Your Test is Only as Good as Your Hypothesis How to Cra! Solid, Data-Driven Hypotheses

ConversionXL Live 2015

@ContentVerve

If we get a responsive website, we’ll make more money…

@ContentVerve

@ContentVerve

$$$ Cost of responsive redesign $$$ $$$ Serious loss of revenue $$$ $$$ Cost of “damage control” $$$ $$$ Cost of double work $$$

@ContentVerve

@ContentVerve

I’m a split test junkie!!!

@ContentVerve

Dumbass…

@ContentVerve

A solid test hypothesis is an informed solu"on to a real problem - not an arbitrary guess.

@ContentVerve

It’s the star"ng point for further research.

@ContentVerve

1. The change you want to make 2. How this change will affect your prospects 3. The impact you expect to see as a result

@ContentVerve

1. The change you want to make 2. How this change will affect your prospects 3. The impact you expect to see as a result

@ContentVerve

X into, ______ X I can get more By changing ______ X and thus increase ______. X prospects to ______

@ContentVerve

Hardwire Your Brain to Think in Hypotheses

@ContentVerve

Think in terms of “op"miza"on hypotheses” not “test hypotheses”

@ContentVerve

And now….. A Really Stupid A/B Test!

@ContentVerve

@ContentVerve

@ContentVerve

@ContentVerve

Control:

Treatment:

2 months data & 490 downloads

@ContentVerve

Control:

Treatment:

2 months data & 490 downloads No Real Difference

@ContentVerve

@ContentVerve

@ContentVerve

40% Check prices

37% Learn more

23% Book

@ContentVerve

40% Tjek prices

37% Learn more

23% Book

@ContentVerve

By adding a pricing page, I can bridge a serious gap in the decision-making process, mo"vate more prospects to say yes and thus increase the number of leads.

@ContentVerve

@ContentVerve

X into, ______ X I can get more By changing ______ X and thus increase ______. X prospects to ______

@ContentVerve

Total number of days to run the test: 314 days

@ContentVerve

If I can get a li# of 33%

Total number of days to run the test: 35 days

@ContentVerve

@ContentVerve

@ContentVerve

@ContentVerve

By adding a privacy policy, I can reduce fric"on, get more prospects to give me their info, and thus increase sign-ups.

@ContentVerve

Control:

Treatment:

100% privacy - we will never spam you!

@ContentVerve

Control:

Treatment:

18.70%  drop  in  sign-­‐ups Sta$s$cal  Confidence  Level  96%

@ContentVerve

By removing the word “spam” and adding a guarantee, I can reduce fric"on, get more prospects to give me their info, and thus increase sign-ups.

@ContentVerve

Control:

Treatment:

We guarantee 100% privacy. Your information will not be shared

@ContentVerve

Control:

Treatment:

19.47% increase in sign-ups Sta$s$cal  Confidence  Level  96%

@ContentVerve

By removing the guarantee, and adding the word “spam” I can increase fric"on, get less prospects to give me their informa"on, and thus decrease sign-ups.

@ContentVerve

@ContentVerve

24%  drop  in  downloads Sta$s$cal  Confidence  Level  99%

1 month of data / 349 conversions @ContentVerve

1. Conduct Research

2. Hypothesize

3. Create Treatment

4. Test Treatment

5. Analyze

6. Run Follow-Up Experiment @ContentVerve

1. Conduct Research

2. Hypothesize

3. Create Treatment

4. Test Treatment

5. Analyze

6. Run Follow-Up Experiment

@ContentVerve

1. Conduct Research

2. Hypothesize

3. Create Treatment

4. Test Treatment

5. Analyze

6. Run Follow-Up Experiment

@ContentVerve

Conversion Research

@ContentVerve

Device Category

Image Removed

@ContentVerve

Browser

Image Removed

@ContentVerve

Browser - Desktop Sessions

Image Removed

@ContentVerve

Other reports: Top landing pages Site Search New vs. returning Nav overview Site Speed Unbounced sessions Unbounced pr. device New vs. returning pr. device Learn to use advanced segments!

