Your SlideShare is downloading. ×
  • Like

Thanks for flagging this SlideShare!

Oops! An error has occurred.

×

Now you can save presentations on your phone or tablet

Available for both IPhone and Android

Text the download link to your phone

Standard text messaging rates apply

Applied Marketing Analytics - Paramore University 4.16.13

  • 1,336 views
Published

 

  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Be the first to comment
    Be the first to like this
No Downloads

Views

Total Views
1,336
On SlideShare
0
From Embeds
0
Number of Embeds
1

Actions

Shares
Downloads
11
Comments
0
Likes
0

Embeds 0

No embeds

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
    No notes for slide

Transcript

  • 1. Applied Marketing AnalyticsParamore University 2013
  • 2. About MarketingSherpa• Established 13 years ago• Independent research and publishing organization focused on the marketing community• From 2008 to 2012: • 36,980 companies and marketers surveyed (cumulative) • 3831 charts and tables • 4,847 pages of insights and analysis • 1,857 pages of research supported tactics and recommended actions
  • 3. About MECLABS• Parent organization of MarketingSherpa and other research brands• World’s largest independent research lab focused exclusively on marketing and sales • More than 15 years of research partnership with our clients • 1,300 experiments • Over 1 billion emails tested • 10,000 landing pages tested • 5 million telephone calls • 500,000 decision maker conversations
  • 4. Questions? #ParamoreU
  • 5. Let’s talk about… Messaging
  • 6. How would your team improve the messaging?
  • 7. What’s going into our marketing decisions… Instead of analytics data to make marketing decisions, we rely on: Gut instincts Our own Intuition Historical spending The status quo Testing What works Brand awareness The need to be known Brand perception The need to be loved Purchase intention What’s been decided Willingness to What’s recommended recommend HIPPO (Highest paid What’s commanded person’s opinion) OtherSource: ©2012 MarketingSherpa Marketing Analytics Benchmark SurveyMethodology: Fielded November 2012, N= 682
  • 8. How it changes based on role Instead of analytics data to make marketing decisions, we rely on: The Troops My WhatNonmanagement marketing personnel Need to be known Gut Works The Middle Mgr Be My What What we Marketing manager or supervisor Known Gut Works already did The Chief Need to My What What we Chief Marketing Officer or Senior Executive be known Gut Works Already did Brand awareness Gut instincts Testing Historical Spending Source: ©2012 MarketingSherpa Marketing Analytics Benchmark Survey Methodology: Fielded November 2012, N= 682
  • 9. How would your team improve the messaging? Improving the message should be no big deal, right?
  • 10. Case Study: Background Case Study ID: Pier 1 Imports Protocol Number: A-TP1002Research Notes:• Background: B2C home products company with a significant online and retail presence• Goal: To increase the click-through rate from the email to the landing page• Primary research question: Which email design will generate the most click-throughs?• Approach: A/B/C/D split test (variable cluster)
  • 11. Case Study: The Challenge
  • 12. Case Study: Campaign Designs Team A Original
  • 13. Case Study: Campaign Designs Team B Original
  • 14. Case Study: Campaign Designs Team C Original
  • 15. Case Study: Which team would you vote for? Team A Team B Team C Original
  • 16. Case Study: Results 52% Decrease in Clickthrough Team A’s design decreased clicks by 51.8% Email Designs CTR Rel. Diff. Original 36.70% - Team A 17.68% -51.83% Team B 29.91% -18.50% Team C 24.07% -34.41%
  • 17. Case Study: Background Case Study ID: Protected Protocol Number: TP1457Research Notes:• Background: B2C, B2B tax services brand with both online and offline products• Goal: To increase online product purchases• Primary research question: Which e-commerce product detail page will produce more purchases of the product being showcased?• Approach: A/B split test (variable cluster)
  • 18. Case Study: Challenge About the original: • Standard e-commerce style product page • Call to action and product imagery above the fold • Supporting information tabbed and organized
  • 19. Case Study: Campaign About the new design: • Completely broken mold with heavy design elements • Call to action BELOW the fold • Product imagery totally eliminated
  • 20. Case Study: Results 83% Increase in Purchases The new design increased purchases by 83.79% Product Page Version Product Conv. Rate Control 8.27% Double Control 9.93% Treatment 1 18.25% Relative Difference: 83.79%
