Summary
Nearly a dozen concepts that aligned with business and customer goals were design, tested, and shipped over the course of 18 months resulting in an est. ~$15 million of revenue over a four-year span.*
* Based upon projections for 2020-2024.
Context
Ancestry.com has two product lines: family history record subscriptions, and DNA tests. This particular onboarding initiative was specific to the family history record subscription business. This was a team effort. Special thanks to Liz Brown, Josh Harman, Jasmine Sarkhazi, Jeff Alton, and the Groot Team for their endless collaboration, support, and feedback as we worked on these initiatives together.
Timeline
February 2019–October 2020
Total Elapsed Time: ~18 months
Customer & Business Problems
FOR THE CUSTOMER
As a new customer, I don’t know how Ancestry.com works, where to start, or what to do next.
FOR THE BUSINESS
As a business, I need to increase bill-thru rate of new customers and find ways to encourage long-term customer retention.
Project KPIs
- Increase bill-thru rate—the number of free-trialers who complete their trial period with a subscription—by 5%.
- Reduce day-0 cancellations.
Learn + Map
Together, my PM and I began by mapping out existing onboarding flow(s) that led customers into the family history record subscription product. We then gathered any existing data we could find—both qual. and quant. From there, we broke the customer journey flows down into sections that appeared to provide the best opportunity for improvements to the KPIs—while also improving the customer experience. Finally, we generated a list of questions/hypotheses that we wanted to find answers to.
ACTIVITIES
- Identifying desired outcomes & KPIs
- Mapping out the existing flow(s)
- Gathering previous qual. research
- Gathering previous quant. research
- Compiling unanswered questions
- Conducting intent surveys
- Conducting user observation
- Competitive & Comparative Analysis
DELIVERABLES
- Overview of existing flow & entry points
- Summary of existing qual. research
- Gathering previous qual. research
- List of questions to answer
- Survey of results summary & next steps
- Summary of user observations
- Comparative & Competitive Ana. Results
Existing Flow Diagramming
My PM and I spoke with stakeholders in marketing, product, and development to get as complete a picture over the system-level flow of the current onboarding experience. He and I created the following chart to explain the complexity of the onboarding experience at a quick glance. Before creating the graphic, Liz, Josh, and I—mostly Liz—printed out as many pages of the flow we could and annotated them with sticky notes.
Competitive & Comparative Analysis
This project was highly iterative. We conducted generative research (1. Learn + Map), tested prototypes with potential customers (2. Design + Test), shipped concepts via A/B tests (3. Launch + Learn) and began the process over again. These competitive and comparative analyses were a part of a second round of onboarding research I conducted.
Scope Clarification Doc
The scope and objective of the onboarding initiative wasn't this clear at the beginning of the project, and this was created afterward to help other teams better understand our space. Initially, my PMs and I had to narrow down what sections of the flow/audiences we were going to focus on. It was messy.
List of Questions / Hypotheses
After having documented the existing flow. Theses were some of the questions we sought to answer through our research.
Design + Test
Once we had identified the target audience—free-trialers without family trees—through the quantitative data we gathered, we began to generate concepts to test at each stage in the onboarding process. We began with paper prototypes, and moved on to high-fidelity prototypes—since Ancestry had a design system in place.
ACTIVITIES
- Concept Theme Generation
- Paper prototype & sketch creation
- Visual design mockup creation
- Flowcharting new flow
DELIVERABLES
- Prioritized concepts
- Paper sketches & prototypes
- Visual design mockups
- Updated flow chart(s)
Theme Creation & Early Sketches
After generating the following concepts as a team, we then grouped the ideas into four broad themes and began sketching concepts.
CONCEPT THEMES
Paper Sketches & Flows
Once we had prioritized which concepts we thought would do best at increasing bill-thru rate and reducing day-0 cancellations, I began creating paper prototypes of these concepts so we could test.
Qualitative & Quantitative Testing
After creating paper prototypes, we tested each concept qualitatively. Whichever concepts tested well qualitatively, we then tested quantitatively with a live A/B test on the site.
KEY
Failed Quant.(Learning)
Concepts with this label were tested live on the site in an A/B test and failed to increase bill-thru rate or reduce day-0 cancellations against the control.
Success Quant./Qual.
Concepts with this label were tested live on the site in an A/B test and increase bill-thru rate or reduced day-0 cancellations against the control.
Saved (Not Coded)
Concepts with this label were saved to test later.
