Tubi

Emphasizing incentive and Tubi’s content, the goal was to educate and excite visitors, and obtain emails for contest reminders. I emphasized the prize messaging on the landing page, and displayed links to Tubi’s horror content on the confirmation page.

I started building a foundation of tools that would serve the entire project, and provided guidance to the team on creating the right assets for production. Throughout the contest experience, we’d utilize multiple stacked videos; a consistent background loop and different transparent videos above on the z-axis.

But how would this actually work in the browser?

I experimented with video formats to achieve the stack in the browser at small file sizes. The transparent animated text sits above a looping background and is served as multiple sources (MOV/HEVC and WebM/VP9), letting the browser load the one it needs. With this, I was able to:

• Confirm achieving the desired experience
• Establish what assets need to be created
• Communicate intent to the dev team; was built into the frontend

Through presenting work to Tubi, we collaboratively merged two concepts; introduce the Screaming Rooms and resolve to a timer counting down to the start of the contest. I also evolved the designs and frontend within a framework, while concurrently preparing to build the core contest experience.

We talked through the primary requirements and how the experience needed to account for the following:

• Traffic from social media or QR
• Room selection
• Sample scream
• Record, preview, submit
• Multiple submissions per user
• User’s submission status
• Submission gallery
• Promotional Tubi content
• Allow visitors to view submissions without submitting

I started with concepts for an interactive room picker with, to let the user choose, and afford recirculation for repeat screamers.

Permissions handling, sample scream, and camera initialization

I explored various methods of user interaction to hear the sample and start recording, and also at what point we’d be able to load the camera and its dependencies; to satisfy interaction requirements in the browser with regard to triggering media.

Record, try again, and preview flows

I iterated on several approaches including instant retries with separate recording review, the review and retry in one step, and celebratory positive feedback, while conferring with the dev team about the tech feasibility of grabbing the recording and being able to play it back to the user for review.

Created user journey flowcharts to facilitate design and collaboration

This was useful while working on the flow and mapping user input and multiple user submissions, but an advantage was with our team and collaborating with development, as history states and accounting for how iOS and Android differently handle the user's response to permissions were challenging.

My intent was to create a guided journey of interconnected steps with nudges to help users understand what actions to take and to increase the submission rate. I also expanded the design library and created templates for repeatable patterns as we prepared for building the experience.

The original intent, serving the mic and camera permissions and loading the camera earlier, was to have a seamless transition for the user to perform their scream after the sample scream prompt.

To fix the performance, my solution was to serve the permissions after the sample video, which really is the last available time, as the user is a bit deeper. Then if the camera is allowed, fire it up with the 8th Wall dependency, behind a quick loader.

This was reflected in the final design and revised flow; separating the camera and scream sample gave us seamless video playback, loading the camera if and when it’s needed.

And cued up that QR code.

Handling user submissions

Once a user submits a scream, they land at the Scream Gallery featuring other user videos and links to Tubi’s horror content. Submissions enter a process of moderation, and I collaborated with the dev team to allow users to obtain their approved video.

As things go sometimes, there was no real opportunity with schedule and budget limitations to measure and iterate. I could have assessed the perception of the contest incentive. I could have explored various methods of promoting Tubi’s content to drive account registration. I could have caught a common point of drop-off along the flow (later revealed to be the submission form).

Once live, I noticed that many of the submitted videos were cut off near the end. The pre-determined five-second recording duration was likely too short, and that could have been addressed earlier.

Per the analytics data, about half of the unique visitors submitted their screams, entering the contest. Some users submitted once, and it was interesting to see a spike of users that had submitted three times, which seems to align with a desire to use each of the three Screaming Rooms.