One studio, every surface — Telegram, VK and LINE Mini Apps, full web games, and worlds in Roblox and Fortnite. Your brand, your players, one pipeline.
Mini Apps
Drop a branded game straight into the chats your players never close — Telegram, VK and LINE. One build runs in all three, launches in a single tap, and never asks for an install.
A brand experience that's one click away — playable on any link, on any device, with nothing to download. The open web is where instant play already lives.
A custom island in the world's biggest game, built in UEFN. Players already spend nearly half their time in creator-made worlds — so yours becomes a destination, not a billboard.
Tell us exactly what you need — and let's create it together.
Describe the game, the audience, the brand. Pick a platform, drop a reference, and we'll turn it into something players can't put down.
Be ready. Soon.
STRATEGY
The Attention Arbitrage: Why a Skippable Ad Can't Compete With 11 Minutes of Play
@RELI·Field notes·5 min read
A banner earns one click per 2,000 views. A video ad gets skipped in five seconds. A branded game holds attention for minutes — and the data says that's the only number that converts to memory.
Here is a number that should unsettle every brand still pouring budget into display: the average banner ad on the Google Display Network is clicked 0.46% of the time. In the worst-performing categories it sinks to roughly 0.10%, and historically across all formats it has bottomed out near 0.05% — one click for every 2,000 impressions you pay for. The other 1,999 people scrolled past a rectangle they never registered.
Video is supposed to be the upgrade. It isn't, not by the metric that matters. On YouTube, the skip button appears at 5 seconds, and 78% of viewers take it the instant it does. You bought a 30-second story and delivered a 5-second interruption — a toll people pay to get to the thing they actually came for.
Now hold that against a branded game. On Roblox, branded experiences hold an average of 11 minutes of engagement per session. Not 11 seconds of viewability. Eleven minutes of a player choosing to be inside your brand. That gap — between a click that almost never happens and minutes that reliably do — is the arbitrage. This is the trade @RELI exists to make.
Stop buying impressions. Start buying attention.
The advertising industry spent a decade optimizing the wrong number. Impressions are cheap and abundant, which is exactly why they're worthless: an impression only certifies that a pixel loaded, not that a human looked. The discipline that fixed this is attention measurement, and the firm that built the standard — Lumen Research, working with dentsu — gives us an apples-to-apples currency: attentive seconds per 1,000 impressions (APM).
Measured that way, the formats invert. Traditional online display generates about 1,416 attentive seconds per 1,000 impressions. Online video roughly quadruples it at 6,100. And intrinsic in-game advertising — ads native to the play environment — lands at 3,442 attentive seconds per 1,000 impressions, with gaming overall averaging 6,736. The screen people are paid to ignore is the one they can't stop watching.
78%Of viewers skip a skippable video ad the moment they can — Wave.video
~11 minAverage engagement per session on Roblox branded experiences
Attentive seconds per 1,000 impressions, by ad format. In-game and gaming environments dominate the formats brands actually spend on.Source: Dentsu & Lumen, Gaming & Advertising Attention Report 2024
Attention isn't a vanity metric — it converts to memory
The reason this reframe holds up is that attention is not the goal; it's the leading indicator. Lumen's research found a direct link between attentive seconds and brand recall: the longer someone actually looks, the more they remember, and the more they remember, the more likely they are to choose you. In the dentsu/Lumen gaming study, in-game ads drove 49% average prompted brand recall at 98% viewability — figures that traditional display, viewed for under three seconds, cannot approach.
This is the quiet collapse of the cheap-impression model. A banner served 2,000 times to win one click has, by definition, failed to be remembered 1,999 times. A game that holds a player for minutes is encoding your brand into the part of the experience they'll describe to a friend. You are no longer renting space next to attention. You are the thing attention is pointed at.
To achieve an average attention of 22 seconds is a significant breakthrough for our industry.— Niraly Jadeja, Lumen Research — on in-game ad engagement
Impressions vs. earned attention
@RELI
The math of earned attention
Put the formats side by side and the buy stops being a cost-per-impression decision and becomes a cost-per-minute-of-attention one. The table below uses the verified, sourced figures — no extrapolation, no rounding in our favor.
Notice what the engagement column is doing. A banner's best case is a fractional click. A video's defining behavior is the skip. A branded game's defining behavior is the session — measured in minutes, repeated by choice, and increasingly shared. That is not a cheaper impression. It is a categorically different asset.
This is no longer a niche audience
The reflex objection — "gaming is for a narrow demographic" — is a decade out of date. Newzoo counted 3.42 billion players worldwide in 2024, inside a games market worth $187.7 billion. That is not a subculture; it is roughly 40% of the planet, spanning every age bracket and spending bracket your media plan cares about.
The buy you've been making: 2,000 impressions to earn one click no one remembers. The buy that's now available: minutes of earned, recalled, shareable attention inside an experience that carries your brand. Same audience. Different physics.
3.42B players · a $187.7B audience
@RELI
Why @RELI
Knowing the arbitrage exists is the easy part. Capturing it means shipping an actual game — designed, built, instrumented, and live where your audience already is. That is what @RELI does: an AI studio that builds branded, revenue-ready games for Telegram, VK, and LINE Mini Apps, the open web (HTML5), Roblox, and Fortnite via UEFN. Not a media placement you rent for a quarter — an owned asset that compounds attention every time someone plays.
The traditional funnel asks you to interrupt people and hope they don't skip. A branded game inverts it: people come to you, stay for minutes, and leave remembering your name. The cheapest impression in the world is still worthless if no one sees it. The most valuable thing in marketing is the attention you actually earn.
Stop buying impressions. Start building attention.
The arbitrage is real and the data is in: minutes of earned play beat seconds of ignored ad. @RELI builds branded, revenue-ready games for Telegram, VK, LINE, the open web, Roblox, and Fortnite — owned assets that turn your audience's attention into something you keep. Let's build the game your competitors are still renting banners next to.
The Roblox Land Grab: What Brands' Visit Counts Reveal About Branded Games vs. Traditional Ads
@RELI·Field notes·6 min read
Vans built a virtual skatepark and 100 million people showed up. Nikeland pulled 21 million in under a year. Stack the real visit counts side by side and the verdict is hard to argue: branded games aren't a marketing experiment — they're the channel.
In September 2021, Vans did something no apparel brand had attempted at scale: it built a virtual skatepark inside Roblox and waited to see if anyone would show up. More than 100 million people did. They didn't watch a 15-second pre-roll and bounce. They customized avatars, designed skateboards, and shredded virtual bowls — logging over 8 million hours inside a branded experience that has no expiration date.
Two months after Vans opened its doors, Nike followed with Nikeland. By Nike's own Q3 FY22 earnings call, 6.7 million people had visited. By September 2022 — under a year in — that number hit 21 million visitors, and had been favorited by roughly 118,000 players. These are not ad impressions. They are visits. Sessions. Hours. The unit of measurement marketers have chased for a century, finally delivered at internet scale.