  • 21. And the moral of the study is… 83% 52% 19% 34%
  • 22. And the moral of the study is… “To know what people really think, pay regard to what they do, rather than what they say.” - René Descartes 83% 52% 19% 34%
  • 23. Key Point• There are no expert marketers, there are just experienced marketers and expert testers
  • 24. Key Point• There are no expert marketers, there are just experienced marketers and expert testers• There is a critical element in the testing and optimization process that you can access now and use without 10 hours of teaching and 10 weeks of systems changes
  • 25. To run a good test… Research Question Treatments Metrics Validity (Hypothesis)
  • 26. You need a premise… Research Question Treatments Metrics Validity (Hypothesis)
  • 27. You need a premise… “ Which sources of information do you ” actively use to better understand your prospects and customers? Research Question (Hypothesis)Source: ©2012 MarketingSherpa Marketing Analytics Benchmark SurveyMethodology: Fielded November 2012, N= 775
  • 28. You need a premise… “ Which sources of information do you ” actively use to better understand your prospects and customers? Research The backbone of expert testing Question (Hypothesis) is analytics examinationSource: ©2012 MarketingSherpa Marketing Analytics Benchmark SurveyMethodology: Fielded November 2012, N= 775
  • 29. Case Study: ObservationExperiments designed with the strategic use of analytics examination on averageproduced more valuable results compared to those that did not • 2010 Homepage Tests Test 2 (questionnaire style, -63%) vs. Test 1 (images, 0%) Test 5 (reducing process friction, +22%) vs. Test 4 (flash treatment, -80%) • 2010 Online Product Page Tests Test 3 (product qualifiers emphasis, +9%) vs. Test 1 (page arrangement, 0%) • 2010 PPC Landing Page Tests Retail 2 (reducing process friction, +533%) vs. Retail 1 (above/below fold, 8%) Shared 3 (reducing process friction, +34%) vs. Shared 1 (page arrangement, 0%) • 2010 Banner Ad Landing Page Tests Test 2 (product specific value, +28%) vs. Test 1 (page alignment, -62%)
  • 30. How helpful are analytics, anyway? Faster growth of our business (40%) Competitive advantage (37%) Common basis for decision making (47%) Better marketing Better utilization of resources (43%) messages (67%) Better risk management (19%) More accurate and precise response to More complete understanding of market customer needs (44%) conditions and trends (40%) Predicting customer behavior (38%) Complete understanding of the marketing purchase cycle (37%)Source: ©2012 MarketingSherpa Marketing Analytics Benchmark SurveyMethodology: Fielded November 2012, N= 682
  • 31. How helpful are analytics, anyway? Faster growth of our business (40%) Competitive advantage (37%) Common basis for decision making (47%) Better don’t need a test toutilization You marketing Better of resources (43%) messages (67%) leverage your analytics Better risk management (19%) More accurate and precise response to More complete understanding of market customer needs (44%) conditions and trends (40%) Predicting customer behavior (38%) Complete understanding of the marketing purchase cycle (37%)Source: ©2012 MarketingSherpa Marketing Analytics Benchmark SurveyMethodology: Fielded November 2012, N= 682
  • 32. How much analytics does your org collect? 14% Vast quantities17%Limited 25% 40% Average 79% Average or more Significant
  • 33. How much analytics does your org collect? 14% Vast quantitiesLimited 25% 40% Average 79% Analytics are available in the 17% majority of organizations Average or more Significant
  • 34. So what’s the problem, then? 52% 19% 34%
  • 35. Are you able to leverage your org’s analytics? 9% 2%access No No tools,6% skills Rarely 37% Routinely & Effectively 46%Occasionally The only exception is Marketing Agencies or Consultancies
  • 36. Maybe it’s just too much to process…
  • 37. Maybe you’ve never been the data type… If only this calculator had excel built into it…
  • 38. Maybe you feel like its all or nothing…
  • 39. Key Point• There are no expert marketers, there are just experienced marketers and expert testers• There is a critical aspect of testing and optimization that you can access without 10 hours of teaching and 10 weeks of systems changes• To take advantage of today’s data, all you need to do is see it with a different perspective• Today, we’re going to discuss four key principles that will help you see today’s data with new eyes
  • 40. POINT 1: Stop focusing on the “how many”
