SlideShare a Scribd company logo
1
Ahead of the curve
How 23andMe improved UX
with Performance Edge
+
Courtney Ball, Antonio Contreras,
and Whelan Boyd
2
Introductions
Courtney Ball
Sr. Manager, Web Marketing
Antonio Contreras
Sr. Web Developer
Whelan Boyd
Sr. Product Manager
● Customize the widgets on your page
to your preference
● This webinar is recorded and you will
receive the link with the slides in the
next few days
● We will have time for questions! Please
submit them in the Q&A box on the left
side of the screen
Housekeeping
4
Agenda
1. Why performance matters
2. How to think about measuring performance
3. How experimentation affects performance
4. Performance Edge
5. Case study: 23andMe
6. Q&A
5
Why
Performance
Matters
6
The business impact of speed
13% increase in sales
when page load time
was cut by half
Found that 53% of mobile
visits were abandoned
if page took > 3 seconds
to load
7% increase in
conversion when page
load time reduced by
850 ms
Source: https://developers.google.com/web/fundamentals/performance/why-performance-matters/
7
How to measure
performance
8
Measure impact to user experience
Source: Google’s Web Fundamentals
First Paint
(FP)
First Contentful
Paint (FCP)
First Meaningful
Paint (FMP)
Time to Interactive
(TTI)
9
Best practices for measuring
Free Online Tools
Synthetic Traffic
Easy to run
Actionable Insights
Real User Monitoring
Production Traffic
Segment by Device, Location, etc.
Robust Analytics
10
How website
experimentation
affects performance
11
Standard Web Experiments
1. Snippet size increases with
more experiments
2. For specific visitors, the
downloaded snippet includes all
project variations vs just active
variations
3. Larger snippet results in more
time to download and execute
4. Resource-constrained browsers
have to execute several tasks
Performance Challenges
for Web Experiments
Standard
CDN
Script
Experiment A
variation 1: CTA: Get Started
variation 2: CTA: Try it out
Experiment B
variation 1: CTA: Promo Code
variation 2: CTA: No Promo Code Task load:
URL Targeting
Audience
Targeting
Traffic Splitting
Apply Variations
Event Tracking
12
Introducing
Performance
Edge
13
Only returns
visitor-specific variations
Standard Web Experiments
Task load:
Experiment
Delivery
Network
Apply Variations
Event Tracking
Key experiment decisions
are moved to the Edge
Much smaller file to download
and execute in the browser
with less tasks to perform
Performance Edge Experiments
Standard
CDN
Script
Experiment A
variation 1: CTA: Get Started
variation 2: CTA: Try it out
Experiment B
variation 1: CTA: Promo Code
variation 2: CTA: No Promo Code
URL Targeting
Audience
Targeting
Traffic Splitting
Apply Variations
Event Tracking
CTA: Get Started
Script
URL Targeting
Audience
Traffic Splitting
14
Why is it so fast?
Less data
Far smaller file
loading over the wire
Reduce download time
Less code
Only code necessary
for active variations
Fraction of the execution
time in browser
Faster machines
Powerful Cloudflare Workers
execute expensive tasks
Located extremely
close to end visitor
15
Proper Measurement Make things smaller Performance Edge
● User-focused Metrics
● Synthetic and real traffic
● Lighter download
● Shorter execution
● Quick implementation
with any CDN
● Push logic to the Edge
Summary
16
23andMe
Case Study
17
23andMe: Intro
Courtney Ball
Sr. Manager, Web Marketing
Antonio Contreras
Sr. Web Developer
18
23andMe: Overview
Company size: 500+ employees
Location: Sunnyvale, California
Mission: Help people access, understand and benefit
from the human genome.
What we offer: 23andMe gives individuals easy and
affordable access to their genetic information, offering
a growing list of reports on traits, wellness, carrier
status and genetic ancestry.
Consumer Research Therapeutics
23andMe: Testing Overview
TV
AUDIO
DISPLAY
SOCIAL
Testing focus for 2020:
● Optimize kit mix.
● Optimize mobile experience.
● Optimize personalization
through audience targeting.
23andMe Customer Acquisition / Marketing Funnel
SEARCH
SITE
Consideration
Conversion
Awareness
20
23andMe: Approach to improving user experience
through a site redesign
Create new designs
(multiple options)
Consumer insights
conducts qualitative
research
Identify key areas of
opportunity through a
site audit
A/B test against
existing page via
Optimizely
Iterative testing
(if needed)
21
Test: Redesigned Product Page
Test Variant #1Control
Round 1
22
Test: Redesigned Product Page
Test Variant #1Control
Round 1
Winner
23
Test: Redesigned Product Page
Control Test Variant #1
Round 2
Test Variant #2
24
Test: Redesigned Product Page
Control Test Variant #1
Round 2
Test Variant #2: Winner
25
Why we needed performance edge as we scaled
A growing point of internal tension…
1. More testing = slower page speed
2. More emphasis placed on performance
Eagerness to try new tactics to improve site
performance ahead of the holiday season.
+
Frequency A/B
Tests
2016 2017 2018 2019 2020
Year
NumberofTestsLaunched
20
40
60
0
26
How we implemented
Performance Edge at 23andMe
1. We had an existing Worker on Cloudflare
that we were using to load our Optimizely
JavaScript
2. We updated the existing Worker for
Optimizely Edge and created a separate
Worker for the old Optimizely JavaScript
3. We used an integration environment to
test performance and root out errors
Cloudflare
CDN / Worker
Client / Browser 23andMe
Web
Server
27
Speed improvements with Performance Edge
Our performance statistics improved across the board:
Standard Snippet ⚡Optimizely Edge Snippet
First Contentful Paint Mobile: 2.7s
Desktop: 2.5s
Mobile: 1.2s
Desktop: 1.1s
First Meaningful Paint Mobile: 3.0s
Desktop: 2.7s
Mobile: 1.6s
Desktop: 1.5s
Time to Interactive Mobile: 6.7s
Desktop: 6.2s
Mobile: 4.4s
Desktop: 5.4s
Key Benefits:
dramatic performance improvements on mobile and we can continue scaling our testing program without slowing down our site!
28
Questions?
Ask them
in the webinar console