The question this raises isn't whether branded games work. The visit counts already answered that. The real question is why any brand still treats a virtual world as an experiment rather than a channel — and what the comparable numbers reveal when you stack them side by side.
The visit counts are an attention ledger
Traditional advertising buys borrowed attention. You rent a slot in front of content someone actually wanted, and you pay whether or not they look. Branded games invert the model: the brand becomes the content people choose. That distinction shows up brutally in the data. Vans World's 100 million visits weren't bought impression by impression — they compounded, because the experience is permanent and players come back.
100M+Visits to Vans World on Roblox
21MNikeland visitors in under a year
8M+Hours players spent inside Vans World
Compare the engagement profile. A skippable YouTube ad fights for five seconds. A billboard gets a glance at 60 mph. Inside these worlds, the median session runs minutes, not seconds — and Walmart Discovered alone has logged more than 2.1 million hours of play. That's not attention you interrupted. It's attention you earned, and it keeps accruing long after the launch budget is spent.
The branded land grab
@RELI
What the numbers look like side by side
Individually, each campaign is a press release. Together, they're a pattern. Every one of these is a real, sourced figure — luxury, apparel, big-box retail, and QSR all clearing tens of millions of visits from a single build.
Branded experience
Visits
Notable metric
Source
Vans World
100M+
8M+ hours logged
The Drum, Aug 2025
Gucci Town
40.6M+
Permanent, launched May 2022
Marketing Dive / WWD
Walmart Discovered
30M+
2.1M+ hours, 96% approval
Retail Tech Innovation Hub, Dec 2024
Chipotle Burrito Builder
25.6M+
Record digital sales day
Chipotle / The Metakey
Nikeland
21M+
~118K favorites
The Drum, Sept 2022
Gucci Garden
~20M
In a 2-week run (2021)
Roblox Wiki / Marketing Dive
Source: Compiled from The Drum, Marketing Dive, WWD, Retail Technology Innovation Hub, and Chipotle newsroom, 2021–2025.
Notice what's missing: a category ceiling. A skate brand, a luxury house, a discount retailer, and a burrito chain all cleared 20 million-plus. The platform doesn't care about your vertical — it cares whether you built something worth playing.
Why the audience math favors games now
The old knock on Roblox was that it was a babysitting app. That's no longer true, and the company's own filings prove it. In Q1 2025, Roblox reported 97.8 million daily active users and 21.7 billion hours engaged — and users aged 13 and over now make up 62% of DAUs and 64% of all hours. The 13+ cohort grew 36% year over year. The audience aged up while you weren't looking.
Roblox Q1 2025: the platform is now majority teen-and-older by both users and time spent.Source: Roblox Corp. Q1 2025 Shareholder Letter (SEC Form 8-K).
This matters because reach without the right age is vanity. The same quarter that Roblox crossed 97.8 million DAUs, it confirmed the bulk of engagement now comes from teens and adults with their own spending power — exactly the buyer a brand wants standing inside its store, not just walking past its banner.
The shift in one line: traditional ads rent attention by the impression and lose it the instant the budget stops. A branded game owns a permanent storefront that keeps drawing visits — and keeps converting them — for years.
Players don't just tolerate brands here — they prefer them
The skeptic's last defense is that engagement is hollow: people show up for the free stuff, not the brand. The data says otherwise. In a March 2024 study Roblox ran with Latitude across 2,100 U.S. users, 75% said they're more likely to notice brands that advertise on Roblox versus elsewhere, and 73% saw those brands as category leaders. Eighty-two percent said they appreciate brands that provide in-experience content.
75% of respondents said they're more likely to notice brands that advertise on Roblox versus elsewhere — and 73% see those brands as category leaders.— Roblox + Latitude brand study, March 2024
That's the inversion no pre-roll can buy. On most channels, advertising is a tax the audience tolerates. Inside a well-built game, the brand presence is the thing people came for — and it lifts favorability instead of eroding it. Chipotle proved the commercial end of this in 2022: its Burrito Builder experience drove a record digital sales day while racking up 25.6 million lifetime visits. Play became purchase.
Visits a media buy can't deliver
@RELI
The land grab is on — and most brands are late
Here's the uncomfortable part. Vans, Nike, Gucci, and Walmart didn't just run campaigns; they claimed permanent real estate in a world where 97.8 million people show up daily. Every visit they bank now is a visit a competitor can't have. This is a land grab, and the parcels are filling up. The brands stacking 20-, 40-, 100-million-visit numbers are building compounding assets while everyone else is still buying impressions by the thousand.
And it's no longer a single-platform bet. The same playbook now runs on Fortnite's UEFN, on the open web as HTML5, and inside Telegram, VK, and LINE Mini Apps — each with its own audience already in motion. The winning move isn't choosing one. It's owning a branded game wherever your customers already spend their hours.
The brands posting 20-, 40-, 100-million-visit numbers didn't get lucky — they shipped a game worth playing and have been compounding visits ever since. @RELI builds branded, revenue-ready games across Roblox, Fortnite (UEFN), the open web (HTML5), and Telegram, VK, and LINE Mini Apps — the channels where your customers already spend their hours. Let's build the experience that earns your visit count.
Why Brands Are Wasting $500K on Custom Fortnite Maps — And the Engagement Math That Fixes It
@RELI·Field notes·5 min read
Big brands keep paying $300K–$500K for bespoke Fortnite islands that peak sub-200 players. A $20–50K integration into an existing hit drove 721,000. Here's the engagement math behind the gap — and how to build on the right side of it.
A brand walks into Fortnite with a $400,000 budget and a beautiful idea: its own island. Custom architecture, custom quests, a logo on the loading screen. Six weeks of agency work later, the map ships. And then the dashboard tells the truth. Peak concurrent players: sub-200. Sometimes sub-100. Sometimes sub-50.
This is not a hypothetical. It is the dominant failure pattern in branded Fortnite right now, documented by the developers who build these maps and watch them empty out. The diagnosis is uncomfortable for any CMO who just signed a six-figure SOW: the most expensive thing you can build in Fortnite is the thing nobody can find.
The fix isn't a bigger budget. It's a different equation. And once you see Epic's engagement-payout math, the whole "build your own island" strategy starts to look like setting money on fire in a very well-rendered campfire.
The $500K island that 50 people visit
Per a Digiday investigation into Fortnite ad spend, a custom branded island runs $300,000 to $500,000 to build. For that, brands get an ownable world they fully control. What they often don't get is players. As multiple creators told Digiday, even the biggest brands on Earth routinely peak "sub-200, sub-100, sometimes sub-50" concurrent players on their bespoke islands.
The problem is structural, not creative. Fortnite has roughly 198,000 creator islands competing for attention. A brand-new map starts with zero discovery signal, zero retention history, and zero algorithmic momentum — and Fortnite's Discover ranking rewards exactly those things. You can buy a gorgeous island. You cannot buy your way out of a cold start.