  • 41. Stop focusing on the “how many”1. The goal of any kind of customer research is to enable the marketer to anticipate customer response to a particular message or approach.2. Therefore, the primary usefulness of examining analytics, or even testing, is not in answering “how many?” but rather in answering, “why so?”
  • 42. Case Study: Background Case Study ID: Protected Protocol Number: TP4067Research Notes: Background: A medical provider specializing in treating chronic pain. They are the sole providers of an innovative procedure and pain management plan. Goal: To plan a content marketing strategy based on the copy focus that generates the most appeal in condition-based searchers. Primary research question: Which subject matter focus (copy) will achieve a higher click-through rate? Approach: A/B Multifactor Split Test
  • 43. Case Study: Campaign [Condition] Sufferer? [Condition] Sufferer? Free access to [part]pain resources Compare available treatments, from the experts in [part] health. from the experts in [part] health. Company.com/[condition] Company.com/[condition] [Condition] Sufferer? [Condition] Sufferer? Learn about the causes & solutions, How to recognize the symptoms, from the experts in [part] health. from the experts in [part] health. Company.com/[condition] Company.com/[condition]
  • 44. Case Study: Results[Condition] Sufferer? [Condition] Sufferer?Free access to [part] pain resources Compare available treatments,from the experts in [part] health. from the experts in [part] health.Company.com/[condition] Company.com/[condition][Condition] Sufferer? [Condition] Sufferer?Learn about the causes & solutions, How to recognize the symptoms,from the experts in [part] health. from the experts in [part] health.Company.com/[condition] Company.com/[condition] 73% more 99% more* What you need to understand: Customers will more likely engage with this company trying to understand the problem as opposed to immediately looking for a solution
  • 45. What we have discovered1. The goal of any kind of customer research is to enable the marketer to anticipate customer response to a particular message or approach.2. Therefore, the primary usefulness of examining analytics, or even testing, is not in answering “how many?” but rather in answering, “why so?”3. Ultimately, analytics from observation and experimentation can enable the marketer to see cognitive trails left by the visitor’s mind.4. These cognitive trails give us clues for how they will respond, even when tracking isn’t always available in another medium
  • 46. Remember this page? About this page: • Call to action • Hero image • copy
  • 47. Not this…• Analytics shouldn’t be points in a game that you can only watch from a distance
  • 48. But this… Marketing Intuition So-so Opinion Research Ok Analytics should be about Behavioral gathering businessWinner Observation & intelligence BEFORE a Experimentation major online (or offline) campaign.
  • 49. Example: Hidden insights in web analytics Web Analytic Cognitive Clues Time on page Are visitors engaged with the content? Are they confused with the process? What are visitors interested in? Click tracking Are they confused with something we are saying? Do we have the wrong focus? Bounce rate Are there too many distractions? Is there too much (or little) information? Segment-level data What motivates individual visitor types? Where are the deeper optimization opportunities? Form event tracking What form fields cause anxiety or confusion? How much friction will your visitor put up with? Traffic patterns Who is coming and where are they coming from? Can we be more relevant to the visitor?
  • 50. Example: Hidden insights in tests conducted Customer Behavior Customer Theory Which headline will What does my customer generate a higher want the most? response? Which testimonial will What makes my customer generate the most especially anxious? response? Which call to action will What is my customer comfortable generate a higher with at this stage of the buying response? cycle?
  • 51. POINT 2: Simplify your perspective of analyticsso it gets used
  • 52. Need a moment?
  • 53. Need a moment? Massive information presented equally assaults the mind
  • 54. Simplify your perspective1. When you focus on the “why so”, all The Who analytics can be organized into four Source categories The What2. Each analytics category reveals a different Result aspect of the visitor’s story The Where and When3. Different perspectives (categories) can be Amount combined to create a single understanding of the person that encounters our The Why messaging Nature
  • 55. Like a Bento Box… The Where Amount Source The Who and when The Why Nature Results The What