@ContentVerve

Slides Removed

Home Page

Category Page

Product Page

Cart

Checkout

Confirma"on

@ContentVerve

Home Page

Category Page

Product Page

Cart

Checkout

Confirma"on

@ContentVerve

By reorganizing the product page content, I can create a more logical user expereince, reduce cogni"ve fric"on, get more prospects to configure a product and thus increase purchases.

@ContentVerve

@ContentVerve

Top-Selling Product Categories

@ContentVerve

Top-Selling Products

@ContentVerve

@ContentVerve

Slide Removed

1.

2.

80% drop off

3.

54% drop off

4.

14% drop off

5.

66% drop off

6.

4% drop off

@ContentVerve

How to Build Funnels Manually Go to “All Pages” in GA

Type URL Structure

Write Down Unique Pageviews

@ContentVerve

@ContentVerve

@ContentVerve

@ContentVerve

“In the economy of ac"on, effort is a cost… Laziness is built deep into our system.” - Daniel Kahneman

@ContentVerve

@ContentVerve @ContentVerve

@ContentVerve @ContentVerve

@ContentVerve @ContentVerve

Al?

@ContentVerve

Michael

[email protected]

Al? Al? Daddy?

@ContentVerve

@ContentVerve

@ContentVerve @ContentVerve

Run Clean Tests, So You Are Acutally Ge$ng Insight!

@ContentVerve

@ContentVerve

@ContentVerve

Slide Removed

Qualita"ve Research

@ContentVerve

Interviews with: - Customers / Poten"al customers - Sales - Customer service / support

@ContentVerve

Interviews with:

- Customers / Poten"al customers - Sales - Customer service / support

@ContentVerve

Sales 1. What are the top 3 ques"ons you get from poten"al customers? 2. What do you answer to these ques"ons? 3. What is the biggest barrier keeping people from buying your product? 4. Are there any specfic selling points that work par"cularly well? If so, which ones? 5. Have we missed anything important? Got something to add? @ContentVerve

Customer service / Support 1. What are the top 3 ques"ons you get from poten"al customers? 2. What do you answer to these ques"ons? 3. Are there any par"cular aspects of X that people don’t understand? 4. What aspects of X do people like the most/least? 5. Have we missed anything important? Got something to add? @ContentVerve

Does A/B Tes"ng Make Sense on Your Website?

@ContentVerve

Your Test is Only as Good as Your Hypothesis The be%er the research, the be%er the insight The be%er the insight, the be%er the hypothesis The be%er the hypothesis, the be%er the test The be%er the test, the be%er the learning

@ContentVerve

Don’t be the Split Test Junkie

Be The Conversion Plumber

@ContentVerve

Thank You For Listening! ConversionXL Live 2015

Your test is never better than your hypothesis 2 - ConversionXL

@ContentVerve. Browser - Desktop Sessions. Image Removed ... How to Build Funnels Manually. Type URL. Structure. Write Down. Unique. Pageviews ...

13MB Sizes 0 Downloads 166 Views

Recommend Documents

Your test is never better than your hypothesis 2 - ConversionXL
How to Cra Solid, Data-Driven Hypotheses. ConversionXLLive 2015 ... prospects to ______ and thus increase ______. X. X. X. X .... Customer service / support.

Better Later than Never? An Experiment on Bargaining ...
Dec 2, 2016 - On the other hand, if the number of periods is small, the unique equilibrium ..... from business, economics and law degrees. ..... screen L-type sellers by continuously raising offers from a median of 800 to prices of 2750 ± 250.

Test your Google knowledge
Google is a trademark of Google Inc. All other company and product names may be trademarks of the respective companies with which they are associated.

Test your Google knowledge
e. All of the above. 2. What type of information is NOT viewable with Google Earth? http://earth.google.com/ a. Topographical features b. Traffic conditions c. Schools d. Crime statistics. 3. True or False: The headlines you see on Google News are se

As is Your Food So is Your Mind
started interrogating my servants, the scared culprit must have taken refuge under you. Being compassionate, as Saints by nature are, you have come to return ...