TreeAssist: Improving time to value

TreeAssist:
Improving time to value

Designing value-first onboarding with AI-powered relative suggestions to accelerate activation and free-trial conversion.

Designing value-first onboarding with AI-powered relative suggestions to accelerate activation
and free-trial conversion.

TL;DR

Problem:

41% of new users dropped off before experiencing product value. New users were overwhelmed by admin-heavy onboarding and a blank family tree canvas, preventing activation which is the strongest predictor of free trial conversion.

Solution:

Designed TreeAssist, an AI-assisted onboarding experience that replaced form-heavy setup with in-product relative suggestions and progressive guidance, helping users build their tree faster and learn by doing.

Impact:

+18%

Free trial starts

+18%

Free trial starts

+18%

Free trial starts

+8%

Free trial --> subscription conversion

+8%

Free trial --> subscription conversion

+8%

Free trial --> subscription conversion

+26%

Users building trees with 10+ nodes

+26%

Users building trees with 10+ nodes

+26%

Users building trees with 10+ nodes

-13%

Reduction in onboarding drop off

-13%

Reduction in onboarding drop off

-13%

Reduction in onboarding drop off

My Role:

Worked with Lead Product Designer on end-to-end design of activation experiments, partnering with PM, data science, and engineering to design and ship three onboarding experiments that improved early user success and monetisation.

My Role:

Worked with Lead Product Designer on end-to-end design of activation experiments, partnering with PM, data science, and engineering to design and ship three onboarding experiments that improved early user success and monetisation.

Stakeholders:

New Users, Head of Tree Product, Head of Product Experience

Timescale:

2 Months end-to-end , 2-week experiment cycles

Deep Dive

Deep Dive

Business problem:

Improve early user activation to increase free trial starts and downstream subscription conversion, unlocking growth from new genealogy users.

User Problem:

41% of new users dropped before experiencing product value. New users felt overwhelmed by admin-heavy onboarding and a blank tree canvas, leaving them unsure how to begin building their family tree.

Hypothesis:

If we replace form-heavy onboarding with AI-assisted relative suggestions and guided tree building, users will reach value faster, increasing activation, tree depth, and free trial conversion.

Getting Started Experiments

Getting Started Experiments

Experiment 1

—> Reveal immediate value

Intent: Reinforce product value immediately after signup.


Change: Surfaced a user’s likely birth record using signup data, showing that Findmypast could instantly uncover real family history. This experiment could only apply to users with British or Irish ancestry.


Result: 60% task acceptance rate, validating that early value moments increase engagement.

Experiment 2

—> Guide the next action

Intent: Reduce post-onboarding confusion and help users continue building their tree.


Change: Introduced suggested mothers and spouses directly on the tree canvas, giving users an obvious next step without requiring manual record searches.


Result: +32% hint engagement post experience, increasing tree expansion behaviour.


Experiment 3

—> Interactive Onboarding

Intent: Help users reach the core product faster.


Change: Reduced onboarding from 8 forms to 3, then shifted users into AI-assisted tree building with suggested relatives.


Result: –13% reduction in onboarding drop-off and higher activation.

Next steps

Next steps

Introduce a personalised task list within the tree canvas to guide new users through key actions during their free trial, helping them discover more product value and increasing the likelihood of subscription.

Trade-offs & Constraints

Trade-offs & Constraints

Prioritised activation over data completeness
Reducing onboarding steps meant collecting less upfront profile information. This improved early engagement and tree creation but resulted in lower short-term profile richness.

Experience limited by record availability
TreeAssist suggestions relied on historical record matches, which meant the experience worked best for users with British ancestry, where record coverage was strongest.

Advanced users needed faster access to complex tools
Simplifying onboarding improved outcomes for new users but created friction for experienced genealogists who preferred quicker access to advanced search and editing features.

Reflection

Reflection

Helping users experience value early is often more impactful than collecting perfect data upfront.
Designing for momentum can unlock both engagement and monetisation.

Descope dungeon

Descope dungeon

Managing latency in activation moments - We descoped this in interest of release speed
TreeAssist suggestions required anywhere between 2–30 seconds to generate, creating a potential drop-off point. I proposed progressive, personalised content during the wait state to maintain engagement, but the team prioritised delivering the core activation improvements within the release timeline.

1

1

findmypast.co.uk

3:30

We’re checking our records for

information on your relatives

Browse while you wait

2 of 5

Tips to grow your family tree:

Fill out any information you know, even if it is just a guess.

Look out for hints that can help you learn more about your family.

Ask your family members to help with any information they might know.

1

1

findmypast.co.uk

3:30

We’re checking our records for

information on your relatives

Browse while you wait

3 of 5

Newspaper headlines on your birthday

1

1

findmypast.co.uk

3:30

We’re checking our records for

information on your relatives

Browse while you wait

4 of 5

Did you know that...

We have over XXX million records and XXX million newspapers to explore.

1

1

findmypast.co.uk

3:30

We’re checking our records for

information on your relatives

Browse while you wait

5 of 5

Try Findmypast free for 7 days

Build your family tree and discover countless family stories

Access all our records and newspapers, including the 1921 Census

Search other family trees and share yours with family and friends

Learn more

Want to work together? ashni.dave111@gmail.com

Resume

Want to work together?

ashni.dave111@gmail.com

Resume