<aside> 📈

Instagram’s first TV app beat its MVP launch targets — exceeding the downloads goal by 10% and the average time spent goal by nearly 2X — with strong positive press and social sentiment.

</aside>

As functional content design lead for Instagram's first TV app, I owned the content foundation from scratch and held the quality bar for the entire product, including guiding a second content designer's work through stakeholder alignment and final decisions. Every surface needed patterns for a screen, input method, and viewing context Instagram had never shipped for before, and building that foundation meant helping bring one of the world's largest social platforms to the living room for the first time.

My work spanned foundational UX patterns (login, onboarding, navigation, settings, profiles, error states), the cross-device mechanics unique to TV (QR-code activation, two-factor security for a shared living-room device, casting content over from a phone) and the go-to-market language that introduced the product itself, from App Store copy to mobile growth prompts. I also developed the channel naming strategy for merchandising interests-based content in a TV-native browsing experience.

Some of my work

Amazon Fire TV app store. I drove messaging strategy and crafted the copy for this critical stage of the funnel, anchoring on emotional benefits instead of feature lists to make the case for reels on the biggest screen in the house and drive downloads. This was also one of the first GTM assets to reflect the bolder, more mature brand voice I’d led the work to develop.

Amazon Fire TV app store. I drove messaging strategy and crafted the copy for this critical stage of the funnel, anchoring on emotional benefits instead of feature lists to make the case for reels on the biggest screen in the house and drive downloads. This was also one of the first GTM assets to reflect the bolder, more mature brand voice I’d led the work to develop.

Post-MVP login screen. To improve download-to-login conversion, I drove a headline exploration of alternatives to the co-watching value prop we'd launched with, testing the hypothesis that it read too narrow for people trying the app solo. I built consensus for shipping this version, which leads with one of Instagram's greatest strengths: the scale and quality of personalized content. The subhead simply signals account options without over-explaining security.

Post-MVP login screen. To improve download-to-login conversion, I drove a headline exploration of alternatives to the co-watching value prop we'd launched with, testing the hypothesis that it read too narrow for people trying the app solo. I built consensus for shipping this version, which leads with one of Instagram's greatest strengths: the scale and quality of personalized content. The subhead simply signals account options without over-explaining security.

First-time user 2-factor auth. This modal follows a first login already verified via mobile 2FA, asking whether to require it every time to keep your account private on TV. Because this language was so crucial to get right, we debated every detail — including whether "two-factor authentication" read as too technical — and I drove the cross-stakeholder alignment that got us to a shipped decision.

First-time user 2-factor auth. This modal follows a first login already verified via mobile 2FA, asking whether to require it every time to keep your account private on TV. Because this language was so crucial to get right, we debated every detail — including whether "two-factor authentication" read as too technical — and I drove the cross-stakeholder alignment that got us to a shipped decision.

End of feed content system. Rather than one generic message, I built a system that customizes the end-of-feed moment to the feed you're in: "caught up with reels from people you follow," "reels your friends are into," "reels for [channel]." Each pairs with an optimistic, encouraging CTA designed to keep sessions going instead of ending them, one of the small systems doing real work on session length.

End of feed content system. Rather than one generic message, I built a system that customizes the end-of-feed moment to the feed you're in: "caught up with reels from people you follow," "reels your friends are into," "reels for [channel]." Each pairs with an optimistic, encouraging CTA designed to keep sessions going instead of ending them, one of the small systems doing real work on session length.

Casting from mobile. This key mobile-to-TV growth lever didn't use Google's actual casting protocol, so the team had concerns labeling it that way. My solution: a plain-language CTA that’s just as clear, matches terminology on the TV app and content patterns on the sheet, and was validated by leads via dogfooding.

Casting from mobile. This key mobile-to-TV growth lever didn't use Google's actual casting protocol, so the team had concerns labeling it that way. My solution: a plain-language CTA that’s just as clear, matches terminology on the TV app and content patterns on the sheet, and was validated by leads via dogfooding.

Launch promo unit. This notif was our only mobile promo unit, so it was essential for top-of-funnel. But IG notif standards ruled out ad-like language that might oversell the MVP experience anyway. I fit awareness of the new app and its Fire TV requirement into two tight lines that were the first intro for many people.

Launch promo unit. This notif was our only mobile promo unit, so it was essential for top-of-funnel. But IG notif standards ruled out ad-like language that might oversell the MVP experience anyway. I fit awareness of the new app and its Fire TV requirement into two tight lines that were the first intro for many people.

Priming for network scan request. Research showed people reflexively dismiss OS permission prompts without context, so this bridge screen sets expectations before that ask. I chose the softer "Continue" over a more explicit CTA like "Scan your network," for plain, low-friction language to get people through.

Priming for network scan request. Research showed people reflexively dismiss OS permission prompts without context, so this bridge screen sets expectations before that ask. I chose the softer "Continue" over a more explicit CTA like "Scan your network," for plain, low-friction language to get people through.

Grammys growth campaign. Timed to re-engage people who'd downloaded the TV app but gone quiet, this notif had to spark urgency without pushy language, so I anchored on cultural FOMO and led with a familiar TV call to action. For the banner, I pushed for leaning entirely on The Grammys' own "Music's Biggest Night" tagline to carry the message through with authority.

Grammys growth campaign. Timed to re-engage people who'd downloaded the TV app but gone quiet, this notif had to spark urgency without pushy language, so I anchored on cultural FOMO and led with a familiar TV call to action. For the banner, I pushed for leaning entirely on The Grammys' own "Music's Biggest Night" tagline to carry the message through with authority.

Channels explore screen and naming. LLM-generated channel titles were accurate but flat, generic phrases like "Music and concerts" that didn't feel built for TV. I led the strategy to fix that: a tone framework and competitive analysis translated into shared design principles, then a revamped system prompt. I tested four persona-driven variants and landed on a clear winner, titles with a conversational, sometimes playful edge, true to the brand.

Channels explore screen and naming. LLM-generated channel titles were accurate but flat, generic phrases like "Music and concerts" that didn't feel built for TV. I led the strategy to fix that: a tone framework and competitive analysis translated into shared design principles, then a revamped system prompt. I tested four persona-driven variants and landed on a clear winner, titles with a conversational, sometimes playful edge, true to the brand.