Make it Personal: Accelerating Your Personalization Game Plan
Download as PPTX, PDF0 likes603 views
The document discusses strategies for enhancing personalization in business through a variety of methods, including audience discovery and adaptive audience strategies. It emphasizes the potential monetary value of effective personalization and provides examples from successful campaigns, such as Crate & Barrel, to illustrate the benefits. The session aims to equip attendees with tangible ways to implement a strong personalization strategy and accelerate its effectiveness.
Make it Personal: Accelerating Your Personalization Game Plan
- 1. Congratulations to our Outperform Award Winners! Most Dramatic Business Impact Most Customer- Obsessed Company Culture Most Transformative Innovation Most Inspiring Social Impact
- 2. 2 Sessions today will be recorded and will be available after Opticon Join the conversation on Twitter at #Opticon19 We want to hear from you! Give feedback and rate sessions on the mobile app
- 3. Anuj Lal Sr. Manager, Global Strategy Consulting - Optimizely Ryan McGredy Sr. Manager, Solutions Engineering - Optimizely Make it Personal: Accelerating Your Personalization Game Plan Opticon 2019
- 4. Personalization is hard. Cliché I know..
- 5. 5 Tales from the field 1. Audience creation & definition is difficult ○ “Which audiences should we personalize for?” 2. Campaign ideation is intimidating ○ “We have entire teams working on this, where should we even begin?” 3. Soliciting input from other parts of the organization can be unwieldy ○ “How do we get more people involved w/o triggering governance problems?” 3 “worries” w/ Personalization Strategy
- 6. 6 Tales from the field 1. Audience creation & definition is difficult ○ “Which audiences should we personalize for?” 2. Campaign ideation is intimidating ○ “We have entire teams working on this, where should we even begin?” 3. Soliciting input from other parts of the organization can be unwieldy ○ “How do we get more people involved w/o triggering governance problems?” 3 “worries” w/ Personalization Strategy Why not let machine learning help you develop valuable audiences? Why not think of recommendations as your first personalization campaign? Why not use re-usable shared code to help scale & permissions to help govern
- 7. Personalization is hard. Cliché I know.. ..but it can be a lot easier to get started than you think.
- 8. We want you to leave this session knowing two things: 1. The tenants of a strong personalization strategy 2. Tangible & easy ways to accelerate that strategy (real examples)
- 9. 9 Agenda 1. Why? Revenue drivers 2. Who? Audience discovery 3. Pro-Tip: Adaptive Audiences 4. What? Campaign ideation 5. Pro-Tip: Recommendations 6. Where? Opportunity analysis 7. Pro-Tip: Extensions
- 10. 10 Agenda 1. Why? Revenue drivers 2. Who? Audience discovery 3. Pro-Tip: Adaptive Audiences 4. What? Campaign ideation 5. Pro-Tip: Recommendations 6. Where? Opportunity analysis 7. Pro-Tip: Extensions
- 11. 11 “In the online world, businesses have the opportunity to develop very deep relationships with customers...so that you can use that individualized knowledge of the customer to accelerate their discovery process.” Jeff Bezos “Accurately predicting the movies Netflix members will love is a key component of our service” Neil Hunt, Chief Product Officer The value of personalization is in the trillions of dollars Personalization at scale has the potential to create $1.7 trillion to $3 trillion in new value. Based on projections by McKinsey Global Institute
- 12. 12 “In the online world, businesses have the opportunity to develop very deep relationships with customers...so that you can use that individualized knowledge of the customer to accelerate their discovery process.” Jeff Bezos “Accurately predicting the movies Netflix members will love is a key component of our service” Neil Hunt, Chief Product Officer The value of personalization is in the trillions of dollars Personalization at scale has the potential to create $1.7 trillion to $3 trillion in new value. Based on projections by McKinsey Global Institute
- 13. “ ”- Anuj Lal Let’s make this real.
- 14. “ ”Optimizely Strategy Consulting A personalization campaign is an experiment that is trying to move a KPI or a set of KPIs for a GROUP of audiences.