EXISTING PROTOTYPES • THEME #1 • INCREASE GUIDANCE / GIVE CONTEXT
Hypothesis: If we offer a quick and easy way to get going with adding information to their family tree—users will be more likely to enter into—and stay within—the discovery flywheel. Also, if customers know how much time they have left, urgency will lead to engagement.
Concept:
Chatbot for Tree Building
Failed Quant.(Learning)
This was an existing onboarding code prototype that hadn't been tested yet. I didn't create this.
Concept:
Free Trial Countdown
Success Quant. (Learning)
This was an existing onboarding code prototype that hadn't been tested yet. I didn't create this.
Tested with potential customers who had never used Ancestry before.
PAPER PROTOTYPES • THEME #2 • OPTIMIZE VALUE
Hypothesis: By providing value earlier in the onboarding experience—such as yearbook photos or surname insights—users will feel rewarded and will be more likely to continue on to enter into the discovery flywheel (Seeing a record that has family info in it, and then seeing another etc.).
Concept:
Chatbot for Tree Building
Failed Quant. (Learning)
Sketch credit: Liz Brown
Concept:
Yearbook Discovery
Saved (Not Coded)
Concept:
Record Tour
Saved (Not Coded)
Concept credit: Josh Harman
Tested with potential customers who had never used Ancestry before.
PAPER PROTOTYPES • THEME #3 • INCREASE DISCOVERY RATE
Hypothesis: By decreasing the time to first discovery—hint, first value, or first search—customers will see a quicker return on investment, will be willing to invest more, and eventually be more likely to bill through (Sign up for a record subscription).
Concept:
Obituary Hints
Saved (Not Coded)
Concept:
Early Intro to Search
Success Quant. (Learning)
Concept:
Next Best Steps
Success Quant. (Learning)
Tested with potential customers who had never used Ancestry before.
PAPER PROTOTYPES • THEME #4 • REDUCE PAYWALL FRICTION
Hypothesis: By clarifying and simplifying the value propositions for each plan—and making upgrading easy—users will be more likely to choose a plan, create an account, and eventually bill through (Sign up for a record subscription).
Concept:
Simple Plan Select
Success Qual. (Learning)
Concept:
Gifting Records
Saved (Not Coded)
Concept:
Option to upgrade
Saved (Not Coded)
Selection of Final Deliverables
After iterating and testing with customers repeatedly, we were able to identify and build those concepts on the site and test them against the existing onboarding experience on production.
LABELLING RECORDS WITH PLAN NAMES DURING FREE TRIAL
Hypothesis: If we show potential customers which records belong to which plans, and give them an upfront and simple way to sign up for a free trial, they will be more likely to purchase a subscription when the time comes.
REDESIGNING THE PLAN SELECTION PAGE AND PAYWALL PAGES
Hypothesis: If we simplify and unify the paywall pages and plan selection page, customers will be more likely to sign up for a free trial or subscription.
PROVIDING NEXT BEST STEPS PROMPTS & GUIDANCE
Hypothesis: If we show users how to begin getting hints and how to use search, then they will be more likely to have success finding records. If they have greater success finding records, then they will be more likely to bill-thru after their trial.
ANIMATED "HOW ANCESTRY WORKS" VIDEO
Hypothesis: If we offer a quick overview of how building a family tree allows Ancestry to find records about their family, then when we ask them to start building a family tree, they will know why they are building a tree. If they build a tree, then they will be more likely to receive Hints (discoveries) and eventually bill-thru.
Launch + Learn
Several of these concepts were launched at once, while a majority were A/B tested as individual concepts. I supported the developers in implementation across 7 different dev teams and monitored the analytics in relation to the project KPIs.
ACTIVITIES
- Design to development hand off
- Design implementation support
- Quantitative data analysis
DELIVERABLES
- Dev support services
- KPI data reports
The Outcome
The collection of concepts that won both qualitatively and quantitatively were launched to production. These concepts were all featured in the pitch deck that helped Ancestry sell to Blackrock for $4 billion dollars. Here is why:
KPI RESULTS
+6.3%
+8.5% on mobile
Bill-thru Rate
+18%
N/A
Onboard Completion Rate
-8%
-12.1% on mobile
Day 0 Cancel Rate
NET PROFIT
~$15 Million in revenue over the next 4 years alone*.
* Based upon projections for 2020-2024.
Lessons Learned
- Failure is learning in disguise.
- Many small improvements can net large overall improvements for the customer and the business.
- It is important for me to pace myself and not get so carried away in the excitement of new projects that I burn myself out.
- Cross-team alignment is hard, but it is a lot easier—both the teams you serve, and yourself—if you focus on how what you’re building can help them reach their goals.