$300K–$500KCost to build a custom branded island (Digiday)
sub-50Peak concurrent players some big-brand islands hit (Digiday)
198,000Creator islands competing for discovery in 2024 (Tubefilter / Epic)
A $500K ghost town
@RELI
The integration that 721,000 people actually played
Now run the other play. Instead of building a ghost town, Popeyes integrated into GoodGamers' already-popular "Murder Mystery" map — a Pickle Menu activation layered into a game players were already grinding. Integrations like this cost $20,000 to $50,000, roughly a tenth of a custom build.
The result, per Digiday: 721,000 unique players moved through the Popeyes integration. GoodGamers' own campaign reporting adds 62.1M+ impressions and 23.2M+ engagements. The brand didn't fight the discovery algorithm — it rented an audience that already existed.
Brands don't need to rely solely on PR or influencers. An exciting integration in a top game, paired with marketing, guarantees both attention and impact.— Zack Billingham, GoodGamers (via Digiday)
Custom island vs. integration: the real math
Approach
Build cost
Players reached
Cost per outcome
Custom branded island
$300K–$500K
Often sub-200 peak concurrent
Catastrophic
Integration into a hit map
$20K–$50K
721,000 unique (Popeyes / Murder Mystery)
~10x cheaper, ~1000x reach
Source: Digiday, "Fortnite advertising friction" (2025); GoodGamers campaign data
The disconnect isn't that brands are dumb. It's that they're measuring the wrong thing. As TMA's Sami Barnett told Digiday, brands chase "a full, ownable moment," while creators live and die by concurrent players. Neither side is wrong — they're just measuring different things. The brands optimizing for an ownable logo placement are paying island prices for billboard results.
Why integrations win: follow Epic's payout pool
To understand why the existing-hit strategy works, follow the money Epic itself moves. Under Creator Economy 2.0, Epic pools 40% of net Item Shop revenue into a single monthly engagement payout, then splits it across every published island by share of player activity (Music Business Worldwide). That one mechanism reveals what Fortnite actually rewards: not polish, not ownership — engagement, measured in retained playtime.
The flywheel it created is enormous. Epic confirmed at Unreal Fest in June 2026 that it has paid creators over $1 billion in engagement payouts since UEFN launched in 2023 (Tubefilter). In 2024 alone, players spent 5.23 billion hours in creator and brand islands — 36.5% of all Fortnite playtime — and Epic distributed $352 million for it. That share has since climbed toward 47%.
Where Fortnite's billion-dollar creator payout actually wentSource: Tubefilter (Jan 2025; June 2026), citing Epic Games
Here is the strategic punchline. When you integrate into a hit, you're plugging into a game that is already winning Epic's engagement-payout formula — already retaining players, already ranking in Discover, already pulling from that 40% pool. When you build a custom island, you're starting that race from a dead stop, alone, against 198,000 competitors. The math doesn't reward your logo. It rewards minutes played.
The lesson isn't "don't do Fortnite." It's don't build a destination — build inside a destination. Brands that treat UEFN like a billboard buy ownership and get silence. Brands that treat it like a game design problem buy engagement and get hundreds of thousands of players.
How @RELI builds for the engagement loop, not the logo
This is the exact problem @RELI is built to solve. We don't sell brands a beautiful empty island. We build branded games engineered around the engagement loop — the retention mechanics, the reasons-to-return, and the integration paths that make Fortnite's payout formula work for you instead of against you. The 40% engagement pool isn't a footnote in our pitch; it's the design target.
And because @RELI builds across Fortnite (UEFN), Roblox, the open web in HTML5, and Telegram, VK, and LINE Mini Apps, we pick the surface where your audience already lives — then build the game that keeps them there. Revenue-ready, engagement-first, and priced like a campaign that's supposed to work.
Build inside the engagement loop, not a $500K ghost town
@RELI builds branded, revenue-ready games engineered for the metric Fortnite actually pays for — retained playtime — across UEFN, Roblox, HTML5, and Telegram, VK, and LINE Mini Apps. Stop buying empty islands. Let's build the game your audience already wants to play.
The 950-Million-User Channel Hiding Inside a Chat App
@RELI·Field notes·6 min read
How Telegram Mini Apps became 2024's biggest acquisition engine — and why a branded game now beats the ad you were about to buy.
In July 2024, a tap-to-earn game about a cartoon hamster running a crypto exchange did something no marketing budget on Earth could buy. Hamster Kombat hit 300 million players — and it got there in roughly five months.
To put that in perspective: at the time, Telegram had 950 million monthly active users. Hamster Kombat had quietly captured nearly a third of an entire messaging platform's user base — not through app-store rankings, not through TV spots, but through a link you tap inside a chat you already had open.
That is the story most marketers missed in 2024. The biggest acquisition engine of the year wasn't a new ad network or a TikTok format. It was a games surface hiding inside an app a billion people already use every day. And the playbook that built it is now available to brands.
The channel that grew up overnight
Telegram crossed 1 billion monthly active users in March 2025, up from 950 million the previous July — and founder Pavel Durov noted the average user opens the app 21 times a day and spends 41 minutes inside it. For comparison, those are the engagement numbers of a social network, not a utility. People don't visit Telegram. They live in it.
In August 2024, Telegram turned that captive attention into a distribution platform. It launched a Mini App Store — a native, in-app catalog sitting between the Channels and Media tabs in search. At launch, Telegram reported that more than 500 million of its users already interacted with Mini Apps every month. No download. No App Store review. No 30% platform tax to Apple. Just a web app that opens inside the chat.
1BTelegram MAU (March 2025)
41 minAvg. time in app per day
500M+Monthly Mini App users at store launch
The chat app is the channel
@RELI
Anatomy of an explosion
Hamster Kombat wasn't a fluke — it was the second act. The pattern started with Notcoin, a deliberately dumb tap-the-coin game that launched January 1, 2024 and pulled in 35 million players in about three months, peaking above 6 million daily active users. Notcoin proved simple mechanics plus a token could onboard millions into a brand-new ecosystem at near-zero acquisition cost.
Then Hamster Kombat picked up the torch and ran further. Durov said the game had reached 239 million players by July 5, 2024; its team claimed 250 million days later; and by July 30 the whitepaper put it at 300 million. That's a growth curve no paid-media plan can replicate — because the distribution was the product. Every player was one tap away from inviting the next.
How fast Telegram Mini App games reached scale — reported player countsSource: Decrypt, The Block (2024)
The lesson isn't "make a crypto game." It's that a game built where attention already lives spreads at a velocity paid acquisition cannot match — because the channel, the install, and the share button are all the same surface.
The math traditional marketing can't win
Here's why this matters for any brand with a customer-acquisition line item. On the open app stores, getting a single install means paying for it — and those prices keep climbing. In 2024, the average cost per install for mid-core mobile games ran $4.50 on iOS; hardcore and RPG titles hit $6 on iOS. North America averaged $5.28 per install across categories. Every user is a toll you pay before they've even opened the app.
Inside Telegram, the dynamic inverts. There's no install to buy — the game opens in the chat — and where paid promotion exists, it's a fraction of app-store rates. One analysis pegged Telegram customer acquisition cost at 90–95% lower than traditional app stores, with Mini App ad views starting around $0.40 per thousand. The arithmetic is brutal for the old model.