- 15. 15 Goal Orientation (B2B/Lead-Generation Example) Identify the KPIs that you aim to influence through personalization and define the metrics that will measure them Increase Customer Lifetime Value (LTV) Increase Customer Acquisition for Product Type (Close Rate %) Increase Quote/Demo Form Submission Rate (Submission %) Define Goals That Will Validate Hypothesis About User Behavior
- 16. 16 Goal Tree – Revenue (B2B) RetentionAcquisition
- 17. 17 Goal Tree - Retail and E-Commerce Revenue: $2.0m Revenue Per Visitor: $1.00 Average Order Value: $50.00 Average quantity: 2.00 Average Price per Unit: $25.00 Conversion Rate: 2.00% Add to Cart Cart Checkout Rate Visitors: 2.0m User acquisition: 1.5m User retention 33% UpsellsCross-sell ... generally
- 18. 18 Goal Tree - Retail and E-Commerce Revenue: $2.0m Revenue Per Visitor: $1.00 Average Order Value: $50.00 Average quantity: 2.00 Average Price per Unit: $25.00 Conversion Rate: 2.00% Add to Cart Cart Checkout Rate Visitors: 2.0m User acquisition: 1.5m User retention 33% UpsellsCross-sell ... generally
- 19. 19 Goal Tree - Retail and E-Commerce Revenue: $2.0m Revenue Per Visitor: $1.00 Average Order Value: $50.00 Average quantity: 2.00 Average Price per Unit: $25.00 Conversion Rate: 2.00% Add to Cart Cart Checkout Rate Visitors: 2.0m User acquisition: 1.5m User retention 33% UpsellsCross-sell ... generally
- 20. 20 Reasons for Goal Tree: Comprehensive: ● Visually analyzes the entire site’s opportunity to weigh opportunity cost Alignment: ● Allows teams to explicitly align testing goals to business goals; keeps ideas tightly focused Productive Brainstorming: ● Strategic foundation for creative brainstorming *Orange = KPIs that can be influenced by digital optimization via Optimizely KPIs that can be influenced using upsell or cross-sell strategies Goal Tree - Retail and E-Commerce Revenue: $2.0m Revenue Per Visitor: $1.00 Average Order Value: $50.00 Average quantity: 2.00 Average Price per Unit: $25.00 Conversion Rate: 2.00% Add to Cart Cart Checkout Rate Visitors: 2.0m User acquisition: 1.5m User retention 33%
- 21. 21 Key Learnings 1. A personalization campaign is an experiment where you are trying to move a KPI or a set of KPIs for a GROUP of audiences 2. A goal tree is an effective way to evaluate the purpose(s) of your personalization strategy – the “why?”
- 22. 22 Agenda 1. Why? Revenue drivers 2. Who? Audience discovery 3. Pro-Tip: Adaptive Audiences 4. What? Campaign ideation 5. Pro-Tip: Recommendations 6. Where? Opportunity analysis 7. Pro-Tip: Extensions
- 23. 23 A framework we like to use at Optimizely to categorize different types of audiences is: Demographic, Contextual, & Behavioral
- 24. 24 Personalized Audiences Framework for categorization Demographic: Who are they? (Can be gleaned using first party or third party data) • Store Preference • Vertical (I-I) • Gender (I-I) • Member/Non-Member • Demographic (I-I) • CRM Data (I-I) Contextual: What can we infer from the context of the browser? • Marketing Campaign • Referral Source • Device • New/Returning • Time of Day • Browser Behavioral: What is their intent? • Viewed category of content • Entered certain city/zipcode • Past Browse Behavior • Clicked ‘Plans’ twice but didn’t convert • Browsed at a particular time of day
- 25. 25 Actionable & Big: Do Both Pursue a variety of both to maximize reach/quality Obviously Actionable Large Need for Creativity Granular Visitor Cohort; New, Returning, Active, Loyal Browsed Twice; Product Category Large Geos; Coastal Urban, State, Key Cities Past Purchasers Second Priority
- 26. 26 Current challenge: If we wanted to pick an audience to start with, we’d have to guess at exactly what the right granularity is:
- 27. 27
- 28. 28 Agenda 1. Why? Revenue drivers 2. Who? Audience discovery 3. Pro-Tip: Adaptive Audiences 4. What? Campaign ideation 5. Pro-Tip: Recommendations 6. Where? Opportunity analysis 7. Pro-Tip: Extensions
- 29. 29 Simply define your audience using keywords
- 30. How it Works! Interest Score Our estimate of a visitor’s interest in the keyword-defined persona. Predicted Intent The new Audience Condition that evaluates a visitor’s Interest Score. Key Ideas Compute Score Optimizely compares the visitor’s profile to the keywords to generate an Interest Score. Audience Builder Customers supply keywords in plain English that define the persona. Browsing History Visitors browse product, article, or landing pages and we construct a profile based on the text content.