  • 56. The who (source)• If you want to know where people are Common Metrics coming from Referrers• These analytics often give clues to Search Terms the motivation of your visitors and allow Countries and Languages you to understand how many different types of visitors are viewing the same Top Landing Pages message • i.e. the kind of experience or content the visitor is expecting.
  • 57. The what (result)• If you want to know what people Common Metrics do once they get to a page Conversions/Purchases• These analytics are like mile markers Clicks on your highway to conversion Next Pages • What markers must people take Downloads to get to the end of the road? • At what markers do people get off the highway and get off track?
  • 58. The where and when (amount)• If you want to know the Common Metrics amount of in each part of your Pageviews, Visits process (including the purchase category) Visitors Impressions Total Revenue
  • 59. The why (nature) • If you want to know what people Common Metrics are experiencing (or selecting) while viewing your messaging Event, eye tracking Clicks/page • Use this group of analytics to find Time on page big problems/disconnects people may be experiencing in your Transaction Details messaging or experience.
  • 60. Best sources of “why so” information Medium: Organic/SEO Where Amount Source Who and when % of Total Traffic (64%) Top Website Referrers (47%) Keyword clicks (45%) Unique Search terms (46%) Inbound links (36%) Branded vs. Non-branded (25%) Nature Results Keywords triggering search Keyword Rankings (63%) (51%) CTR (47%) Keyword movement Term Conv. Rate (38%) (37%) ROI (33%) Why What
  • 61. Best sources of “why so” information Medium: Video Marketing Where Amount Source Who and when Placements on share sites Video SEO (21%) (21%) Embeds on non-video Most viewed video sharing sites (14%) segments (33%) Nature Results Comments, Likes, +1s Conv Rate (27%) (33%) Video Ad Clicks (23%) Video ratings (20%) Play rate (40%) Play-through rate (21%) ROI (15%) Video shares (29%) Why What
  • 62. Better sources of “why so” information Medium: Email Where Amount Source Who and when Clicks per email (55%) List Size (48%) Deliv. rate (55%) Inbox placement rate (16%) Nature Results Clicks per link in email (49%) Open Rate (78%) Complaint Rate (25%) CTR (78%) Unsub rate (65%) Social Sharing rate (21%) Conv Rate (55%) ROI (28%) Why What
  • 63. Better sources of “why so” information Medium: Social Marketing Where Amount Source Who and when Views (55%) Top Influencers (26%) RSS (23%) Nature Results Brand Sentiment (23%) Social Reach (61%) Engagement/Post/Tw (33%) Traffic from social (49%) Sales (23%) ROI (20%) Why Conv Rate (27%) What
  • 64. Better sources of “why so” information Medium: Paid Search Where Amount Source Who and when Clicks (66%) Avg Ad Pos.(41%) Impr. Share (28%) Nature Results Quality Score (36%) CTR (66%) CPCl (65%) CPConv(44%) Conv Rate (54%) CPLd(43%) ROI (40%) Profit per click (18%) Why What Profit per Impr.(10%)
  • 65. Average sources of “why so” information Medium: Display Advertising Where Amount Source Who and when Clicks (61%) Reach (27%) Frequency (27%) Impr. Share (21%) Nature Results CTR (62%) CPM (33%) CPConv(40%) Conv Rate (45%) ROI (33%) Lost Impr. Share (6%) Why What
  • 66. Average sources of “why so” information Medium: Content Marketing Where Amount Source Who and when Views (55%) RSS (23%) Nature Results Comments/Post (29%) Leads (48%) Downloads (41%) Conv. Rate (40%) ROI (23%) Likes, Tweets, +1s Why Shares (45%) What
  • 67. POINT 3: Start with the minimum (not max)effective dose
  • 68. As for the rest of us…• Not everyone has a “Data” on board to do the most daring of analyses
  • 69. Forget the analysis paralysis• You don’t need to throw everything (including the kitchen sink) at something to get messaging the performs positively
  • 70. Case Study: Background Case Study ID: RegOnline Homepage Test Protocol Number: TP1428Research Notes: Background: Event management software company that lets users create online registration forms and event websites to manage their events. Goal: To increase number of completed leads on home page. Primary research question: Which process will yield a higher conversion rate? Approach: A/B Multifactor Split Test
  • 71. Case Study: Challenge• Already tested and optimized by the local design/dev team over the past year
  • 72. Case Study: Campaign• Two easily accessible pieces of data (nav summary, time on page) From here… To here… To here… ~2 minutes ~2 minutes <1 minute And back again…