More Related Content

PPTX
Why Things Go Off the Rails and How to Prevent Product-Engineering Angst
PPTX
How We Do It: Proven Website Personalization Strategies
PDF
Failure is an Option: Scaling Resilient Feature Delivery
PDF
Optimizely's Vision for Product Development Teams
PDF
How FOX Tests Everything from Mobile, Web, to Living Room Devices
PPTX
Test Everything: TrustRadius Delivers Customer Value with Experimentation
PPTX
Experimentation Excellence: Mastering the basics
PPTX
How to Use Quant and Qual Feedback to Rapidly Improve Your Product
Why Things Go Off the Rails and How to Prevent Product-Engineering Angst
How We Do It: Proven Website Personalization Strategies
Failure is an Option: Scaling Resilient Feature Delivery
Optimizely's Vision for Product Development Teams
How FOX Tests Everything from Mobile, Web, to Living Room Devices
Test Everything: TrustRadius Delivers Customer Value with Experimentation
Experimentation Excellence: Mastering the basics
How to Use Quant and Qual Feedback to Rapidly Improve Your Product

What's hot (20)

PPTX
AMC Networks Experiments Faster on the Server Side
PPTX
How The Zebra Utilized Feature Experiments To Increase Carrier Card Engagemen...
PDF
[Webinar] Innovate Faster by Adopting The Modern Growth Stack
PDF
Atlassian's Mystique CLI, Minimizing the Experiment Development Cycle
PPTX
7 Habits of Highly Effective Personalisation Organisations | Optimizely ANZ W...
PPTX
Improve your content: The What, Why, Where and How about A/B Testing
PDF
How to feature flag and run experiments in iOS and Android
PDF
The Optimizely Experience Keynote by Matt Althauser - Optimizely Experience L...
PDF
The Future of Optimizely for Technical Teams
PDF
The Future of Building Good Products: Progressive Delivery and Experimentation
PDF
Optimizely Experience Customer Story - Atlassian
PPTX
Optimizely NYC Developer Meetup - Experimentation at Blue Apron
PDF
How Clorox Experiments Across Brands to Turn Visitors into Consumers
PDF
SplitMetrics answers burning questions on mobile A/B testing
PPTX
Zillow + Optimizely: Building the Bridge to $20 Billion Revenue
PDF
Optimizely Agent: Scaling Resilient Feature Delivery
PDF
Evolving Experimentation from CRO to Product Development
PDF
Apply A/B Testing with NGINX Routing Policy
PDF
Better Decisions with Experimentation
PDF
Experimentation at Blue Apron (webinar)
AMC Networks Experiments Faster on the Server Side
How The Zebra Utilized Feature Experiments To Increase Carrier Card Engagemen...
[Webinar] Innovate Faster by Adopting The Modern Growth Stack
Atlassian's Mystique CLI, Minimizing the Experiment Development Cycle
7 Habits of Highly Effective Personalisation Organisations | Optimizely ANZ W...
Improve your content: The What, Why, Where and How about A/B Testing
How to feature flag and run experiments in iOS and Android
The Optimizely Experience Keynote by Matt Althauser - Optimizely Experience L...
The Future of Optimizely for Technical Teams