- 31. Browsing History: Tags Tags Page Tags can capture any text elements, search terms, invisible meta tags, or data layers. One time setup! Visitor Profile Most recent 10 page views are used to construct profile of this visitor. Key Ideas
- 32. 32
- 33. 33 Demo 1. Adaptive Audiences
- 34. 34 Key Learnings 1. Traditionally, organizations suffer from “analysis paralysis” trying to define audiences. 2. Using machine learning to define your audiences for you is a quick way to start experimenting and iterating.
- 35. 35 Crate & Barrel Success Story ● Running a number of major campaigns using Adaptive Audiences today ○ Furniture vs housewares ○ Parents of Kids vs Infants ● In all instances, bounce rates dropped. ● Conversion rates (purchase rate) improved substantially. Across all three, it was more than a 20+% improvement. ● Revenue per visitor was a double digit improvement as well.
- 36. Using Optimizely’s new personalization offering, we have been able to increase conversion rates on the homepage by 20%+ while reducing bounce rates by up to 30%. Within one month of using Adaptive Audiences → Crate & Barrel had 16 different audiences they are using. CHRISTINE GARVEY Senior Manager, Personalization & Optimization
- 37. 37 Agenda 1. Why? Revenue drivers 2. Who? Audience discovery 3. Pro-Tip: Adaptive Audiences 4. What? Campaign ideation 5. Pro-Tip: Recommendations 6. Where? Opportunity analysis 7. Pro-Tip: Extensions
- 38. A/B Testing Finding the AVERAGE best experience OVERALL for everyone
- 39. Personalization Exploring the BEST experience for each audience
- 44. Category Affinity Create relevant offers and product placements across key points of your experience based on affinity shown for product or content types from prior browsing behavior Customer Experience as a Competitive Advantage Tailor your experiences to target your customers and test how each experience resonates Symmetric Messaging & Session Cohorting Ensure that customers receive messaging that is consistent and applicable based on search terms, creating a cohesive customer journey Localizations Localize web experiences based on customer location Behaviorally Triggered Campaigns Based on previous behavior, target specific customers with personalized experiences in real time ABM Create relationships with your most important customers across your experience through 1-to-1 messaging and value propositions based on which company or industry they are in Loyalty Tiers Push your loyal customers to the most important groups by offering unique benefits based on current tier of loyalty as well as ensure consistent messaging for a seamless experience Abandoned Cart & Purchase History Personalize based on individuals purchase history and cart abandonment Key Personalization Use-Cases
- 45. 45 How do we think about our audiences & exactly what we should do for each one of them?
- 46. 46 Reminder: Audience Framework Framework for categorization Demographic: Who are they? (Can be gleaned using first party or third party data) • Store Preference • Vertical (I-I) • Gender (I-I) • Member/Non-Member • Demographic (I-I) • CRM Data (I-I) Contextual: What can we infer from the context of the browser? • Marketing Campaign • Referral Source • Device • New/Returning • Time of Day • Browser Behavioral: What is their intent? • Viewed category of content • Entered certain city/zipcode • Past Browse Behavior • Clicked ‘Plans’ twice but didn’t convert • Browsed at a particular time of day
- 47. 47 Consider where users are in their journey with you Demographic Who are they? (First party or third party data) • Geo, Vertical, Gender, Member Non-Member, Demographic, CRM Data, Customer Tier, or Loyalty Behavioral What is their intent? ● Viewed category of content, Entered certain city/zipcode, Past Browse Behavior, Viewed lead form, Scrolled 50% down the page, but didn’t engage Contextual What can we infer from the context of the browser? • Marketing Campaign, Referral Source, Device, New/Returning, Time of Day, Browser NEW VISITORS RETURNING VISITORS KNOWN VISITORS Awareness Consideration Service LoyaltyPurchase
- 48. 48 Who are they? What is important to them? What are they looking for? What can you do? New Visitor ● Looking for info/education ● Someone who can sympathize or relate to her individual problem or need ● Acknowledge her “newness” to brand, ● Aim to get her contact information ● Educate her on offering & return visit Return Visitors ● Ability to continue her shopping session from where she last left ● Incorporate her past site behavior position offering ● Acquire her contact information A Repeat/Loyal Customer Engaged) ● Recognition and benefits as a customer ● Use rewards or awards to him/her for being a good customer ● Surface your ‘premier’ loyalty programs ● Encourage him/her to share or promote the brand Connecting your users to a “mission” Connecting the ‘Who?’ to the ‘What?’
