Surprise Me! (TED App)

Challenge

Users have trouble discovering interesting content in the TED app and even superfans aren’t aware of the breadth of TED’s content offerings.

Objective

Build a new app feature that exposes users to a breadth of content they'd enjoy but not normally seek out.

My Role

UX Lead

Timeline

7 weeks

Results

Check it out →

+ 15% video starts
+ 35% video shares
+ 50% account conversions

OPPORTUNITY
——

Users have trouble discovering content in the app and aren’t interested in recommendations from a new personalization engine. Prior research found content discovery is primarily impeded by a lack of category and thematic browsing.

While the team works on long-term solutions⏤overhauling category taxonomies, implementing category landing pages and retooling the personalization engine⏤we needed an interim solution.

DISCOVERY RESEARCH
——

To learn why recommendations don’t resonate with users and uncover quick-win ideas to boost content discovery, I led a round of discovery user research.

Analytics Audit
Parsed app analytics so usage trends could emerge.

Social Audit
Analyzed mobile engagement trends of 25M+ TED followers across LinkedIn, YouTube, Facebook and Instagram.

Current-State Testing & Interviews
Led 5 moderated tests of live app to explore the ‘why’ behind known usage patterns and uncover blindspots.

Surveys
Polled 1,500+ app users and social media followers via 2 concept-testing surveys.

Research Findings

Users don’t look for content. They stumble upon it.

——

Since social platforms are users’ go-to source for content, the bar is high for how publishers package and deliver media. Most content consumed is happened-upon, not sought-out.

Over-personalization backfired.
——
Tailoring content to user interests limits the joy of finding something interesting you previously thought was boring. This ability to offer a new way of thinking makes TED different—we expose users to topics they wouldn’t normally seek out.

CONCEPT
——

After affinity mapping research findings, we opted to create video-randomizer feature users swipe through like a dating app or deck of cards. To counteract the personalization issue, we leaned into randomness and switched the burden of discovery so the platform does the heavy lifting, not the user.

DESIGN KPIs
——

Boost Discovery

Expose users to a breadth of content they'd enjoy but not normally seek out.

Unique Impressions
Increase # videos users see (not watch) in a session

Play Starts
Increase # of video plays per user/session

Shares
Increase # of shares per user/session

Boost Retention

Encourage users to save as they swipe, to sustain browsing while curating content for future consumption.

Monthly Retention
Increase # of sessions per user/month

Saves
Increase # of saves per user/session

Account Acquisition
Increase % of signed-in users

PLANNING
——

Design & Testing Timeline

Since there was no time allocated for user testing, I created a timeline to show stakeholders how it could easily be baked into the design process.

The timeline covers all facets of design⏤from paper prototyping to recruiting usability testers to hi-fi designs.

DESIGN

——

Play & Experimentation

I kicked off the design process by creating a sandbox file to explore what I wanted (and didn’t want) the experience to feel like.

Considerations:

What actions will be most common?
Upon finding an interesting video, what will users do next?
What gestures feel most natural?

Content Scaffolding & Hierarchy

What info do users need to quickly decide if a video is of interest?

Choosing a video is hard because imagery of speaker contributes very little. And, unlike an article, you can’t scan text and subheads to see the content is worth your time.

USABILITY TESTING

——

Which design performs better?

I led 8 sessions testing 2 different prototype designs.

Key Questions:
How well do users understand what system does and how to operate?

How easily are gesture operations discovered?

Which version sustains attention longer?

Design with lightest cognitive load kept users engaged longer, as browsing paths disrupted flow and discouraged engagement.

The most important design element proved to be having an information-bearing title with nothing below the fold.

MVP DESIGN

——

An immersive experience focused on spontaneity and simplicity.

Gives users feeling of control and choice while exploring a vast amount of content.

Prepping designs for handoff

I captured and annotated all potential use cases to:

Communicate path logic to engineering teams

Ensure all states and use cases were accounted for

Ensure accurate translation and implementation of new pages and components

RESULTS
——

1-month post launch

15%

increase in video starts (2.7M)

35%

increase in video shares (42K)

252%

increase in video downloads

50%

increase in account conversions (57K)

16%

increase in video completions (874K)

100k

feature adopters

Rachel Costas

Surprise Me! (TED App)

Users have trouble discovering interesting content in the TED app and even superfans aren’t aware of the breadth of TED’s content offerings.

