PixelStar
Activating Times Square with the world’s largest social feed. How TSX Entertainment created a $2.56M business by empowering users from around the world to be seen.
Objective
As TSX prepared for the launch of its 18,000 sq. ft, 8K, wraparound digital display, the product team began discovery with the goal of activating Times Square. The team sought out opportunities to begin delivering on TSX’s core mission to create a global epicenter for entertainment and cultural events.
Problem Statement
Empower Times Square tourists and people around the world to showcase their original content on the TSX Screen through an accessible, interactive user experience, so they can make memories and be seen.
Methodology
Using the Outcome-Driven Innovation product framework, the team discovered and defined PixelStar: a consumer-facing feature that allows end users to upload their own video or photo and schedule a 15-second slot on the TSX Screen in Times Square. The feature is available for Android and iOS, and is enabled by an AI-assisted content moderation workflow and companion NextJS app which receives UGC (user generated content) and displays it dynamically on the TSX Screen.
PixelStar was developed in an agile scrum environment and launched New Year’s Eve 2022. Several key enhancements have been released since, with a healthy feature roadmap extending well into 2024.
Results
A little more than one year from launch, PixelStar is delighting fans and disrupting the digital billboard space.
Discovery
TSX Entertainment’s core mission is to create an epicenter for entertainment and cultural events, where artists, innovators, and creators connect with fans globally, in powerful new ways. The fully realized vision for TSX spans 9 floors and 74,500 sq. ft of retail and entertainment space at TSX Broadway, and has been dubbed the word’s first web3 building.
In Spring of 2022, as the first phase of exterior construction was wrapping up, the product team prepared for the launch of TSX’s massive 18,000 sq. ft, 8K, wraparound digital billboard. We focused on answering the question: “How might we activate Times Square, to delight fans, and begin delivering on the TSX brand vision?”
The Opportunity
The team leveraged audience and market research to size the potential opportunity, and gain insight into our target users and competitors.
Audience Research
Design Sprint
Contributors and stakeholders from product, design, and engineering came together for a 4-day design sprint to explore the problem and discover potential design solutions.
Understand & Define
- How Might We?
- JTBD Market Definition Canvas
- JTBD Stories
Diverge & Sketch
- Job Map
- Crazy 8’s
Decide
- Revise Crazy 8’s
- Storyboards
Prototype Planning
- Paper Prototype
- Establish Prototype & Validation Plan
The team concluded the sprint with alignment on three concepts to prototype and test at higher fidelity. While this is a departure from the traditional design sprint which concludes with user testing a single concept, it was beneficial in TSX’s early stage discovery. We were cautious not to narrow down concepts too quickly, and wary of potential false signals that can result from testing an unfamiliar concept at too low a fidelity.
Validation
After the design sprint, I created medium fidelity prototypes for each concept. I presented them to peers and stakeholders, and made revisions to ensure the high-level experience and value offer resonated with the TSX brand. After several rounds of internal feedback, I teamed up with my product partner to conduct a round of in-person, moderated user testing and interviews. The purpose of these early tests was primarily to gauge value and comprehension.
Of the three concepts tested, Me-saic consistently received the highest value rating. It was also the leading favorite among stakeholders. Post-testing interviews provided insight into what specifically testers found to be valuable, and how we might evolve the concept to increase that value even further.
PixelStar is Born
The me-saic concept ultimately evolved into the product now known as PixelStar, which allows anyone in the world to upload their own video or photo and schedule a 15-second slot on the TSX billboard in Times Square. This evolution was guided by several key insights:
Users see value in being featured on the screen, but they want to be the star of the show
The average length of an ad Times Square is 15 seconds
The top NYC tourist attractions incorporate photos ops to memorialize the experience and encourage social validation
User Testing
After aligning on the PixelStar concept and creating a high-fidelity prototype, we conducted 3 rounds of remote, unmoderated user testing to measure usability and value. The target test segment included a mix of NYC locals as well as tourists who had visited Times Square within the past year.
5-Point Value Scoring
Testers were fairly polarized in their responses to value scoring questions. We were most interested in respondents who responded with a 3 or lower, while also expressing positive sentiment verbally throughout the test. We hypothesized that price sensitivity, or other barriers to entry could make interested prospective users less likely to act.
Sentiment Highlights
Usability & Comprehension
We observed some comprehension issues in the first round of testing. Most testers spent a significant amount of time on the PDP (product detail page) before moving forward, and some were unable to move forward at all. We revised the UX copy to convey the offer as succinctly as possible, and introduced additional supportive visual content. By the third round of testing, we observed faster comprehension of the offer, less time spent on the PDP, and no drop-off.
Another theme we observed was testers looking for clarifying information about their current context and/or reassurance about upcoming steps in the user journey. We looked for opportunities to empower users with the information they need to complete their current step in the journey, as well as to set expectations about the following step(s).
Implementation
Once we had clear alignment on the in-app user experience, we began one-week sprint cycles with iOS and Android engineering. In addition to daily standups, our typical sprint cycle included a regular cadence of agile ceremonies:
Review
- Demo progress from the previous sprint
Plan
- Groom & point backlogged tickets for the upcoming sprint
Discover
- Groom & point backlogged tickets for the upcoming sprint
Reflect
- Bi-weekly retro to reflect on the past sprint and identify opportunities
Opportunities for design and flow optimizations typically arose during technical discovery, giving product and design the remainder of the sprint to make revisions and document requirements.