- 49. 49 Most difficult part in practice: Building content to map to each of these personas or audiences
- 50. 50 Most difficult part in practice: Building content to map to each of these personas or audiences Why not think about adding a recommendations algorithm for your first personalization campaign?
- 51. 51 Recommendations Use-Cases for All Verticals Retail, B2B/Lead-Generation, Media 1. Retail/eCommerce Website ○ In the checkout funnel, show accessories that complement the items a visitor is purchasing ○ On product detail pages, show alternative items that are related to the product a visitor is browsing ○ Highlight crowd favorites on the homepage 2. B2B/Lead-Generation Website ○ Show visitors whitepapers, infographics, blog posts, and other content based on their browsing behaviors ○ Suggest knowledge base articles or community posts to reduce support call volume 3. Media Website ○ Show visitors articles, sectionals, videos, and other content based on their browsing behaviors
- 52. 52 Agenda 1. Why? Revenue drivers 2. Who? Audience discovery 3. Pro-Tip: Adaptive Audiences 4. What? Campaign ideation 5. Pro-Tip: Recommendations 6. Where? Opportunity analysis 7. Pro-Tip: Extensions
- 53. INTERNAL USE Browsing History: Tags Tags Page Tags can capture any text elements, search terms, invisible meta tags, or data layers. One time setup! Visitor Profile Most recent 10 page views are used to construct profile of this visitor. Key Ideas
- 54. 54 Demo 1. Adaptive Audiences 2. Optimizely Recommendations
- 55. Recommendations V1 CHALLENGES ● Could not be implemented by the customer w/o assist from services + ongoing maintenance from services. ● Was not built into the Optimizely Web interface. ● Therefore, often wasn’t incorporated into our Personalization pitch
- 56. 56
- 57. 57 You don’t have to create new copy or content, it already exists on your site today
- 58. 58 Key Learnings 1. A personalization campaign is all about connecting users to a MISSION. 2. Recommendations are a great way to 1) show value early, 2) back into implementation, & 3) force up-front discovery – so you can quickly learn from your first personalization campaign
- 59. 59 Recommendation Use-Cases for All Verticals Retail, B2B/Lead-Generation, Media 1. Retail/eCommerce Website ○ In the checkout funnel, show accessories that complement the items a visitor is purchasing ○ On product detail pages, show alternative items that are related to the product a visitor is browsing ○ Highlight crowd favorites on the homepage 2. B2B/Lead-Generation Website ○ Show visitors whitepapers, infographics, blog posts, and other content based on their browsing behaviors ○ Suggest knowledge base articles or community posts to reduce support call volume 3. Media Website ○ Show visitors articles, sectionals, videos, and other content based on their browsing behaviors
- 60. 60 Optimizely Recommendations Co-browse Co-buy Popular Recently viewed Collaborative Filtering Items that other visitors who viewed this item also viewed (ideal for PDP) Items that other visitors who purchased this item also bought (ideal for PDP, Cart) Items most frequently viewed or bought across the entire catalog (ideal for Homepage) Items a visitor had previously browsed (“pick up where you left off”) Unique recommendations for each visitor, based on their unique behavior on your site in the past