The Future of Building Good Products: Progressive Delivery and Experimentation
Optimizely Experience Customer Story - Atlassian
Optimizely NYC Developer Meetup - Experimentation at Blue Apron
How Clorox Experiments Across Brands to Turn Visitors into Consumers
SplitMetrics answers burning questions on mobile A/B testing
Zillow + Optimizely: Building the Bridge to $20 Billion Revenue
Optimizely Agent: Scaling Resilient Feature Delivery
Evolving Experimentation from CRO to Product Development
Apply A/B Testing with NGINX Routing Policy
Better Decisions with Experimentation
Experimentation at Blue Apron (webinar)
Ad

Similar to Ahead of the Curve: How 23andMe Improved UX with Performance Edge (20)

PPTX
How BiggerPockets runs experiments faster with performance edge
PDF
Maximizing site speed with mercy corps
PDF
o2script(UK) - Top Site Execution_ Compelling Devices and Systems.pdf
PDF
Real User Monitoring: Getting Real Data from Real Users in the Real World - S...
PPTX
Supercharging Optimizely Performance by Moving Decisions to the Edge
PPTX
Rl Partner Webinar To Share
PPT
Gitte Ottosen - Agility and Process Maturity, Of Course They Mix!
PDF
Cutting-edge Performance Testing on eCommerce Websites
PPTX
Core web Vitals: Web Performance and Usability
PDF
Web.dev extended : What's new in Web [GDG Taichung]
PDF
Smashing Conference Short Talk on Sustainable Webdesign and coding practices
PPTX
Quality Jam: BDD, TDD and ATDD for the Enterprise
PPTX
New in Core Web Vitals 2024 - Boost User Experience and SEO - Macaw Digital.pptx
PDF
Designers Guide to Web Performance Yotta 2013
PDF
Web performance e-book
PDF
A Designer's Guide to Web Performance
PDF
2.1 Automation Nation: Keeping your Process Builders in Check
PPTX
Testing for performance
PDF
New in Core Web Vitals 2024 - Boost User Experience and SEO - Macaw Digital.pdf
PPTX
Continuous Performance Testing
How BiggerPockets runs experiments faster with performance edge
Maximizing site speed with mercy corps
o2script(UK) - Top Site Execution_ Compelling Devices and Systems.pdf
Real User Monitoring: Getting Real Data from Real Users in the Real World - S...
Supercharging Optimizely Performance by Moving Decisions to the Edge
Rl Partner Webinar To Share
Gitte Ottosen - Agility and Process Maturity, Of Course They Mix!
Cutting-edge Performance Testing on eCommerce Websites
Core web Vitals: Web Performance and Usability
Web.dev extended : What's new in Web [GDG Taichung]
Smashing Conference Short Talk on Sustainable Webdesign and coding practices
Quality Jam: BDD, TDD and ATDD for the Enterprise
New in Core Web Vitals 2024 - Boost User Experience and SEO - Macaw Digital.pptx
Designers Guide to Web Performance Yotta 2013
Web performance e-book
A Designer's Guide to Web Performance
2.1 Automation Nation: Keeping your Process Builders in Check
Testing for performance
New in Core Web Vitals 2024 - Boost User Experience and SEO - Macaw Digital.pdf
Continuous Performance Testing
Ad