  • 73. Case Study: Campaign• With that data, the team created a messaging experience that FORCED visitors to 1 Essential Product read and see certain and Company piece of information Overview before others and Details WITHOUT negatively effecting SEO 2 Pricing Info
  • 74. Case Study: Results 89.8% increase in conversion The treatment generated 89.8% more completed leads Versions Conversion Rate Rel. diff Control 0.3% - Treatment 0.5% 89.8% * What you need to understand: The team was able to achieve a substantial lift by understanding how the customer responds to information when presented in a certain sequence
  • 75. Forget the analysis paralysis• You don’t need to throw everything (including the kitchen sink) at something to get messaging the performs positively• To get the effect of an analysis for a minimal amount of effort, transform your analytics bento box and into a pyramid
  • 76. Analytics Pyramid Not all analytics categories are created equal…
  • 77. What we have discovered…• The analytics (or clues) that are more telling to create effective messaging are at the bottom and the metrics that need less are at the top Like the old food Pyramid…
  • 78. Analytics usage for message creation Which of the following do you routinely use to create different message types? Keyword SOURCE Website activity NATUREPerformance or previous message RESULT New vs. Returning visitor SOURCE Purchase history NATURE Referral channel SOURCE Location SOURCE Date of last visit RESULT Device SOURCE Comprehensive testing strategy OtherSource: ©2012 MarketingSherpa Marketing Analytics Benchmark SurveyMethodology: Fielded November 2012, N= 602
  • 79. Analytics Pyramid• The key: The more you combine and utilize source and nature based analytics, the better performance potential you’ll have with your messaging
  • 80. Analytics Pyramid Examples• Example 1: Messaging that doesn’t take Source analytics into consideration is a message that has no clear target.• Example 2: If you see a great Amount of visitors that show a common Result (like leaving the critical path in a certain direction), then you may have found a major disconnect with the messaging • but you still need more to know what causes it
  • 81. POINT 4: Stop focusing on tools and startfocusing on visibility
  • 82. Tools seem to be top of mind… In which areas are you planning additional investments? Please select all that apply. Purchase of analytics tools/platforms/software Training Data integration with other applications Staffing of in-house personnel/analysts Hiring of external analysts/consultants OtherSource: ©2012 MarketingSherpa Marketing Analytics Benchmark SurveyMethodology: Fielded November 2012, N= 233
  • 83. Though few come close to using the same set• 17 Experts, Digital Marketing (not just SEO)• 40+ tools referenced• Very few consistent results
  • 84. With no clear decision on free vs. paid Are you satisfied with the PRECISION of your analytics systems? Paid Tools Competitive intelligence tools CRM systems Live chat tracking tools Attribution management software Social media monitoring tools Offline call management and tracking systems Marketing automation software SEO management tools Email marketing analytics software PPC bid management tools Web-integrated call management and tracking… Web (clickstream) analytics tools Dissatisfied Neutral SatisfiedSource: ©2012 MarketingSherpa Marketing Analytics Benchmark SurveyMethodology: Fielded November 2012, N≤79
  • 85. With no clear decision on free vs. paid Are you satisfied with the PRECISION of your analytics systems? Free Tools Competitive intelligence tools CRM systems Live chat tracking tools Attribution management software Social media monitoring tools Offline call management and tracking systems Marketing automation software SEO management tools Email marketing analytics software PPC bid management tools Web-integrated call management and tracking systems Web (clickstream) analytics tools Dissatisfied Neutral SatisfiedSource: ©2012 MarketingSherpa Marketing Analytics Benchmark SurveyMethodology: Fielded November 2012, N≤409
  • 86. What marketers really want from Oz…If I only had __________, my marketing efforts would be substantially more effective Advanced customer behavior analysis (37%) Complete quantitative understanding of the entire marketing and purchase cycle (36%) Predictive analytics (33%) Competitive trends insights (30%) Integration of online and offline data (29%) Customer sentiment/Voice of customer (27%) Visibility info pipeline (funnel) performance (26%) Cross-channel view of results (24%) Lifetime value measurement (24%) Social media and Web 2.0 measurement (24%) A/B and multivariate testing (21%) Real-time reporting (18%) Custom report creation (16%)
