metrics will
be available along with optimizing goal conversions based on
your test pages.
• Improved statistical engine for analyzing experiments, which
will help making decisions faster about the winning/loosing
pages.
• Dynamic Traffic Allocation functionality: traffic will be shifted
away from low-performing variations, over to higher
performing ones.
• Makes you wait for 2 weeks before it shows a winner.
• Tests expire after 90 days
Google Content Experiments
NEW FEATURES
25. • No multivariate testing
• Limited to 5 variations
• Best for landing pages
• Can’t test global elements
• No E-Commerce Transaction Goals
• 12 test per profile at once
• Each test requires it’s own code
Google Content Experiments
LIMITATIONS
26. • No multivariate testing
• Limited to 5 variations
• Best for landing pages
• Can’t test global elements
• No E-Commerce Transaction Goals
• 12 test per profile at once
• Each test requires it’s own code
Google Content Experiments
LIMITATIONS