DreamStream - Dream11's creator console for live sports

01. Context.

Why Dream11 needed a live room.

Our fans treat match day like a festival, but they had absolutely nowhere to hang out together on our app.

Leadership gave us a simple mandate: build a communal space where fans can experience the match together.

What we were building on

250M registered users
10M+ daily actives
12 years of sports DNA
A user base obsessed with cricket

The tools were totally fragmented.

Watching a match at home meant staring at five different apps at once. We wanted to pull all that scattered stadium energy into a single screen.

app · 01

Dream11

Fantasy - where the match actually matters.

app · 02

Live Score

Ball-by-ball on Cricbuzz or Google.

app · 03

OTT App

The actual broadcast of the match.

app · 04

YouTube

Creator reactions and hot takes.

app · 05

The banter with friends.

Five apps, one match. DreamStream collapses the whole room into a single surface.

Who we were solving this for.

We used engagement tools as the bridge between playing fantasy sports and feeling like you're in a crowd.

problem / 01

The Viewer

Watching alone and bored.

no shared room

problem / 02

The Creator

No dedicated platform for sports streams.

no home turf

problem / 03

Dream11

We owned the transaction, but completely missed the live community.

missing the room

Building the console.

What I owned

The Go Live flows - designing the setup.
The layout - figuring out what sits where.
The hierarchy - keeping chat accessible.
Feature surfaces - Squad Goals and Shout Outs.
Foundations - the core components.

What my team handled

Engineering - the live video pipeline.
Creator success - creator contracts.
Internal CS - manual moderation strategies.
PM - roadmap scoping.
Design - adjacent UX support.

02. Discovery.

Four quick inputs for maximum speed.

With only 60 days, we skipped the massive research phase and went straight for quick, dirty answers.

01 / benchmark

Competitive benchmarks

Twitch, YouTube Live, Kick and OBS - to see how creators actually work.

02 / interviews

Creator feedback

We literally grabbed creators from our onboarding cohort visiting the office for fast feedback.

03 / expert call

Twitch design call

One hour with Charlene from Twitch - don't repeat mistakes mature platforms already solved.

04 / shadowing

CS team shadow

Sat with the internal moderation team - moderation is a hidden nightmare we had to design around.

Takeaways from Twitch.

This 1-hour call saved us weeks of trial and error:

01. Creator-First

If the tools are annoying, creators won't stream. Fix friction first.

02. Chat As Content

Most viewers lurk, but chat still needs to look alive.

03. Engagement Earns Attention

Give creators interactive tools that create instant talking points.

04. Community Identity

People love feeling like insiders. Badges work.

05. Growth Beyond Live

Growth happens after the stream ends, when clips get shared.

06. Quality First

If the audio is bad, people leave. Period.

Testing with creators.

We hacked together a playable prototype just to see where our mental models completely clashed with the creators' habits.

Figma Make - playable go-live prototype · open in new tab ↗

Who saw it - stakeholders across product, content and ops; three creators already onboarded for contracts; one moderation lead.

What it asked - could a creator go from login to live in under a minute, using their existing OBS muscle memory? Where did the flow break down for them?

Creator go-live flow - 9 steps

01Login

02Empty landing

03Click go live

04Select match

05Fill match info

06Key generated

07Paste in OBS

08Pipeline created

09You are live

What we kept, and what we cut.

What we kept · 5

Passing the stream key directly to OBS.
Keeping controls right next to the video.
Letting creators customize their UI.
Using colored usernames to make chat punchy.
Relying entirely on OBS as the backbone.

What we killed · 5

Building our own broadcast tools.
Complex affinity systems for launch.
Deep engagement layers like digital gifting.
Rigid badge goals.
Building a custom overlay system.

Three realizations that changed everything.

reframe 01

Chat is raw material, not the feature.

Assumed: creators care most about chat features. Learned: they only care if the tool helps them make better content - so chat became something for the creator to react to.

reframe 02

Recognition before money.

Assumed: we need tipping on day one. Learned: smaller creators want loyalty first, big ones monetize elsewhere - so we held tipping and pushed recognition instead.

reframe 03

Motion over button density.

Assumed: fans want a massive, complex dashboard. Learned: they just want to feel the energy of the room - so we optimized for visible motion.

03. Decisions.

Four hard constraints we embraced.

Saying "yes" to these limitations is the only reason we shipped on time.

01 / tech

Engineering timelines

We designed around the engineers so they could focus purely on video stability.

02 / dependency

OBS as non-negotiable

We built around OBS rather than trying to compete with it.

03 / ops

Manual moderation

Our poor CS team handled the massive load of moderating manually for V1.

