Upload
marketingexperiments
View
1.330
Download
1
Embed Size (px)
DESCRIPTION
View the full replay here: http://www.marketingexperiments.com/marketing-optimization/what-your-customers-want.html
Citation preview
#webclinic
Join the conversation on Twitter
#webclinic
What Your Customers Want: How to predict customer behavior for maximum ROI
#webclinic
Join the conversation on Twitter
#webclinic
#webclinic
Today’s team
Dr. Flint McGlaughlin Managing Director
Jon Powell Senior Research Manager
#webclinic
5
Marketing takes time and resources
The Ideal Customer The Marketer
The Message
No Thanks
Time
Energy
Budget
? How can I know that my campaigns will be successful before I invest resources?
#webclinic
6
Background: Medical provider specializing in treating a particular chronic pain. Goal: To plan a content marketing strategy based on which approach generates more appeal in condition-based searchers. Primary Research Question: Which content approach will achieve a higher clickthrough rate? Approach: A/B Multifactor Split Test
Experiment ID: Research Partner Identity Protected Location: MarketingExperiments Research Library Test Protocol Number: TP4067
Research Notes:
Experiment: Background
#webclinic
Experiment: Medical Condition Keywords
Medical Condition Keywords: • Keyword 1 • Keyword 2 • Keyword 3 • Keyword 4 • Keyword 5 • Keyword 6
Ad Abbreviated Med. Conditions: • Thyroid pain • Sciatic nerve • Blood Pres. • Kidney cysts • Colonic pain • Lumbar pain
*Keywords have been randomly selected and optimized to protect partner identity
#webclinic
Experiment: Treatments
Treatment Design:
• Template test in 6 condition-
based ad groups
• Headline is specific to
[condition] ad group
• Variable is description line one
• Value corresponds to content
approach
[Condition] Sufferer? Free access to kidney pain resources from the experts in kidney health. [Display URL]/[condition]
T1: Body Part Pain Resources
#webclinic
Experiment: Treatments
[Condition] Sufferer? Compare available treatments, from the experts in [specialty] health. [Display URL]/[condition]/[condition]
T2: Treatment Options
Treatment Design:
• Template test in 6 condition-
based ad groups
• Headline is specific to
[condition] ad group
• Variable is description line one
• Value corresponds to content
approach
#webclinic
Experiment: Treatments
[Condition] Sufferer? Learn about the causes & solutions, from the experts in [specialty] health. [Display URL]/[condition]
T3: Causes and Solutions*
Treatment Design:
• Template test in 6 condition-
based ad groups
• Headline is specific to
[condition] ad group
• Variable is description line one
• Value corresponds to content
approach
*Audience member submission
#webclinic
Experiment: Treatments
[Condition] Sufferer? How to recognize the symptoms, from the experts in [specialty] health. [Display URL]/[condition]
T4: Recognize Symptoms
Treatment Design:
• Template test in 6 condition-
based ad groups
• Headline is specific to
[condition] ad group
• Variable is description line one
• Value corresponds to content
approach
#webclinic
Experiment: Side-by-side
[Condition] Sufferer? Free access to specific pain resources from the experts in [specialty] health. Display URL/[condition]
T1: specialty Pain Resources [Condition] Sufferer?
Compare available treatments, from the experts in [specialty] health. Display URL/[condition]
T2: Treatment Options
[Condition] Sufferer? Learn about the causes & solutions, from the experts in [specialty] health. Display URL/[condition]
T3: Causes and Solutions* [Condition] Sufferer? How to recognize the symptoms, from the experts in [specialty] health. Display URL/[condition]
T4: Recognize Symptoms
*Amy’s submission
#webclinic
Version CTR Rel. diff Stat. Conf
[Specialty] Pain Resources 0.15% - -
Treatment Options 0.17% 14.1%
Causes & Solutions* 0.26% 71.8%
Symptoms 0.30% 99.7%
99.7% Increase in clickthrough rate ‘Symptoms’ generated 99.7% more clicks than the ‘[Specialty] Pain Resources’
What You Need to Understand: The ‘Symptoms’ content approach is most appealing to visitors making condition-based searches.
*Audience submission
Experiment: Results
99.9%
99.9%
65.8%
#webclinic
14
The Ideal Customer The Marketer
The Message
No Thanks
Time
Energy
Budget
? How can I know that my campaigns will be successful before I invest resources?
How can we predict success?
#webclinic
Today we are going to learn 3 principles to help us predict the success of our campaigns.
#webclinic
PRINCIPLE #1: All you need to know is enough to predict
#webclinic
Key Principles
PRINCIPLE #1:
Customer research can generate an overwhelming amount of data. We can cut through it with this simple maxim:
All we need to know is enough to predict customer behavior.