Channel
Typical cost to acquire a user
Friction to first session
App Store paid install (iOS, mid-core game, 2024)
~$4.50 CPI
Download + open + onboarding
App Store paid install (iOS, RPG/hardcore, 2024)
~$6.00 CPI
Download + open + onboarding
North America avg. CPI (all apps, 2024)
$5.28
Download + open
Telegram Mini App (in-chat)
~90–95% lower CAC; ads from $0.40 CPM
One tap, no download
Source: Business of Apps / Mapendo CPI benchmarks 2024; RichAds Telegram Mini App pricing 2025
And the virality compounds the savings. A Mini App can drop a shareable leaderboard, screenshot, or AI-generated clip straight into a user's Telegram Stories — so every session has a built-in invitation back into the same network of a billion people. You're not buying the next user. The last one brings them.
More than 500 million people interacted with Mini Apps every month — before Telegram even built them a storefront.
What's real after the hype settles
Skeptics will note the tap-to-earn bubble deflated — token prices crashed and Mini App usage cooled from a September 2024 peak near 1.44 billion monthly actives down to a more durable 150–190 million by mid-2025. Fair. But that's the crypto-speculation layer correcting, not the channel disappearing. The underlying surface is bigger than ever: Telegram is past a billion users, roughly half of them touch Mini Apps, and the TON network behind them grew from ~4 million to 128 million accounts in about a year.
Strip away the tokens and what remains is the actual prize: a frictionless, viral, install-free games surface sitting inside one of the world's largest messaging apps — and the same mechanics work on VK, LINE, and the open web. The brands that win the next phase won't be chasing airdrops. They'll be shipping genuinely good branded games that turn a billion-person chat app into an owned acquisition channel.
950M users, one tap away
@RELI
Where @RELI comes in
Hamster Kombat proved the ceiling; the hard part is building the thing — a game polished enough that people choose to play it, instrumented to convert and retain, and engineered to live natively inside Telegram, VK, or LINE Mini Apps (or HTML5, Roblox, and Fortnite). That's the entire reason @RELI exists: an AI studio that builds branded, revenue-ready games as acquisition channels, not gimmicks.
The traditional funnel asks you to pay $5 a head to interrupt someone scrolling somewhere else. A branded game asks them to come to you — and bring their friends. That's not a better ad. It's a different category of marketing entirely.
Hamster Kombat reached 300 million players where a billion people already are. @RELI builds the branded, revenue-ready game that turns Telegram, VK, LINE, HTML5, Roblox, or Fortnite into your owned acquisition channel — not another line item on an ad invoice. Let's build the game that brings users to you.
The 100-Million-Player Web Nobody Markets On: Inside the Instant-Play HTML5 Game Boom
@RELI·Field notes·5 min read
Poki crossed 100M monthly players and 625M total in a year — built by ~65 people with no outside funding. The open-web HTML5 channel is the most under-priced distribution in entertainment, and it's the one @RELI ships to with a single click.
Here is a number that should rearrange how you think about reaching customers through games: 100 million people play on Poki every month, and not one of them downloaded an app to do it. They clicked a link. The game loaded. They played.
In March 2026, Poki announced it had welcomed 625 million players across 2025 — and that it hit 1 billion gameplays in a single month in June 2025. The company that did this employs roughly 65 people and has never raised a dollar of outside funding. For context, Poki's 100 million monthly players put it within shouting distance of PlayStation Network's 119 million.
This is the part of the games market almost nobody markets on — the open, instant-play HTML5 web. No app store, no install screen, no 4-gigabyte download. And for any brand trying to actually reach an audience instead of renting attention from one, it is the most under-priced distribution channel in entertainment.
The boom hiding in your browser tab
Web gaming spent a decade being written off as Flash-era nostalgia. Then mobile browsers got fast, HTML5 matured, and the numbers quietly went vertical. Bloomberg called web gaming the games industry's hottest new platform of 2025. The platform data backs the headline.
Poki isn't alone. Belgium's CrazyGames hit a record 35 million monthly users in 2024 — its best year since launching in 2013 — serving over 300 million gameplays a month across 4,000-plus titles. These platforms share one trait that the rest of gaming would kill for: zero friction between intent and play.
100MPoki monthly active players (2025)
1B+Poki gameplays in a single month (June 2025)
~65People who built it, with $0 outside funding
One click. Zero installs.
@RELI
The app-store tax nobody puts on the invoice
Every brand that has tried to reach people through a downloadable game has paid a hidden tax: the install funnel. Someone sees your ad, taps it, and then has to want your product enough to wait through an app store page, a download, and a sign-up. Most don't.
The benchmarks are brutal. According to Adjust's 2024 data, mobile games convert at an installs-per-mille (IPM) of just 4.99 — meaning roughly 5 installs for every 1,000 ad impressions. Even once a user lands on the store page, AppTweak pegged first-half-2024 conversion at about 25% on the App Store and 27% on Google Play. The funnel leaks at every step.
HTML5 deletes the funnel. There is no store page, no download bar, no permission prompt. The ad is the game, or one tap from it. The thing you're marketing and the thing the user experiences become the same object.
Reaching a player: app install vs. instant-play web
The reach isn't theoretical. SYBO's Subway Surfers, ported to the open web on Poki, has logged over 1.7 billion gameplay sessions and 21,000 combined years of playtime in seven years on the platform. When SYBO and Poki launched new Subway Surfers puzzle titles for the web in 2026, they crossed 100 million players globally on those alone.
Indie hits scale the same way. Monkey Mart, built by a small studio called TinyDobbins, became one of Poki's most-played titles — part of an ecosystem that generated 8.1 billion gameplays a year when Poki counted 500 million players. Top studios on the platform now earn up to €1 million annually. The open web isn't a consolation prize. It's a stage with stadium-sized audiences.
Monthly reach of instant-play web platforms vs. a console networkSource: Poki (2025), Sony PlayStation Network (2025), CrazyGames (2024)
Web gaming has undergone a renaissance thanks to advances in mobile browsers and HTML5 — titles start instantly on phones, tablets, or PCs, no installs needed.
Why this is a marketing channel, not a games story
Step back to the macro. Newzoo put the global games market at $187.7 billion across 3.42 billion players in 2024, rising to an estimated $188.8 billion and 3.6 billion players in 2025. Games are how more than 3 billion people spend their attention — more than any social network, streaming service, or sport.
Brands already know attention lives in games. What they've gotten wrong is the format. A 30-second pre-roll inside someone else's game rents you a sliver of attention you don't own. A branded game you own flips the equation: the player chooses to engage, stays for minutes instead of seconds, and the experience is your product — not an interruption before it.
The instant-play web is the only channel where your ad, your product demo, and your retention loop can be the exact same click. A 100M-MAU audience reaches it with zero install friction — and almost no brands are building for it yet.
Where @RELI comes in
This is precisely the gap @RELI is built for. We're an AI game studio that builds branded, revenue-ready games for businesses — and the open web (HTML5) is one of our native shipping targets, alongside Telegram, VK, and LINE Mini Apps, Roblox, and Fortnite via UEFN.