Build a new app feature that exposes users to a breadth of content they'd enjoy but not normally seek out.

UX Lead

7 weeks

Check it out →

+ 15% video starts + 35% video shares + 50% account conversions

OPPORTUNITY——

OPPORTUNITY——

Users have trouble discovering content in the app and aren’t interested in recommendations from a new personalization engine. Prior research found content discovery is primarily impeded by a lack of category and thematic browsing.

While the team works on long-term solutions⏤overhauling category taxonomies, implementing category landing pages and retooling the personalization engine⏤we needed an interim solution.

DISCOVERY RESEARCH ——

DISCOVERY RESEARCH ——

To learn why recommendations don’t resonate with users and uncover quick-win ideas to boost content discovery, I led a round of discovery user research.

——

——

Since social platforms are users’ go-to source for content, the bar is high for how publishers package and deliver media. Most content consumed is happened-upon, not sought-out.

——

——

Tailoring content to user interests limits the joy of finding something interesting you previously thought was boring. This ability to offer a new way of thinking makes TED different—we expose users to topics they wouldn’t normally seek out.

CONCEPT——

CONCEPT——

After affinity mapping research findings, we opted to create video-randomizer feature users swipe through like a dating app or deck of cards. To counteract the personalization issue, we leaned into randomness and switched the burden of discovery so the platform does the heavy lifting, not the user.

DESIGN KPIs——

DESIGN KPIs——

Boost Discovery

Expose users to a breadth of content they'd enjoy but not normally seek out.

Boost Retention

Encourage users to save as they swipe, to sustain browsing while curating content for future consumption.

PLANNING——

PLANNING——

Since there was no time allocated for user testing, I created a timeline to show stakeholders how it could easily be baked into the design process.

The timeline covers all facets of design⏤from paper prototyping to recruiting usability testers to hi-fi designs.

DESIGN

DESIGN

——

——

I kicked off the design process by creating a sandbox file to explore what I wanted (and didn’t want) the experience to feel like.

What info do users need to quickly decide if a video is of interest?

Choosing a video is hard because imagery of speaker contributes very little. And, unlike an article, you can’t scan text and subheads to see the content is worth your time.

USABILITY TESTING

USABILITY TESTING

——

——

I led 8 sessions testing 2 different prototype designs.

Key Questions:How well do users understand what system does and how to operate? How easily are gesture operations discovered? Which version sustains attention longer?

Design with lightest cognitive load kept users engaged longer, as browsing paths disrupted flow and discouraged engagement. The most important design element proved to be having an information-bearing title with nothing below the fold.

MVP DESIGN

MVP DESIGN

——

——

Gives users feeling of control and choice while exploring a vast amount of content.

I captured and annotated all potential use cases to:

I captured and annotated all potential use cases to:

Communicate path logic to engineering teams Ensure all states and use cases were accounted for Ensure accurate translation and implementation of new pages and components

RESULTS——

RESULTS——

15%

15%

35%

252%

50%

16%

100k

+ 15% video starts
+ 35% video shares
+ 50% account conversions

OPPORTUNITY
——

OPPORTUNITY
——

DISCOVERY RESEARCH
——

DISCOVERY RESEARCH
——

CONCEPT
——

CONCEPT
——

DESIGN KPIs
——

DESIGN KPIs
——

PLANNING
——

PLANNING
——

Key Questions:
How well do users understand what system does and how to operate?

How easily are gesture operations discovered?

Which version sustains attention longer?

Design with lightest cognitive load kept users engaged longer, as browsing paths disrupted flow and discouraged engagement.

The most important design element proved to be having an information-bearing title with nothing below the fold.

Communicate path logic to engineering teams

Ensure all states and use cases were accounted for

Ensure accurate translation and implementation of new pages and components

RESULTS
——

RESULTS
——