How can we predict success?
#webclinic
Knowing enough to predict
Our researchers call this knowledge the “Customer Theory.” Customer Theory Definition: The Customer Theory is an understanding of the customer that enables us to predict their total response to a given offer.
#webclinic
Knowing enough to predict
[Condition] Sufferer? Free access to [specialty] pain resources from the experts in [specialty] health. [Display URL]/[condition]
T1: [Specialty] Pain Resources [Condition] Sufferer? Compare available treatments, from the experts in [specialty] health. [Display URL]/[condition]
T2: Treatment Options
[Condition] Sufferer? Learn about the causes & solutions, from the experts in [specialty] health. [Display URL]/[condition]
T3: Causes and Solutions* [Condition] Sufferer? How to recognize the symptoms, from the experts in [specialty] health. [Display URL]/[condition]
T4: Recognize Symptoms
*Amy’s submission
? What can we predict based on these results?
#webclinic
Experiment #2: Background
Background: Medical provider specializing in treating chronic pain. Goal: To plan a content marketing strategy based on which approach generates more appeal in condition-based searchers. Primary Research Question: Which content approach will achieve a higher clickthrough rate? Approach: A/B Multifactor Split Test
Experiment ID: Research Partner Identity Protected Location: MarketingExperiments Research Library Test Protocol Number: TP4068
Research Notes:
#webclinic
Admin Note: Reference Test Protocol
From Test Protocol
#webclinic
General Hypothesis: “Based on what we learned from the previous content approach test, if we use a symptom content approach while matching the control's specificity to each ad group, we can achieve a higher click-through rate.” -Quoted from Test Protocol
Experiment #2: Control Keyword 1 Treatment Read about our xx-minute treatment for keyword pain relief. [Display URL]/Keyword_1
Keyword 2 Treatment Keyword pain? You Have Options. Ask Our Experts About [procedure]. [Display URL]
Keyword 3 Treatment Relieve [keyword] Pain with Our Unrivaled xx-Minute Procedure. [Display URL]/Keyword 3
Degenerative Keyword Disease Relieve Degenerative [keyword] Pain w/ Our Unrivaled Revolutionary Program [Display URL]/Degenerative
Keyword 4 Treatment New 53-Minute Procedure, Reclaim Your Life From [keyword] Pain. [Display URL]/Keyword
Relief From Keyword Pain Read about our 30-minute treatment for total freedom from [keyword] pain! [Display URL]
Control Ads
#webclinic
Experiment #2: Treatment 1 Keyword 1 Symptoms Read about our XX-minute treatment For keyword pain relief. [Display URL]/keyword
Keyword 2 Symptoms Keyword 2? You Have Options. Ask Our Experts About [Procedure™]. [Display URL]
Keyword 3 Symptoms Relieve [keyword] Pain with Our Unrivaled XX-Minute Procedure. [Display URL]/keyword
Keyword 4 Symptoms New xx-Minute Procedure, Reclaim Your Life From [keyword] Pain. [Display URL]/keyword
Keyword 5 Symtpoms Read about our XX-minute treatment for total freedom from [keyword] pain! [Display URL]
Treatment 1 Ads
How treatment will test hypothesis: “If treatment 1 wins, we will learn that the symptom content approach is most effective only when used in the headline.” -Quoted from Test Protocol
#webclinic
How treatment will test hypothesis: “If treatment 2 wins, we will learn that the symptom content approach is most effective when used in the description and when the description is specific to the ad group.” -Quoted from Test Protocol
Experiment #2: Treatment 2 Keyword 1 Treatment Read about our xx-minute treatment for keyword pain relief. [Display URL]/keyword
Keyword 2 Treatment Read about keyword Symptoms and see how our experts can help! [Display URL]
Keyword 3 Treatment Read about keyword Symptoms and see how our experts can help! [Display URL]/keyword
Degenerative Keyword Disease Read how our keyword experts can help relieve Degenerative keyword Symptoms! [Display URL]/Degenerative
Keyword 4 Treatment Read about keyword Symptoms and see how our experts can help! [Display URL]/keyword
Relief From Keyword Pain Read about keyword Symptoms and see how our experts can help! [Display URL]
Treatment 2 Ads
#webclinic
Experiment #2: Treatment 3
How treatment will test hypothesis: “If treatment 3 wins, we will learn that the symptom content approach is most effective when used in BOTH the headline and description and when the description is specific to the ad group.” -Quoted from Test Protocol
Keyword 1 Symptoms Read about our xx-minute treatment for keyword pain relief. [Display URL]/keyword
Keyword 2 Symptoms Read about keyword Symptoms and see how our experts can help! [Display URL]
Keyword 3 Symptoms Read about keyword Symptoms and see how our experts can help! [Display URL]/ /keyword
Keyword 3 Symptoms Read about keyword Symptoms and see how our experts can help! [Display URL]/ /keyword
Keyword 4 Symptoms Read about keyword Symptoms and see how our experts can help! [Display URL]
Treatment 3 Ads
#webclinic
Experiment #2: Results
Version CTR Rel. diff Stat. Conf
Control 0.28% - -
T1- Headline Change 0.26% -5.8%
T2- Description Change 0.21% -26.5%
T3- Both Change 0.39% 40.6%
40% Increase in clickthrough rate Adding ‘Symptoms’ to BOTH headline and description produced a 40% increase
What You Need to Understand: By applying the insight from the previous test and inserting ‘symptoms’ into both the headline and description, the team was able to create more successful treatments across all ad groups.