More from Optimizely (20)

PDF
Clover Rings Up Digital Growth to Drive Experimentation
PPTX
Make Every Touchpoint Count: How to Drive Revenue in an Increasingly Online W...
PPTX
The Science of Getting Testing Right
PPTX
Autotrader Case Study: Migrating from Home-Grown Testing to Best-in-Class Too...
PPTX
Empowering Agents to Provide Service from Anywhere: Contact Centers in the Ti...
PPTX
Experimentation Everywhere: Create Exceptional Online Shopping Experiences an...
PDF
Building an Experiment Pipeline for GitHub’s New Free Team Offering
PDF
Overcoming the Challenges of Experimentation on a Service Oriented Architecture
PPTX
Making Your Hypothesis Work Harder to Inform Future Product Strategy
PPTX
Kick Your Assumptions: How Scholl's Test-Everything Culture Drives Revenue
PPTX
Experimentation through Clients' Eyes
PPTX
Shipping to Learn and Accelerate Growth with GitHub
PDF
The Future of Software Development
PPTX
Practical Use Case: How Dosh Uses Feature Experiments To Accelerate Mobile De...
PDF
Run High Impact Experimentation with High-quality Customer Discovery
PDF
Using Empathy to Build Custom Solutions at Scale
PPTX
How to find data insights that will drive a 10X impact
PPTX
Targeted Rollouts: How to Release Features to Multiple Audiences
PDF
Deploying Fearlessly in a Continuous World
PPTX
Detecting incorrectly implemented experiments
Clover Rings Up Digital Growth to Drive Experimentation
Make Every Touchpoint Count: How to Drive Revenue in an Increasingly Online W...
The Science of Getting Testing Right
Autotrader Case Study: Migrating from Home-Grown Testing to Best-in-Class Too...
Empowering Agents to Provide Service from Anywhere: Contact Centers in the Ti...
Experimentation Everywhere: Create Exceptional Online Shopping Experiences an...
Building an Experiment Pipeline for GitHub’s New Free Team Offering
Overcoming the Challenges of Experimentation on a Service Oriented Architecture
Making Your Hypothesis Work Harder to Inform Future Product Strategy
Kick Your Assumptions: How Scholl's Test-Everything Culture Drives Revenue
Experimentation through Clients' Eyes
Shipping to Learn and Accelerate Growth with GitHub
The Future of Software Development
Practical Use Case: How Dosh Uses Feature Experiments To Accelerate Mobile De...
Run High Impact Experimentation with High-quality Customer Discovery
Using Empathy to Build Custom Solutions at Scale
How to find data insights that will drive a 10X impact
Targeted Rollouts: How to Release Features to Multiple Audiences
Deploying Fearlessly in a Continuous World
Detecting incorrectly implemented experiments

Recently uploaded (20)