- 61. 61 Recommendations Examples by Page – Retail example Page Sample Algorithms that we may use: Homepage Popular products Ratings-based Affinity Landing Page Popular products Similar products Similar offers Category Page Popular products Similar products Zero results & 404 Pages Product Page Similar products Bought / viewed this… Frequently bought together Cart Page Accessories Frequently bought together
- 62. 62 Top Retailer: Recommendation Strategies & Custom Algorithms Example Recommendation Type Reasoning Recommended for you Most common but effective (add name to personalize) Frequently bought together Increase AOV Recently Viewed Introduce shoppers to new items Browsing History Access Relocate items viewed earlier Related items you’ve viewed Help with ideas Customer who bought this also bought this Social proof and peer-generated recommendations Newer Version Alert viewers of products that have been updates Related to previous purchase Personalize on what they might want / simplify re- purchase Best Selling Items Indirect social proof and way of adding confidence Product Bundles Frequently bought together bundles to increase AOV Rating Highest rating items as social proof
- 63. 63 Agenda 1. Why? Revenue drivers 2. Who? Audience discovery 3. Pro-Tip: Adaptive Audiences 4. What? Campaign ideation 5. Pro-Tip: Recommendations 6. Where? Opportunity analysis 7. Pro-Tip: Extensions
- 64. 64 Ideating Solutions Ideate often at the start of a problem!
- 65. 65 Lack of Focus on Initial Campaigns Pitfalls to Avoid Focus on Top ROI Pages to Start, Progress Incrementally Homepage Landing Page Product Page Checkout 1st iteration 2nd iteration 3rd iteration Launch
- 66. 66 Lack of Focus on Initial Campaigns Pitfalls to Avoid Focus on Top ROI Pages to Start, Progress Incrementally Homepage Landing Page Product Page Checkout 1st iteration 2nd iteration 3rd iteration Launch ...Not Everything at Once Homepage Landing Page Product Page Checkout Launch ● Successful practitioners start with a tight focus on addressing top audiences and pages. ● Pages should be high traffic + high value. ● Audiences relatively high reach.
- 67. 67 Data-Driven Approach: WHO on the site? ● Customers: 1,000 Visitors x 50% Exit Rate = 500 Visitors ● Return Visitors: 5,000 Visitors x 25% Exit Rate = 1,350 Visitors ● New Visitors: 10,000 Visitors x 60% Exit Rate = 6,000 Visitors (audience opportunity) WHERE on the site? ● App Homepage: 30,000 Visitors x 50% Exit Rate = 15,000 Visitors ● Desktop Homepage: 15,000 Visitors x 25% Exit Rate = 3,750 Visitors ● Mobile Homepage: 60,000 Visitors x 60% Exit Rate = 36,000 Visitors (impactful opportunity) Where Can You Have The Largest Impact?? Easiest way to figure out who/where to start is to look at your exit rate on each part of your flow and take it one step further and segment by “Mobile/Desktop”; “New/Returning”; “SEM/Direct”
- 68. 68 Previously, how customers of ours maintained governance of their personalization program is through “shared code.”
- 69. Two main challenges heard over & over: 1. “We know where we want to personalize and for whom (ie. Homepage for ad-campaign audiences), but we don’t want to have to re-create (copy/paste/edit) the code every campaign.” 2. “It’s hard to collaborate & expand/grow personalization across the organization without having the same key stakeholders involved in every decision."