  • 87. What marketers really want from Oz…If I only had __________, my marketing efforts would be substantially more effective Advanced customer behavior analysis (37%) Complete quantitative understanding of the entire marketing and purchase cycle (36%) Predictive analytics (33%) Competitive trends insights (30%) Integration of online and offline data (29%) Customer sentiment/Voice of customer (27%) Visibility info pipeline (funnel) performance (26%) Cross-channel view of results (24%) Lifetime value measurement (24%) Social media and Web 2.0 measurement (24%) A/B and multivariate testing (21%) Real-time reporting (18%) Custom report creation (16%)
  • 88. What marketers want from Oz…If I only had __________, my marketing efforts would be substantially more effective Advanced customer behavior analysis (37%) Complete quantitative understanding of the entire marketing and purchase cycle (36%) Predictive analytics (33%) Competitive trends insights (30%) Integration of online and offline data (29%) Customer sentiment/Voice of customer (27%) Visibility Visibility info pipeline (funnel) performance (26%) Cross-channel view of results (24%) Lifetime value measurement (24%) Social media and Web 2.0 measurement (24%) A/B and multivariate testing (21%) Real-time reporting (18%) Custom report creation (16%)
  • 89. Case Study: Background Case Study ID: Protected Protocol Number: TP1305Research Notes: Background: A company that sells retail and wholesale collector items primarily online Goal: To increase conversion rate, specifically from new customers. Primary research question: Which version of second step in the conversion funnel will produce the highest conversion rate? Approach: A/B variable cluster split test
  • 90. Case Study: Challenge• Their checkout’s messaging came to queue, and Google Analytics was only showing 1 page for a 6 page process Sample of code change• When doing the research, they discovered that getting the details would require some extensive code changes and risks to the current tracking • Legacy • Simulated page tracking
  • 91. Case Study: Campaign Revenue drop offs• Seeing the challenge, the team used their creativity to do two alternatives that would grant them the sight they needed: • Install a limited amount of code from a new tool (mitigate risk, faster turnaround) • Utilize existing data already being captured by other systems• The result was two sources of information that pointed to one particular messaging problem for new customers
  • 92. Case Study: Campaign • Can you see a potentialControl red flag (from the customer’s point of view) in the messaging?
  • 93. Case Study: Campaign • The emphasis on theTreatment detailed terms and conditions was refocused to the satisfaction guarantee that was already in place New focal point
  • 94. Case Study: Results $548,000 Increase in profit per year The new checkout page increased conversion by 4.51% Design Conversion Rate Control 82.33% Treatment 86.04% Relative Difference 4.51% * What you need to understand: While the conversion increase is small, optimizing messaging in this specific step in the sales funnel resulted in a projected $500,000+ increase in profit per year.
  • 95. Visibility is a big problem with big payoff potential Email Content Marketing Social Organic/SEO Display PPC/SEM
  • 96. Visibility is a big problem with big payoff potential Email Content Marketing Social Organic/SEO Display PPC/SEM
  • 97. Visibility isn’t just plugging holes either… Which sources of information do you actively use to better understand your prospects and customers? Please select all that apply. Website activity Customer service feedback Industry blogs, professional journals Transaction data Social media conversations Demographic data Third-party market research Competitive benchmarking Reviews and rankings Focus groups/Customer surveys Brand performance analysis OtherSource: ©2012 MarketingSherpa Marketing Analytics Benchmark SurveyMethodology: Fielded November 2012, N= 775
  • 98. KEY PRINCIPLES
  • 99. Key Principles• There are no expert marketers, there are just experienced marketers and expert testers• Analytics examination is a fundamental aspect of testing that you can access without 10 hours of teaching and 10 weeks of systems changes• Take advantage of today’s analytics by changing your team’s perspective: • POINT 1: Stop focusing on the “how many”, start focusing on the “why so” • POINT 2: Adopt a simplified perspective of analytics to make it usable • POINT 3: Focus on the minimum (not maximum) effective dose • POINT 4: Stop focusing on tools and start focusing on visibility