PDF
Spectral efficient network and resource selection model in 5G networks
PDF
Blue Purple Modern Animated Computer Science Presentation.pdf.pdf
PPTX
VMware vSphere Foundation How to Sell Presentation-Ver1.4-2-14-2024.pptx
PDF
Unlocking AI with Model Context Protocol (MCP)
PDF
How UI/UX Design Impacts User Retention in Mobile Apps.pdf
PPTX
KOM of Painting work and Equipment Insulation REV00 update 25-dec.pptx
PDF
Reach Out and Touch Someone: Haptics and Empathic Computing
PDF
MIND Revenue Release Quarter 2 2025 Press Release
PPTX
sap open course for s4hana steps from ECC to s4
PDF
Building Integrated photovoltaic BIPV_UPV.pdf
PDF
Build a system with the filesystem maintained by OSTree @ COSCUP 2025
PDF
Architecting across the Boundaries of two Complex Domains - Healthcare & Tech...
PDF
Review of recent advances in non-invasive hemoglobin estimation
PPTX
Big Data Technologies - Introduction.pptx
DOCX
The AUB Centre for AI in Media Proposal.docx
PPTX
Effective Security Operations Center (SOC) A Modern, Strategic, and Threat-In...
PDF
Encapsulation_ Review paper, used for researhc scholars
PDF
Peak of Data & AI Encore- AI for Metadata and Smarter Workflows
PDF
Encapsulation theory and applications.pdf
PPT
“AI and Expert System Decision Support & Business Intelligence Systems”
Spectral efficient network and resource selection model in 5G networks
Blue Purple Modern Animated Computer Science Presentation.pdf.pdf
VMware vSphere Foundation How to Sell Presentation-Ver1.4-2-14-2024.pptx
Unlocking AI with Model Context Protocol (MCP)
How UI/UX Design Impacts User Retention in Mobile Apps.pdf
KOM of Painting work and Equipment Insulation REV00 update 25-dec.pptx
Reach Out and Touch Someone: Haptics and Empathic Computing
MIND Revenue Release Quarter 2 2025 Press Release
sap open course for s4hana steps from ECC to s4
Building Integrated photovoltaic BIPV_UPV.pdf
Build a system with the filesystem maintained by OSTree @ COSCUP 2025
Architecting across the Boundaries of two Complex Domains - Healthcare & Tech...
Review of recent advances in non-invasive hemoglobin estimation
Big Data Technologies - Introduction.pptx
The AUB Centre for AI in Media Proposal.docx
Effective Security Operations Center (SOC) A Modern, Strategic, and Threat-In...
Encapsulation_ Review paper, used for researhc scholars
Peak of Data & AI Encore- AI for Metadata and Smarter Workflows
Encapsulation theory and applications.pdf
“AI and Expert System Decision Support & Business Intelligence Systems”