99.9%
99.9%
60.6%
#webclinic
27
Knowing enough to predict
Which? Why? What?
Research Question Theory Question
Value Proposition Question: If I am your ideal customer, why should I buy from you rather than your competitors?
Ideal Customer
Behavior reflects identity
#webclinic
28
201% 2% 29%
Test results are interpreted and second test was created based on
the analyst’s observations
Again, test results are interpreted and the next
round of testing is started for this page
Test is again interpreted and transferrable principles are applied to other offer pages
Knowing enough to predict Customer Theory applied
#webclinic
29
451%
302%
603% 257% 28%
• The discoveries and insights about customer motivation from the three prior tests were applied to other landing pages and used to optimize PPC campaigns.
• The purposeful effort to identify and selectively apply these transferrable insights led to widespread optimization gains.
Knowing enough to predict Customer Theory applied
#webclinic
PRINCIPLE #2: The Internet is a laboratory
#webclinic
Key Principles
PRINCIPLE #1: Customer research can generate an overwhelming amount of data. All we need to know is enough to predict customer behavior. PRINCIPLE #2: The Internet is not just another marketing channel, it is the primary laboratory for conducting customer research.
How can we predict success?
#webclinic
32
The Internet as a lab
The Internet has become the most efficient means of gathering business intelligence BEFORE a major online (or offline) campaign.
Level 1
Level 2
Level 3
The Decision Resolution
“To know what people really think, pay regard to what they do, rather than what they say.” – Descartes (source unknown)
#webclinic
33
The Internet as a lab Traditional marketing process
The Traditional Marketing Process
Business Leader Creates
Product
Agency delivers creative
Marketing Implements
Online marketing has simply been ADDED to the process.
#webclinic
34
The Behavior-Based Marketing Process
Business Leader Creates
Product
Agency delivers creative
Marketing Implements
If we can interrupt the process and test the message, then we can drive results across the business.
The Internet as a lab Behavioral-based marketing process
#webclinic
35
The Internet as a lab
548%
SEO Landing Page
Example
141%
Homepage
Lower investment channels
Higher investment channels
Offline Campaigns
*TV image by Brylle
#webclinic
PRINCIPLE #3: Negative tests can generate positive results
#webclinic
Key Principles
PRINCIPLE #1: Customer research can generate an overwhelming amount of data. All we need to know is enough to predict customer behavior. PRINCIPLE #2: The Internet is not just another marketing channel, it is the primary laboratory for conducting customer research. PRINCIPLE #3: The goal of a test is not (simply) to get a lift but rather to get a learning. Even a negative test can translate into a positive result.
How can we predict success?
#webclinic
38
Background: Company is a publisher of electronic marketing information and offers related services. Goal: Increase registrations for a free email newsletter. Primary Research Question: Which sign-up page will yield the highest conversion rate? Approach: A/B/C Multivariate test involving changes in headline, credibility indicators, and images according to optimization best practices.
Experiment ID: (Protected) Location: MarketingExperiments Research Library Test Protocol Number: #TP1092
Research Notes:
Experiment #3: Background
#webclinic
39
Control
Experiment #2: Control
? What are some problems you see with the control?
#webclinic
40
Treatment 1
Treatment 2 Experiment #3: Treatments
#webclinic
41
• Clearer headline emphasizes the value proposition.
• “Featured Clients” list emphasizes value and reduces anxiety.
• Bolded key terms make body copy easier to read and scan.
• Form fields are reduced to the bare minimum.
• Costumer testimonials reduce anxiety.
• Anti-spam seal reduces anxiety.
Treatment 1
Experiment #3: Treatment 1
#webclinic
42
• Headline is stronger and more quantitative to emphasize the value proposition.
• Added more testimonials.
• Rearranged page to give better site flow.
• Added personal feel with images and hand-written signature.
• “Tell me where to send…” language used.
Treatment 2
Experiment #3: Treatment 2
#webclinic
43
Treatment 1 Treatment 2 Control
Experiment #3: Side-by-side
Which of these treatments had the highest conversion rate?