A @RELI web game is one link. Drop it in an ad, a QR code, a tweet, an email, a retail display — and it opens to play, on any device, no download. You get the reach of the 100-million-player web with a game that carries your brand, your offer, and your data — instead of paying the app-store tax to send players to someone else's storefront.
Poki proved 65 people can reach 625 million players on the open web. The audience is already there, already engaged, already one click away. The only question left is whether the game they play next is yours — or your competitor's.
Poki reached 625 million players on the open web with 65 people and no outside funding — proof the instant-play HTML5 audience is real, engaged, and one click away. @RELI builds branded, revenue-ready games that ship to that channel with a single link: no app store, no download, no friction between your ad and your product. Let's build yours.
Wendy's Declared War on Frozen Beef Inside Fortnite — and Beat Nike at Cannes
@RELI·Field notes·9 min read
With zero paid media, Wendy's turned a Fortnite game mode into a nine-hour product argument — 1.5M minutes watched, 23M+ impressions, a +4.2% sales lift, and the Cannes Grand Prix the jury said the year's best traditional ad didn't deserve. The cleanest proof yet that played attention beats interrupted attention.
Here is a number that should make every CMO with a nine-figure media budget uncomfortable: zero. That is roughly what Wendy's spent on paid media for the campaign that, seven months later, beat Nike's Dream Crazy to win the inaugural Social & Influencer Grand Prix at Cannes Lions. No banner. No pre-roll. No sponsored skin. Just a red-pigtailed avatar, a battle royale game, and nine hours of livestreamed freezer demolition.
On November 29, 2018, Wendy's didn't buy attention inside Fortnite. It earned attention by playing — and in doing so it ran the cleanest real-world experiment we have for the thesis @RELI exists to prove: played attention beats interrupted attention. This is the story of how a quick-service burger chain turned a video game's own lore into a product argument, reached the exact generation that blocks ads, and walked out of the South of France with the most coveted prize in advertising.
The setup: a free game mode that wrote Wendy's ad for it
In late 2018, Fortnite was not a game. It was a country. Epic Games confirmed 200 million registered players in November 2018 — up roughly 60% from the 125 million it had reported just five months earlier in June. For a brand chasing young diners who had largely abandoned cable, this was the largest unmonetized attention pool on the planet.
Then Epic handed Wendy's a gift it could not have written better itself. Update v6.30 introduced a limited-time mode called Food Fight: two teams of 50 players, one defending a giant Durr Burger mascot, the other a Pizza Pit tomato. Here was the detail that mattered: in Fortnite's own canon, Durr Burger stored its beef in freezers scattered across the map. Wendy's entire brand — its single most-repeated claim since 1969 — is "fresh, never frozen" beef. The game had, by accident, built a villain that violated Wendy's core positioning. All Wendy's had to do was show up and play the hero.
The native-advertising move in one sentence Wendy's did not interrupt the game to deliver its message — it found the place where the game's existing story already disagreed with a competitor and let players watch the brand win that argument as gameplay.
The execution: nine hours, 752 freezers, one avatar that looked exactly like the logo
Wendy's created a custom Fortnite avatar — red hair, pigtails, unmistakably the brand's namesake mascot — and dropped into Food Fight on Team Pizza — the side pitted against Durr Burger. But it ignored the combat objective entirely. Instead of battling rival players, the Wendy's character roamed Durr Burger's restaurants and systematically smashed every freezer it could find, refusing to so much as fire at another player. The brand streamed the entire thing live on Twitch for nine straight hours, narrating the rampage as improv comedy and rallying its notoriously sharp Twitter account to cheer the mission on in real time.
The genius is in what's absent. There was no streamer-celebrity buying the audience. As VMLY&R group creative director McKay Hathaway put it, the brand "made a huge impact on Twitch without arguably the most important part of Twitch: the presence of the streamer." The freezer-smashing was the advertisement — indistinguishable from emergent play, impossible to skip, and impossible to ad-block, because it was content the audience had chosen to watch.
Brands are a little intimidated about how to step into the gaming space.— Kristin Tormey, Social Media Manager, Wendy's
The freezer rampage
@RELI
A red-pigtailed Wendy's avatar standing over a smashed Durr Burger freezer mid-stream — the ad that was indistinguishable from play.
9 hrsContinuous Twitch livestream, Nov 29 2018
752In-game freezers destroyed on stream
$0Paid media spend
The results: a funnel built entirely from earned attention
Because the campaign cost effectively nothing to run, every metric it produced is pure earned media — the kind brands normally pay dearly to manufacture. The Twitch stream drew more than 250,000 live views. Across the nine hours, the freezer-smashing amassed more than 1.5 million minutes watched. The campaign generated more than 23 million impressions, and mentions of Wendy's rose 119% across social platforms.
None of that would matter to a finance team if it didn't move product. It did. The campaign is credited with a +4.2% lift in North American sales — the rare in-game brand activation with a documented hard-sales outcome, not just a vanity-metric afterglow.
Fortnite's audience when Wendy's arrived
Fortnite grew from 125M registered players in June 2018 to 200M by November 2018 — the largest unmonetized attention pool on the planet.Source: Epic Games
Campaign scorecard at a glance — Every figure traces to a named industry source; metrics marked '+' are reported as minimums.
Metric
Result
Source
Livestream date
Nov 29, 2018 (9 hours)
Adweek / Warc
Live stream views
250,000+
Digital Training Academy
Minutes watched
1,500,000+
Warc / VMLY&R
Total impressions
23,000,000+
Warc
Lift in brand mentions
+119%
Digital Training Academy
In-game freezers destroyed
752
Digital Training Academy
North American sales lift
+4.2%
Unlocked Meta case study
Paid media spend
~$0
Warc
Source: Warc, Digital Training Academy, Adweek, Unlocked Meta
The verdict: a Cannes jury said it beat the best traditional ad of the year
Seven months later, at Cannes Lions 2019, "Keeping Fortnite Fresh" won the first-ever Social & Influencer Grand Prix. Its competition in that category included Nike's Dream Crazy — the Colin Kaepernick film widely regarded as the year's defining piece of advertising, which took two Gold Lions in the same category. The jury still chose the burger chain's freezer stunt over it.
Jury chair PJ Pereira was explicit about why, and his reasoning is the whole argument for native in-game marketing compressed into two sentences:
What Nike did with 'Dream Crazy' is possibly the best example of purpose-driven marketing ever done. But Cannes should also be about what the next trend should be… It was setting up a new trend instead of being the apex of a previous trend.— PJ Pereira, Social & Influencer Jury Chair, Cannes Lions 2019
Read that again through @RELI's lens. The most prestigious jury in advertising looked at a flawless, big-budget, celebrity-driven broadcast ad and a zero-budget brand pretending to be a player inside a video game — and declared the game the future. The campaign went on to win across multiple categories (the Grand Prix plus additional Lions per Ad Age's coverage) and took Best in Show at the In2 SABRE Awards. The trophy case confirmed what the sales line already proved.
