Radical Transformations How American Express reimagined the customer experience with a data-driven approach to optimization Sa Sabrin ina Pas asin ini Sele Selena Bl Blue Senior Marketing Manager Manager of Editorial Content American Express MECLABS
Session speaker Sabrina Pasin ini Senior Marketing Manager American Express Headshot As Senior Manager of Digital Acquisition at American Express, Sabrina manages the delivery of innovative and high-profile Web projects that maximize the performance and user experience of the American Express Consumer Card website. This has included overseeing a new site redesign, ongoing test and learn campaigns, @s_e_pasini as well as an optimization roadmap to operationalize big data for real-time targeting. Pasini has been with American Express for more than five years and has held a variety of key roles in the digital acquisition group.
About American Exp xpress FORTUNE 500 RANKING: 90 107.2 130+ $33 million countries cards-in-force billon in annual 62,848 employees worldwide revenue
More than 20 credit cards
The Challe llenge: Transforming dig igit ital l exp xperiences Ditch Implement Deliver Rethink the educated strategic targeted and digital guesswork optimization personalized experience for testing roadmap experiences
Challe llenge: Transforming dig igit ital l exp xperiences How do you accomplish this? Implement Ditch Deliver Rethink the strategic educated targeted and digital optimization guesswork personalized experience roadmap for testing experiences
Testin ing progression Segment-based Expert Judgment A/B Testing Multivariate Testing Multivariate Testing
Step 1. . Exp xpert ju judgment
Step 1. . Exp xpert ju judgment
Testin ing progression Segment-based Expert Judgment A/B Testing Multivariate Testing Multivariate Testing
Step 2. . A/B /B testing Experiment Research Notes: Bac Background: Test the existing Card Member-focused homepage with treatments featuring three and four cards in the hero section. Ob Objective: To increase card application conversion rate. Prim rimary ry Res esearch Qu Question: Which page emphasis – specific popular cards or balanced selection across portfolios – will result in higher conversion rates? Tes est t Des Desig ign: 100% redirect from Current Cards Landing to 3 Cards Landing – 4% random, 5% default and 91% targeted 11
Step 2. . A/B /B testing Control
Step 2. . A/B /B testing Treatment 1 Control
Step 2. . A/B /B testing Treatment 1 Treatment 2 Control
Step 2. . A/B /B testing Treatment 1 Treatment 2 Control
Step 2. . A/B /B testing Treatment 1 Treatment 2 Control Resulted in No Clear Winner
Testin ing progression Segment-based Expert Judgment A/B Testing Multivariate Testing Multivariate Testing
Step 3. . Mult ltivariate testing Experiment Research Notes: Bac Background: Test for further improvement on Treatment 1 from the previous A/B test. Ob Objective: To increase card application conversion rate. Prim rimary ry Res esearch Qu Question: Will optimizing each of the key elements on the page drive engagement and increase conversion rate? Tes est t Des Desig ign: Full-factorial MVT with six elements for a total of 256 variations. With business exclusion applied, a a to total of of 144 144 experie iences ar are te tested. 18
Step 3. . Mult ltivariate testing Variable #1. Color scheme
Step 3. . Mult ltivariate testing Variable #2. Page navigation
Step 3. . Mult ltivariate testing Variable #3. Calls-to-action
Step 3. . Mult ltivariate testing Variable #4. Comparison tab
Step 3. . Mult ltivariate testing Variable #5. Lower tool section
Step 3. . Mult ltivariate testing Variable #6. Tool in tab
Step 3. . Mult ltivariate testing Which combination won?