- 70. 70 Agenda 1. Why? Revenue drivers 2. Who? Audience discovery 3. Pro-Tip: Adaptive Audiences 4. What? Campaign ideation 5. Pro-Tip: Recommendations 6. Where? Opportunity analysis 7. Pro-Tip: Extensions
- 71. 71 Demo 1. Adaptive Audiences 2. Optimizely Recommendations 3. Optimizely Extensions
- 72. 72 Optimizely Extensions 1. Extensions let developers create reusable features for non-technical team members 2. Marketers can add custom features to experiments without ongoing developer support 3. Developers can scale by enabling marketers and other business users to do more on their own Enabling scale
- 73. 73 Additionally, see how Banana Republic (GAP Inc.) expanded access to their personalization program via extensions:
- 78. 78 Key Learnings 1. Determining “where” you want to personalize is just as important as deciding “what” you want to personalize use a data-driven approach. 2. Scaling personalization across the entire organization can seem daunting, but with the right tools, it can turn that challenge into a key differentiator for your business by reinforcing a culture of experimentation
- 79. 79 Tales from the field IBM Story
- 80. 80 Tales from the field IBM Story Rollout of Testing & Personalization Center of Excellence ○ 300+ experiments globally Focus on Personalization w/ Extensions ○ 10x 2017 volume Audience integrations
- 81. 81 Agenda 1. Why? Revenue drivers 2. Who? Audience discovery 3. Pro-Tip: Adaptive Audiences 4. What? Campaign ideation 5. Pro-Tip: Recommendations 6. Where? Opportunity analysis 7. Pro-Tip: Extensions
- 82. 82 Takeaways 1. Audience creation & definition is difficult, why not let machine learning assist you in finding your most valuable audiences? 2. Campaign ideation can be prohibitive, but a recommendations engine is another form of personalization, and can be a great way to start, show value early and accelerate initial learnings. Make sure your idea is rooted in the “why?” with a measurable KPI. 3. Expanding access & controls for personalization can be unwieldy, but utilizing the right tools (ie. Optimizely extensions), you can actually use personalization strategy as a way to build excitement internally.
- 83. Thank You! & any questions? Anuj Lal: anuj@optimizely.com Sr. Manager – Global Strategy Consulting Ryan McGredy: ryan.mcgredy@optimizely.com Sr. Manager – Sales Engineering
- 84. Head to the Expo Hall for refreshments and to say to our sponsors Up next...hear how IBM accelerated digital optimization to achieve digital excellence
- 85. APPENDIX:
- 86. Behavioral Audiences Demographic: Who are they? (Can be gleaned using first party or third party data) • Geo • Vertical (I-I) • Gender (I-I) • Member/Non-Member • Demographic (I-I) • CRM Data (I-I) Contextual: What can we infer from the context of the browser? • Marketing Campaign • Referral Source • Device • New/Returning • Time of Day • Browser Behavioral: What is their intent? • Viewed category of content • Entered certain city/zipcode • Past Browse Behavior • Clicked ‘Plans’ twice but didn’t convert • Browsed at a particular time of day
- 87. Things to think about: ● What are your largest audiences? What is the ‘reach’ of these audiences? (% of that page’s visitors) ● What is the unique value of these audiences (or user personas)? ● How feasible is it to target these audiences? Media Sample Audiences Local Sports Fan Fan of specific sports teams, by viewing more than X pieces of related content Topical Affinity Visitors that have viewed a type of content more frequently than others (i.e. Politics, Sports, Crossword, Editorial, Cars, Travel etc.) Hit Paywall Previously Visitors that have visited paywall page and not subscribed International Visitors from an international location Article Meter Visitors at level X in a meter (i.e., read 5 of 7 articles) Video Watchers Visitors that have watched videos more than some threshold (i.e., more than the average by 30%, or more than X times in last 30 days) Attention Span Visitors that on average reach less than 25%, or more than 75%, article completion e-Commerce Sample Audiences Price Based Visitors who have completed a purchase at least once in the last X days where price >=< $X Category Affinity Visitors that have viewed a product of category X at least Y times in the last Z days Abandoned Cart Visitors who triggered Add to Cart at least once but did not Purchase Window Shoppers Visitors that have viewed X number of product pages in the last X days and have not added to cart Repeat Purchasers Visitors that have completed a purchase more than X times in the last X days Coupon Shoppers Visitors that have completed a purchase while using a coupon code Product Filter Visitors that have refined a search, product, or category landing page filter (e.g. size, color, type of product) Behavioral Audiences - Examples
- 88. 88 Goal Tree – Revenue (B2B)
- 89. 89 Goal Tree – Revenue (B2B) Reasons for Goal Tree: Comprehensive: • Visually analyzes the entire site’s opportunity, and help avoid opportunity cost Alignment: • Allows teams to explicitly align testing goals to business goals; keeps ideas tightly focused Productive Brainstorming: • Strategic foundation for creative brainstorming, helping to garner a volume of ideas for prioritization *Orange = Areas available for digital optimization via Optimizely