Navigating Ambiguity
With app development well underway, the platform engineering team was focused on our implementation of sedna GmbH, the hardware and software that powers the TSX Screen. While sedna is a powerful platform used by some of the most notable digital displays around the world, what we were trying to accomplish differed significantly from standard content and media management, and playout use cases.
Platform Requirements
Intake and moderate user-generated-content
Schedule approved content to play at a specific time
Dynamically display scheduled content on the TSX Screen
Generate real-time status updates
Simply put, we were trying to determine how to get user-generated content from the app to our screen, and notify users when their content is displayed.
Ultimately we discovered that while SENDA did not natively support dynamic content, it is compatible with the chromium web browser. So, we built a NextJS web app to intake and display user generated content from a queue.
The platform team defined an inventory schema that allows users to schedule their content to play within a 5-minute window. Each 5 minute block of PixelStar is composed of 20 sellable inventory “seats,” each identifiable by a unique feature number. After a piece of content is moderated and approved, it fills one seat in the block. Once all 20 seats are filled, the block is sold out.
Moderation
PixelStar is public-facing, physical social feed, therefore moderation is crucial for ensuring a safe and inclusive experience. We opted to partner with the incredible team a Partner Hero to ensure all content submitted through the TSX app is double human-moderated within 15 minutes of submission. Users whose content does not pass moderation are given clear feedback citing what part of the content guidelines they are in violation of, and are issued an automatic refund.
Launch
Beta Program
TSX Screen officially powered up on December 6, 2022. At that time, we launched a Beta program, inviting friends and affiliates to try PixelStar free with promo code in exchange for candid feedback. We were able to process over 500 orders across the iOS and Android apps during the beta period, and identified and resolved several bugs concerning device-specific media encoding issues.
Hard Launch
PixelStar officially launched on New Year’s Eve 2022, in time to provide the holiday crowd with the opportunity to go live on the screen leading up to and after the iconic ball drop.
KPIs & Roadmap
Following launch, we measured success based on 3 key metrics and met biweekly to review our progress and align on what actions to take to improve performance.
Daily Sold
- Increased awareness through product marketing and social campaigns
- Released features with high user value to optimize product/market fit and lower barrier to entry
Daily Aired
- Lowered reduction rate by educating users on content guidelines
- Reduced improper rejections with moderator education
Utilization
- Optimized inventory allocation based on demand
- Sought alternative audiences or uses for low-demand times
Daily Sold & Utilization
Daily Aired & Rejection Rate
Since the PixelStar launch, the team released several key features to provide increased value to users and address unforeseen needs & opportunities.
Livestream
PixelStar was initially created to target locals and tourists in Times Square. Following launch, we learned that 38.23% of transactions came from users outside of the US at the time of purchase. For the international user segment, having a way to watch the TSX screen remotely was incredibly important. Customer support tickets revealed that the majority of these users were already relying on Earthcam’s Times Square Cam. TSX ultimately installed our own earth cam so we could provide an optimized, in-app viewing experience with limited downtown and latency.
In-App Editing
Since PixelStar MVP supported only video upload, we anticipated user demand for additional tooling. Customer feedback, as well as a PixelStar partnership program we ran with the Paris Hilton team, helped make the case for two high-value enhancements: photo slideshows and text editing. After evaluating different "build it" and "buy it" solutions, the team ultimately integrated with IMG.LY's advanced APIs to offer a fairly customized experience with maximum speed to market.
Text & Text Design
PixelStar partnerships allow artists and brands to align themselves with larger cultural movements by sponsoring PixelStar for fans, and inviting them to submit content following a prompt or theme. Text and text decoration created simple, yet powerful opportunities for users to respond to prompts and engage with themes.
Photo Slideshows
When asked what existing content they might want to put on the screen, users reported having more photos than videos, and overall lacked video content of high personal significance.
The slideshow feature, also powered by IMG.LY, allows users to select up to 5 photos, which are automatically converted into a 15-second video for display on the TSX Screen.
Upcoming Features
PixelStar Packages
Since launch, we have observed a growing segment of repeat buyers including small brands, business owners, and creators. Users in this segment often re-use the same content, meaning they have to go through the content upload and checkout process multiple times. We are currently developing a package feature that allows users to schedule recurring plays over a duration of time.
Variable Layout
Depending on their intent, users may be more or less reliant on the content they already have on hand. If I wanted to create a “Happy Birthday” PixelStar for my out of town friend, for example, I would likely scroll my camera roll to find a few recent photos of them. The current PixelStar screen display is optimized for square format content, which may not always be optimal for users working with existing content. Variable layout allows users to choose between square, landscape, and portrait aspect ratios, giving them more control over how their content displays on the screen, and empowering them to better leverage existing content.
Final Thoughts
PixelStar celebrated its one year anniversary December 31, 2023, achieving $2.56M in revenue. And I would say we are just getting started. The last 3-6 months have revealed a number of insights into how we might better meet the needs of the core PixelStar user, as well as reach additional segments. Those insights have helped us build out a healthy product roadmap geared towards increasing user value and strengthening product market fit.
As TSX’s flagship product, PixelStar has built the foundation for direct screen to user interactions. I look forward to exploring how we might expand upon that model of interaction, and potentially build out entirely new features or product areas.