The trophy that broke the rule
@RELI
The Social & Influencer Grand Prix — awarded to a video-game stunt over the year's most acclaimed traditional ad.
Why this is the cleanest proof of the @RELI thesis
Strip the campaign to its mechanics and you get a checklist that almost no marketing case in history satisfies all at once:
Native mechanic — Wendy's used Fortnite's own mode, map, and lore. The freezers were already in the game; the brand just played its values out loud.
Genuine entertainment — 1.5 million minutes watched means people chose to spend time with the brand. You cannot buy that the way you buy a 15-second skip-able pre-roll.
Ad-blocker-proof reach — the audience was Gen Z and young millennials inside Twitch and Fortnite, the precise cohort that has trained itself to ignore, skip, and block interruptive advertising.
Hard sales lift — +4.2% in North America, not just sentiment.
A jury verdict — Cannes explicitly ranked it above the year's best traditional ad.
The arbitrage Wendy's reached a 200-million-player audience, generated 23M+ impressions, and lifted sales — for roughly the cost of nine hours of staff time. That is the gap between what interruptive media costs and what played attention costs. The brands that build inside games are buying the same attention at a fraction of the price.
There is, however, an honest caveat — and it points straight at the opportunity. Wendy's pulled this off as a one-time, reactive stunt that depended on Epic happening to ship a game mode that contradicted a competitor. It was lightning in a bottle. The brand was a guest in someone else's world, on someone else's schedule, with no control over the rules, the canon, or whether the moment ever came again.
From borrowed stunt to owned world — what @RELI builds
"Keeping Fortnite Fresh" worked because the brand's argument became playable. But Wendy's had to wait for a competitor's game to accidentally hand it a stage. Imagine running that play on demand — a branded game where your product truth is the core mechanic, your world is the canon, and the freezer-smashing moment isn't a lucky coincidence but the entire design.
That is exactly what @RELI builds: branded games and native in-game marketing where the brand isn't a banner bolted onto play, but the reason the play exists. Wendy's proved the ceiling of borrowing someone else's game for one afternoon. The next Cannes Grand Prix belongs to the brand that owns the game — and gets the 1.5 million minutes, the 23 million impressions, and the sales lift on repeat, not by accident.
Wendy's needed a competitor's game mode to accidentally stage its product story. @RELI builds branded games where your brand's truth is the core mechanic — on demand, on your terms, engineered for the watch-time, the impressions, and the sales lift that Wendy's caught once by luck. Let's design the world your audience chooses to play in.
America's Army: The $5M Game That Out-Recruited Every Other Army Ad Combined
@RELI·Field notes·9 min read
In 2002 the U.S. Army gave away a free video game. Twenty years and 260 million played hours later, it had reached young recruits at 10 cents an hour — versus $5-8 for TV — and, by its own director's account, moved more of them toward enlistment than every other form of Army advertising combined. The clearest proof on record that played attention beats interrupted attention.
In 2005, a U.S. Army colonel named Casey Wardynski did something almost no marketer ever does in public: he put a number on attention. The free first-person shooter his team had built and given away, he said, was reaching young Americans at a cost of roughly 10 cents per person-hour of engagement — versus $5 to $8 per person-hour for television. That is a 50-to-80x efficiency gap, disclosed not by an agency pitching a renewal but by a government program with hard accountability and its own recruits on the line.
The game was America's Army. It launched on July 4, 2002, ran for almost exactly twenty years, and by the time its servers went dark in May 2022 it had logged the kind of numbers that make modern ad-buyers wince. This is the bedrock case for everything @RELI argues to CMOs: when the brand becomes the thing people choose to do, attention stops being something you rent by the impression and starts being something you earn by the hour. America's Army is the oldest, cleanest proof that a branded game can out-perform every interruptive channel on effectiveness AND on cost — at the same time.
The number that should end the debate
Start with the headline comparison, because it reframes the entire economics of advertising. A 30-second TV spot buys you a fragment of divided attention — a viewer half-watching while they check a phone, fast-forward, or leave the room. America's Army bought something categorically different: voluntary, sustained, hours-long engagement inside an Army-built world. Wardynski's own framing was that the program spent on the order of a million dollars a year and delivered a cost per person-hour of about 10 cents.
Cost per person-hour (USD)
Cost per engaged person-hour: America's Army vs. television, as disclosed by the program's director in 2005. The game's hourly engagement cost was roughly 50-80x cheaper than TV's.Source: GameSpot, 2005 (Col. Casey Wardynski)
That is not a brand-lift study or a modeled estimate. It is a per-hour exchange rate between two media, published by the person whose budget and recruiting targets depended on getting it right. And the comparison held up over time: in a 2010 military report, Wardynski and two co-authors went further, writing that the game could reach prospective soldiers at a cost 10 to 40 times cheaper than traditional recruitment advertising.
Why this case anchors the @RELI thesis Almost every advertising case study measures reach or recall — things that happen TO an audience. America's Army measured hours willingly spent and bodies that showed up at a recruiting station. It is the rare campaign where the brand became the activity, attention was earned rather than interrupted, and the organization put its own attribution on the record.
A government built a video game because the old ads stopped working
The origin story matters because it rhymes with what brands face now. By the late 1990s the Army was missing recruiting targets, and the channels it had always relied on — TV, race-car sponsorships, bull-riding, print — were reaching a generation that had stopped paying attention to them. Wardynski's insight was that you could not buy your way back into that audience with louder interruptions. You had to go where they already were, doing what they already loved to do.
So the Army did something radical for a government bureaucracy: it hired professional game developers. Wardynski took the idea to Michael Zyda and Michael Capps at the MOVES Institute at the Naval Postgraduate School, who assembled a studio of veterans from major commercial titles and built the game on the Unreal Engine. The development budget was about $5 million — a rounding error against the Army's overall marketing spend, and a sum that would barely fund a single national TV flight.
July 4, 2002: the U.S. Army ships a free Unreal-engine shooter on Independence Day
@RELI
The distribution model was equally ahead of its time. America's Army was free — downloadable from the Army's website or handed out on CD at recruiting stations — years before free-to-play became the default. There was no paywall, no pre-roll, no banner. The product was the message. And it worked fast: the game racked up 1.5 million downloads in its first six months and quickly became one of the ten most-played online first-person shooters in the world.
260 million hours: what twenty years of played attention looks like
Run a free game for two decades and the engagement compounds in a way no ad campaign ever does — because the audience keeps coming back of their own accord. By 2014 the franchise had passed 13 million registered accounts and more than 260 million hours played across its many versions. By the 2022 shutdown, the Army counted over 20 million total players across the series.
Cumulative hours played (millions)
Cumulative hours played on America's Army, from the Army's own milestone disclosures. The PC version alone had crossed 230.9 million hours by August 2008; the franchise passed 260 million across all titles by 2014.Source: U.S. Army (Guinness World Records, 2009); Wikipedia (America's Army)
The scale was enough to set records. In the Guinness World Records 2009: Gamer's Edition, America's Army claimed five marks at once, including Most Downloaded War Video Game (42.6 million downloads across versions) and Most Hours Spent Playing a Free Online Shooter. When the game hit its 8-millionth registered user in January 2007, Guinness noted the 'Largest Virtual Army' was roughly 15 times the size of the real Army's active-duty force at the time.