#webclinic
44
Vs.
Control Treatment 1
Conversion Rate = 14.26% Conversion Rate = 6.74%
53%
Experiment #3: Treatment 1 results
#webclinic
45
Vs.
Control Treatment 2
Conversion Rate = 14.26% Conversion Rate = 6.84%
52%
Experiment #3: Treatment 2 results
#webclinic
46
Version CR Rel. diff Stat. Conf
Original 14.26% - -
Treatment 1 6.74% -52.7%
Treatment 2 6.84% -52.0%
53% Decrease in conversion The lowest performing treatment decreased conversion by 53%
What You Need to Understand: Through the negative lift in this test, we were able to learn that the customer was already sold by the time they landed on this page.
95%
95%
Experiment #3: Results
#webclinic
47
• Common landing page best practices failed to improve conversion on this original page.
• Began testing removal of elements from the page to match visitor motivation levels.
• If adding elements to increase the value proposition decreased conversion, maybe the traffic to this page was already highly motivated?
Control
Experiment #4: Control
#webclinic
48
• Much of the copy on this page is removed, leaving simple form submission fields.
• No real selling points are included in this design.
Treatment
Experiment #4: Treatment
#webclinic
49
Experiment #4: Side-by-side
Control
Treatment
#webclinic
50
Version CR Rel. diff Stat. Conf
Original 12.09% - -
Treatment 21.54% 78.12%
78% Increase in conversion The treatment increased conversion by 78.12%
95%
Experiment #4: Results
#webclinic
51
What You Need to Understand: Without the “negative lifts” that challenged almost every part of our normal optimization methodology, it would have taken much longer to achieve the eventual 78% increase.
X X
Test Cycle: Results
53% 52% 78%
#webclinic
52
What You Need to Understand: Without the “negative lifts” that challenged almost every part of our normal optimization methodology, it would have taken much longer to achieve the eventual 78% increase.
Test Cycle: Results
53% 52% 78%
#webclinic
53
Summary: Putting it all together
Key Principles
PRINCIPLE #1: Customer research can generate an overwhelming amount of data. All we need to know is enough to predict customer behavior. PRINCIPLE #2: The Internet is not just another marketing channel, it is the primary laboratory for conducting customer research. PRINCIPLE #3: The goal of a test is not (simply) to get a lift, but rather to get a learning. Even a negative lift can translate into a positive result.
#webclinic
54
Optimization Summit 2012 in Denver
#webclinic
55
Audience Questions
#webclinic
56
Question 1
How do we formulate a proper research question?
?
#webclinic
57
Question 2
How do you balance SEO with Value Proposition?
?
#webclinic
58
Question 3
Is it possible to have the same landing page for organic and PPC traffic?
?
#webclinic
59
Question 4
Can you talk about application of customer theory to engaging charitable donations?
?
#webclinic
60
Question 5
How does Value Proposition apply to sales?
?
#webclinic
61
http://whiterimage.com/
Submission 1
Primary Channel: Search Target Audience: Females, 18-40 Page Objective: To read more about a product.
#webclinic
62 http://www.appup.com
Submission 2
Primary Channel: Advertising via display, search, email Target Audience: Consumers who use PCs Page Objective: Click download to get more information and then download the free software with another click.
#webclinic
63
Submission 3 Primary Channel: Organic keywords- home lighting; ceiling lights; table lamps; We just updated the page and lost ranking. Target Audience: Home owners 40+ and contractors. Page Objective: The next action should be to search, and once they search they should buy. We plan to have a sign-up box on the search page so users can enter their email to receive special offers.
http://kewlighting.co.uk/
#webclinic
64
Submission 4
http://www.wwf.org.za/donate
Primary Channel: Homepage (organic, referral, and email) Target Audience: Existing WWF donors (contacts), non-donors (leads) on our database and new leads that can be converted to new donors to WWF-SA Page Objective: Click to another page with more detail/next steps.
#webclinic
65
Submission 5 Primary Channel: Display advertising (Sympatico, Facebook, Olive Network, Casale, Cineplex) and paid search Target Audience: Future students (main target grade 12 grads -18-24, but extending as wide as to 35; mainly in Ontario and western Quebec) Page Objective: To click “book a tour” or “request a publication”
http://www.algonquincollege.com/prospective/
#webclinic
MECLABS conducts rigorous experiments in the new science of optimization. We apply our discoveries to help leaders optimize the financial performance of their sales and marketing programs. Learn more about how you may be a fit for a MECLABS research partnership: • Select Research Partnership
Opportunities on the post-webinar survey
• Contact us directly
[email protected] 1-877-635-0565
MECLABS Research Partnership Opportunities
x