04 / design

No design system

We built basic components instead of a robust system for speed.

Three massive bets in the first week.

bet 01 · the tools

Lean completely into OBS.

Tradeoff · OBS dependencyProof · industry standard

Every established creator already uses OBS. We didn't want to build a clunky native tool and force them to relearn how to stream.

It made OBS a hard dependency for true beginners, but every benchmark proved OBS was the standard - so engineering could focus purely on the live pipeline.

bet 02 · feature priority

Prioritize engagement. Hold off on money.

Tradeoff · no tipping at launchProof · Twitch warning

Creators need an audience before they need a wallet, so V1 led with loyalty tools like Squad Goals. Money was staged for later: engagement (v1) → loyalty (v1.5) → gifting & subscriptions (v2).

Some creators were mad, but we held firm. Twitch design warned us: "Once you introduce a currency you can't take it away."

bet 03 · hierarchy

Keep chat close. Push the rest back.

Tradeoff · quieter featuresProof · Twitch layouts

Creators constantly scan chat while they talk, so we shoved everything else out of the primary view. Playground and Chat flank the stream; On Stage and Highlights sit pushed back.

On Stage became visually quieter and PMs fought it, but the scan-frequency data - and how Twitch handles layouts - won the argument.

DreamStream live console with Playground and Chat columns highlighted in red - the high-scan zones flanking the stream — Live console · red = high-scan zones flanking the stream

The simple four-step loop.

For the stream to feel alive, all four of these steps have to be visible at once.

01Tune inActivity is already happening.

02EngageThey drop a message in chat.

03RewardThe creator calls them out on stream.

04ReturnThey feel seen, so they come back.

↻ every feature asks which beat does this serve?

04. Solution.

Getting creators live in two minutes.

We cut all the fat. A creator can pull out their phone and start broadcasting before the coin toss.

Match select screen — Step 01 · Pick the live match

Stream details screen — Step 02 · Add title & thumbnail

Stream key screen — Step 03 · Generate stream key

Console screen — Step 04 · Followers drop in

Laying out the console by where people look.

We organized the dashboard purely by how often a creator needs to glance at it.

DreamStream live console annotated with 5 colored pointers showing the scan-frequency zones — Live console - annotated with the 5 scan-frequency zones

Stream - center, always. The main focus.
Chat - center-right. Right next to the video.
Activity - left. For managing polls.
Highlights - side rail. A sticky area for Shout Outs.
Session controls - bottom. Kept out of the way.

Closing the loop on recognition.

Everything was designed to drive interaction, not just look pretty.

Squad Goals - a shared goal for the audience to hit together. It gives the room a common objective.

Squad Goal feature visual with creator panel and mobile view — Squad Goals · loop position: engage

On Stage - brings a viewer directly into the live broadcast. When users see they might get pulled "on stage," they come back tomorrow.

On Stage feature visual — On Stage · loop position: engage + return

Chat & Shout Out - real-time tools to acknowledge the audience. It creates an immediate feedback loop.

Chat and shout out feature visual — Chat & Shout Out · loop position: reward

The same room, but on mobile.

Creators got the complex desktop cockpit, but viewers just got a lively, streamlined mobile room.

Viewer mobile - shout-out rail — Shout-out rail

Viewer mobile - user on stage — User on stage

05. Outcome.

The bet paid off: engagement led the way.

Here is how our month-one numbers stacked up.

16.7M

Watch minutes · month 1

"Watching alone, missing the room." Viewers loved it - session times destroyed our older metrics.

457.2K

Peak views · single stream

"No home for sports-first livestreams." Creators adopted it, and we barely got any setup-confusion tickets.

669

Streams · 20 creators

"Owns the transaction, not the room." We finally gave 250M users a reason to stay for the whole match.

Our first successful creator-led live product.
Our CS team handled the manual moderation beautifully.
We got V1.5 approved just four weeks later.

What I'm proud of, and what I'd do differently.

what I'd change

Build the design system properly.

Skipping a design system helped us ship fast, but we paid the price in tech debt later. Next time, I'd force a one-week pause to build real tokens - the tradeoff between speed and scale.

what I'm proudest of

The hierarchy worked.

Putting chat front-and-center while burying cooler features was a tough sell internally. But we trusted our observation data over personal taste - scan-frequency over feature prominence - and it worked.

See more case studies

DreamGuru - Built a two-sided fantasy guru marketplace, free first then paid.

Creator Economy

Watch Along - Changed user behaviour by shifting the primary view to landscape.

Behaviour Design

We built a live sports console in 60 days.