13M+registered player accounts (by 2014)
260M+hours played across all versions
42.6Mtotal downloads (Guinness, 2009)
Translate 260 million hours into the language of advertising and the number becomes almost absurd. At broadcast rates of $5-8 per attentive hour, buying that much genuine engagement on television would have cost on the order of $1.3 to $2 billion. The Army got it for an annual program budget measured in single-digit millions.
America's Army by the numbers — Verified lifetime metrics for the America's Army franchise (2002-2022), compiled from the U.S. Army, Guinness World Records, and Wikipedia.
Metric
Value
Source
Launch date
July 4, 2002
Wikipedia
Final shutdown
May 5, 2022
Vice / Wikipedia
Registered accounts
13M+ (by 2014)
Wikipedia
Total hours played
260M+
Wikipedia
Total downloads
42.6M (by 2009)
Guinness / U.S. Army
Total players (series)
20M+
Vice
Development budget
~$5M
Wikipedia
Versions/updates
41+ (by Jan 2014)
Wikipedia
Source: U.S. Army; Guinness World Records 2009; Wikipedia; Vice
The part that matters: it actually moved recruits
Reach is cheap to claim. Conversion is where most branded entertainment quietly dies — and where America's Army separates itself. The Army didn't just measure hours; it measured bodies walking into recruiting stations, and it kept measuring for years.
By 2005, internal data showed that somewhere between 20% and 40% of new recruits had played the game — and that the figure ran to 20% of incoming cadets at West Point, the Army's most selective pipeline. Wardynski's blunt summary, repeated across reporting, was that the game had more impact on recruits than all other forms of Army advertising combined.
It's designed to give them an inside view on the very fundamentals of being a soldier, and it's also designed to give them a sense of self-efficacy — that they can do it.— Col. Casey Wardynski, America's Army program director (2005)
Third-party measurement backed the internal claims. A 2008 study from MIT found that roughly 30% of all Americans aged 16 to 24 held a more positive impression of the Army because of the game — and, again, that it had more impact on recruits than all other Army advertising combined. Other research into the recruiting funnel found that a meaningful share of new recruits had played the game heavily in the weeks before enlisting, and that young men who played at recruitment fairs came away markedly more favorable than those who didn't.
Share who had played America's Army
Penetration of America's Army among incoming Army populations, per program data reported in 2005. The game had reached a fifth to two-fifths of new recruits — including a fifth of the West Point class.Source: GameSpot, 2005 (Col. Casey Wardynski)
Played attention converts. Interrupted attention leaks. A TV viewer who half-remembers a slogan is a weak signal. A teenager who spent forty hours practicing teamwork, radio discipline, and rules of engagement inside an Army-built simulation is a qualified lead who pre-sold themselves. That is the difference between renting attention and earning it — and it shows up downstream, in the recruits.
It wasn't free, and it wasn't uncontroversial — that's the point
Honesty strengthens the case. The 'free' game was not costless to the institution that built it. Reporting later pegged the total taxpayer investment at roughly $32.8 million spread over about a decade — a real line item, and one critics challenged on ethical grounds, arguing the Army was marketing combat to minors. The program drew sustained criticism from anti-militarization groups and academics throughout its run.
But scrutiny is exactly what makes the numbers credible. This was not a brand grading its own homework in a glossy case study. It was a government program, audited, debated in public, and forced to defend its spending — and the per-hour and per-recruit economics survived twenty years of that pressure. Even taking the full $32.8 million against 260 million-plus hours, the implied cost lands well under a cent and a half per hour of engagement. By any honest accounting, played attention was an order of magnitude cheaper than the alternative.
$32.8M over a decade, 260M+ hours earned — roughly a penny per hour of attention
@RELI
There is also a generational lesson buried here. The audience the Army most needed — young men aged 16 to 24 — was precisely the cohort least reachable by, and most skeptical of, conventional advertising. The game reached them not by interrupting the thing they loved but by becoming it. That is the structural advantage of native, played media: it works on the people who have already opted out of everything else.
What a 2002 government project tells a 2026 CMO
Strip away the camouflage and America's Army is a brand campaign — arguably the most rigorously measured branded-game campaign in history. A brand with a hard objective (recruiting), a skeptical young audience, and accountability for every dollar chose to build a game instead of buying more ads. The result was cheaper attention, deeper engagement, and measurable downstream conversion that its own leadership credited above all other channels combined.
The platforms have changed — today the audience lives in Roblox, Fortnite, and Discord rather than on a CD from a recruiting station — but the physics are identical. Interrupted attention is rented by the impression and forgotten in seconds. Played attention is earned by the hour and converts. The America's Army numbers are two decades old and still the cleanest proof of it on record.
$0.10per person-hour (game) vs. $5-8 (TV)
20-40%of 2005 recruits had played it
30%of 16-24-year-olds viewed Army more positively (MIT, 2008)
The question for any brand sitting on a TV-and-pre-roll budget is no longer whether a branded game can work. A government org settled that twenty years ago, in public, with its own recruits as the scoreboard. The question is whether you'd rather keep paying $5-8 for a divided half-second — or own the hours.
Stop renting attention by the impression. Own it by the hour.
America's Army proved the model in 2002 with a $5M game, 260M+ played hours, and recruits who pre-sold themselves. @RELI builds that same advantage for brands — native games and in-game marketing on the platforms your audience already lives in, measured on engaged hours and conversion, not impressions. Let's build the game that out-performs your media plan.
Gucci's Roblox Empire: The $4,115 Handbag Made of Pixels
@RELI·Field notes·9 min read
A digital-only Gucci bag that launched for $5.50 resold for $4,115 — more than the physical original. The frenzy wasn't a glitch. It was the clearest proof yet that played attention beats interrupted attention, and that a brand can be wanted instead of skipped.
In May 2021, a handbag that cost about $5.50 sold for $4,115. It was not leather. It was not stitched. It could not be carried, insured, or resold for cash. It existed only as pixels inside a video game played mostly by teenagers. And it sold for $715 more than the physical Gucci bag it was modeled on, which retailed for $3,400.
That single transaction — a digital-only Dionysus Bag with Bee changing hands for 350,000 Robux on Roblox's secondary market — is the cleanest proof point in modern marketing for a thesis the entire ad industry keeps trying to ignore: people will pay a brand for attention they choose, while they pay nothing for attention they're forced to give. The same Gen Z audience that installs ad blockers and skips pre-roll in under two seconds spent real money, and waited in line, to wear Gucci inside a game. The advertising was the product. The product was the advertising.
The number that broke the rules
Start with the artifact itself. The Queen Bee Dionysus bag debuted inside the Gucci Garden experience as a limited drop — available for purchase during just two one-hour windows, one hour on May 17 and one on May 18, 2021. The initial price was 475 Robux, roughly $5.50. Scarcity did what scarcity does. Within hours, scalpers were listing the bag on Roblox's resale market for as much as 1 million Robux (about $10,000), and at least one buyer paid 350,000 Robux — approximately $4,115.