Step 3. . Mult ltivariate testing Control Winning Combination
Step 3. . Mult ltivariate testing Control Winning Combination 29% 29% In Increase in Conversion Rate
Ranki king of MVT treatment result lts 144 1 ? Winning Control Combination
Ranki king of MVT treatment result lts 144 119 1 Winning Control Combination
The next xt frontier Experiences One-size-fits-all customized to in interactions mic icrosegments
Testin ing progression Segment-based Expert Judgment A/B Testing Multivariate Testing Multivariate Testing
Bala lancing Act: Busin iness prio iority vs. . customer needs Lower Combined Value Higher Value
Segmentation process Customer Prospect COOKIE: Data Data Site 1048379205 Activity Offsite Model Media Scores 3 rd Step 1. Party Track and Identify Prospect Traits
Segmentation process Customer Prospect COOKIE: Data Data Site 1048379205 Activity Offsite Model Media Scores 3 rd Step 1. Step 2. Party Track and Identify Combine Traits Prospect Traits into Segments
Segmentation process Customer Prospect COOKIE: Data Data Site 1048379205 Activity Offsite Model Media Scores 3 rd Step 1. Step 2. Step 3. Party Track and Identify Combine Traits Displays Custom Prospect Traits into Segments Content in Real Time
Step 4. . Segment-based mult ltiv ivaria iate testin ing Experiment Research Notes: Bac Background: Test personalized homepages for a customer-centric experience. Ob Objective: To increase response and conversion rate. Prim rimary ry Res esearch Qu Question: Will targeting cards to profiles that may have an affinity for these products increase clickthrough and conversion rates? Tes est t Des Desig ign: Target created four campaigns: 100% Redirect from Current Cards Landing to 3 Cards Landing – 4% random, 5% default and 91% targeted. 36
Step 4. . Segment-based mult ltiv ivaria iate testin ing Feature the winning product Display next- highest valued products Match “Member Story” to featured product
Step 4. . Segment-based mult ltiv ivaria iate testin ing Control Treatment
Step 4. . Segment-based mult ltiv ivaria iate testin ing Control Treatment 28% 28% In Increase in Conversion Rate
Segmentation example St Step 1. 1. Cookie assigned
Segmentation example Prospect: Rob obert Assig igned Co Cookie ie: St Step 1. 1. 87 8762 6209 09479 795 Cookie assigned
Segmentation example St Step 2. 2. Traits tracked before Card homepage arrival St Step 1. 1. Cookie assigned
Segmentation example Step 2. St 2. Interest: Travel Traits tracked before Card homepage arrival 3 rd Party: Travel • • Geo: Atlanta • DFA: Delta 300x250 St Step 1. 1. • DFA: Delta 160x600 Cookie assigned
Segmentation example Interest: Travel St Step 2. 2. Traits tracked before Card homepage arrival St Step 1. 1. St Step 3. 3. Cookie assigned Customized Card homepage presented
Segmentation example Interest: Travel St Step 2. 2. Traits tracked before Card homepage arrival St Step 1. 1. St Step 3. 3. Cookie assigned Customized Card homepage presented
Segmentation example Interest: Travel St Step 2. 2. St Step 4. 4. Traits tracked before Prospect continues Card homepage arrival to search website Step 1. St 1. Step 3. St 3. Cookie assigned Customized Card homepage presented
Segmentation example Interest: Travel St Step 2. 2. Step 4. St 4. Traits tracked before Prospect continues Card homepage arrival to search website St Step 1. 1. St Step 3. 3. Cookie assigned Customized Card homepage presented
Segmentation example Interest: Travel St Step 2. 2. St Step 4. 4. Traits tracked before Prospect continues Card homepage arrival to search website Step 1. St 1. Step 5. St 5. St Step 3. 3. Cookie assigned Newly acquired traits Customized Card assigned to prospect homepage presented
Segmentation example Interest: Travel Step 2. St 2. Step 4. St 4. Intent: Cashback Traits tracked before Prospect continues Card homepage arrival to search website • View All Cards: Cashback • BCE Learn More Page • BCE Credit QuickInfo St Step 1. 1. St Step 5. 5. Step 3. St 3. Cookie assigned Newly acquired traits Customized Card • View All Cards: Credit assigned to prospect homepage presented
Segmentation example Interest: Travel Step 2. St 2. St Step 4. 4. St Step 6. 6. Traits tracked before Prospect continues Updated Cards Cards homepage arrival to search website homepage presented Step 1. St 1. St Step 5. 5. Step 3. St 3. Cookie assigned Newly acquired traits Customized Cards assigned to prospect homepage presented Intent: Cashback
Segmentation example Interest: Travel St Step 2. 2. Step 4. St 4. Step 6. St 6. Traits tracked before Prospect continues Updated Cards Cards homepage arrival to search website homepage presented Step 1. St 1. St Step 5. 5. Step 3. St 3. Cookie assigned Newly acquired traits Customized Cards assigned to prospect homepage presented Intent: Cashback
Recommend
More recommend