Ahead of the Curve: How 23andMe Improved UX with Performance Edge

  • 1. 1 Ahead of the curve How 23andMe improved UX with Performance Edge + Courtney Ball, Antonio Contreras, and Whelan Boyd
  • 2. 2 Introductions Courtney Ball Sr. Manager, Web Marketing Antonio Contreras Sr. Web Developer Whelan Boyd Sr. Product Manager
  • 3. ● Customize the widgets on your page to your preference ● This webinar is recorded and you will receive the link with the slides in the next few days ● We will have time for questions! Please submit them in the Q&A box on the left side of the screen Housekeeping
  • 4. 4 Agenda 1. Why performance matters 2. How to think about measuring performance 3. How experimentation affects performance 4. Performance Edge 5. Case study: 23andMe 6. Q&A
  • 6. 6 The business impact of speed 13% increase in sales when page load time was cut by half Found that 53% of mobile visits were abandoned if page took > 3 seconds to load 7% increase in conversion when page load time reduced by 850 ms Source: https://developers.google.com/web/fundamentals/performance/why-performance-matters/
  • 8. 8 Measure impact to user experience Source: Google’s Web Fundamentals First Paint (FP) First Contentful Paint (FCP) First Meaningful Paint (FMP) Time to Interactive (TTI)
  • 9. 9 Best practices for measuring Free Online Tools Synthetic Traffic Easy to run Actionable Insights Real User Monitoring Production Traffic Segment by Device, Location, etc. Robust Analytics
  • 11. 11 Standard Web Experiments 1. Snippet size increases with more experiments 2. For specific visitors, the downloaded snippet includes all project variations vs just active variations 3. Larger snippet results in more time to download and execute 4. Resource-constrained browsers have to execute several tasks Performance Challenges for Web Experiments Standard CDN Script Experiment A variation 1: CTA: Get Started variation 2: CTA: Try it out Experiment B variation 1: CTA: Promo Code variation 2: CTA: No Promo Code Task load: URL Targeting Audience Targeting Traffic Splitting Apply Variations Event Tracking
  • 13. 13 Only returns visitor-specific variations Standard Web Experiments Task load: Experiment Delivery Network Apply Variations Event Tracking Key experiment decisions are moved to the Edge Much smaller file to download and execute in the browser with less tasks to perform Performance Edge Experiments Standard CDN Script Experiment A variation 1: CTA: Get Started variation 2: CTA: Try it out Experiment B variation 1: CTA: Promo Code variation 2: CTA: No Promo Code URL Targeting Audience Targeting Traffic Splitting Apply Variations Event Tracking CTA: Get Started Script URL Targeting Audience Traffic Splitting
  • 14. 14 Why is it so fast? Less data Far smaller file loading over the wire Reduce download time Less code Only code necessary for active variations Fraction of the execution time in browser Faster machines Powerful Cloudflare Workers execute expensive tasks Located extremely close to end visitor
  • 15. 15 Proper Measurement Make things smaller Performance Edge ● User-focused Metrics ● Synthetic and real traffic ● Lighter download ● Shorter execution ● Quick implementation with any CDN ● Push logic to the Edge Summary
  • 17. 17 23andMe: Intro Courtney Ball Sr. Manager, Web Marketing Antonio Contreras Sr. Web Developer
  • 18. 18 23andMe: Overview Company size: 500+ employees Location: Sunnyvale, California Mission: Help people access, understand and benefit from the human genome. What we offer: 23andMe gives individuals easy and affordable access to their genetic information, offering a growing list of reports on traits, wellness, carrier status and genetic ancestry. Consumer Research Therapeutics
  • 19. 23andMe: Testing Overview TV AUDIO DISPLAY SOCIAL Testing focus for 2020: ● Optimize kit mix. ● Optimize mobile experience. ● Optimize personalization through audience targeting. 23andMe Customer Acquisition / Marketing Funnel SEARCH SITE Consideration Conversion Awareness
  • 20. 20 23andMe: Approach to improving user experience through a site redesign Create new designs (multiple options) Consumer insights conducts qualitative research Identify key areas of opportunity through a site audit A/B test against existing page via Optimizely Iterative testing (if needed)
  • 21. 21 Test: Redesigned Product Page Test Variant #1Control Round 1
  • 22. 22 Test: Redesigned Product Page Test Variant #1Control Round 1 Winner
  • 23. 23 Test: Redesigned Product Page Control Test Variant #1 Round 2 Test Variant #2
  • 24. 24 Test: Redesigned Product Page Control Test Variant #1 Round 2 Test Variant #2: Winner
  • 25. 25 Why we needed performance edge as we scaled A growing point of internal tension… 1. More testing = slower page speed 2. More emphasis placed on performance Eagerness to try new tactics to improve site performance ahead of the holiday season. + Frequency A/B Tests 2016 2017 2018 2019 2020 Year NumberofTestsLaunched 20 40 60 0
  • 26. 26 How we implemented Performance Edge at 23andMe 1. We had an existing Worker on Cloudflare that we were using to load our Optimizely JavaScript 2. We updated the existing Worker for Optimizely Edge and created a separate Worker for the old Optimizely JavaScript 3. We used an integration environment to test performance and root out errors Cloudflare CDN / Worker Client / Browser 23andMe Web Server
  • 27. 27 Speed improvements with Performance Edge Our performance statistics improved across the board: Standard Snippet ⚡Optimizely Edge Snippet First Contentful Paint Mobile: 2.7s Desktop: 2.5s Mobile: 1.2s Desktop: 1.1s First Meaningful Paint Mobile: 3.0s Desktop: 2.7s Mobile: 1.6s Desktop: 1.5s Time to Interactive Mobile: 6.7s Desktop: 6.2s Mobile: 4.4s Desktop: 5.4s Key Benefits: dramatic performance improvements on mobile and we can continue scaling our testing program without slowing down our site!
  • 28. 28 Questions? Ask them in the webinar console