The detail that makes industry veterans flinch: Roblox players cannot cash out Robux for dollars. There is no resale arbitrage, no flip-for-profit endgame. The buyer wasn't speculating on an asset. They wanted to own a Gucci bag inside the world where their identity actually lives — and they valued that more than the physical equivalent. As Reddit co-founder Alexis Ohanian noted at the time, this purse 'is not an NFT and thus has no value/use/transferability outside the Roblox world — yet it's worth more than the physical one.'
The virtual bag outsold the real one
A digital-only Dionysus bag that originally sold for ~$6 resold for $4,115 on Roblox's secondary market — more than the $3,400 physical bag.Source: The Fashion Law / NME / Hypebeast (2021)
The contrarian read Every dollar of that $4,115 flowed TOWARD the brand's world, not away from a media budget. Traditional advertising pays platforms to rent a sliver of someone's attention. Here, consumers paid to enter, paid to wear, and competed to keep paying. Inverting who-pays-whom is the entire game.
The venue: 199 million reasons this wasn't a stunt
Gucci didn't pick Roblox for novelty. It picked it for reach that no fashion campaign could otherwise buy. In Q1 2021, the quarter Gucci Garden launched, Roblox reported 42.1 million daily active users and roughly 199 million monthly active users, with daily engagement of 9.7 billion hours — up 98% year over year. Critically for a brand chasing the next generation of luxury buyers, 67% of Roblox users were under 16, and the average user spent about 156 minutes per day — two and a half hours — inside the platform.
Put that against the media reality Gucci was escaping. The audience it most wanted to reach is the same cohort that has structurally opted out of interruptive advertising: ad blockers, skip buttons, sub-second banner glances, TikTok scroll velocity. You cannot buy your way into a 14-year-old's attention with a pre-roll. But you can build a place they choose to spend two and a half hours a day, and put your name on the walls and on their avatar.
42.1MRoblox daily active users (Q1 2021)
~199MMonthly active users at launch
156 minAvg. daily time on platform
Roblox annual Robux spend — a built-in willingness-to-pay engine — Gucci entered a platform where spending on virtual goods was already accelerating into the billions — a population pre-conditioned to pay for digital self-expression.
Period
Consumer spend on Robux
2018
$499 million
2019
$694.26 million
2020 (first 9 months)
$1.24 billion
Source: The Fashion Law, citing Roblox financials (2021)
gucci_garden_rooms
@RELI
The seven-room virtual exhibition where visitors shed their avatars and absorbed the environment's textures.
The two-week stunt that 20 million people walked through
Gucci Garden ran from May 17 to May 31, 2021, a digital companion to the physical centenary exhibition in Florence celebrating Gucci's 100th anniversary. It was not a billboard or a sponsored level. It was a destination — seven themed rooms drawn from Gucci's archival collections. On entry, visitors shed their avatars and became neutral mannequins that absorbed the colors, textures, and patterns of each room as they moved through, leaving as a one-of-a-kind composite of the experience.
The result: roughly 20 million visitors in two weeks. The experience went on to win a Webby Award. And across its 31 purchasable items, Gucci Garden generated 286 million Robux in initial collectible sales — before a single secondary-market resale. This is the part most coverage glosses: the campaign was not a cost center subsidized by a marketing budget. The audience funded it.
There is probably an underestimation of the value attached to individuals who want to express themselves in a virtual world, with a virtual product, through a virtual persona.— Robert Triefus, then EVP of Brand & Customer Engagement, Gucci
From stunt to owned media: the Gucci Town pivot
The genius wasn't the viral resale headline. It was what Gucci did next. Rather than banking a two-week PR spike and moving on, the brand made the bet permanent. In May 2022, it opened Gucci Town — a persistent, owned space on Roblox with mini-games, a Creative Corner, an exhibition plaza, a digital shop, and its own in-experience currency, GG Gems. A limited stunt became standing real estate inside a 200-million-user platform.
And the engagement numbers held up under permanence. According to a peer-reviewed case study by Heo et al. (published in Thunderbird International Business Review), Gucci Town accumulated 47 million visits, an average session duration of 7.56 minutes, and more than 6 million cumulative hours of audience engagement. Triefus framed the ambition plainly: he expected the virtual world to become 'a very significant new revenue stream' — not a marketing line item, but a business.
The Garden-to-Town arc: a 20M-visitor two-week stunt converted into a permanent property that compounded to 47M cumulative visits — the difference between renting attention once and owning the room.Source: Roblox / Heo et al., Thunderbird International Business Review (2024)
7.56 minutes vs. a sub-second glance
Here is the comparison that should reframe every media plan. The industry's accepted standard for a 'viewable' display impression is that 50% of the ad is on screen for one second. One second — and even that is routinely missed, skipped, or blocked. Gucci Town's average session was 7.56 minutes: roughly 453 seconds of chosen, active, leaned-in engagement with the brand as the environment itself.
That is not a 450x improvement in a vanity metric. It is a categorical change in what attention means. A banner glance is interruption tolerated. Seven and a half minutes inside a branded world is interruption inverted — the brand stopped being the thing between the user and what they wanted, and became the thing they wanted.
A glance vs a 7.5-minute visit
Engaged attention, compared. A Gucci Town session delivers multi-minute, opt-in brand time; the industry's minimum 'viewable' display impression is a single second — usually less in practice.Source: Heo et al. (2024) for Gucci Town; IAB/MRC viewability standard for display
Gucci's Roblox era, by the numbers — Verified figures from the Gucci Garden (2021) and Gucci Town (2022) Roblox activations.
Metric
Value
Source
Gucci Garden run
May 17–31, 2021 (2 weeks)
Roblox / Dezeen
Gucci Garden visitors
~20 million
Vogue Business / Roblox
Garden initial item sales
286 million Robux
Roblox Wiki
Dionysus bag launch price
475 Robux (~$5.50)
The Fashion Law
Dionysus bag resale peak
350,000 Robux (~$4,115)
The Fashion Law / NME
Physical bag retail
$3,400
The Fashion Law
Gucci Town visits
47 million
Heo et al. (2024)
Gucci Town avg. session
7.56 minutes
Heo et al. (2024)
Gucci Town cumulative hours
6 million+
Heo et al. (2024)
Source: Compiled from The Fashion Law, Dezeen, Roblox financials, and Heo et al. (2024)
What the resale frenzy actually proved A virtual good can generate MORE genuine desire than its physical twin. That is the precise opposite of ad avoidance. When demand for a branded item exceeds demand for the real product — at a price the buyer can never recoup — the brand has stopped advertising and started being wanted.
gucci_avatar_wearables
@RELI
Avatars wearing Gucci as self-expression — the brand worn by choice, not shown by force.
Why this is the @RELI playbook
Gucci's Roblox arc is the clearest case study for the model @RELI builds for brands: the advertising is the destination. No pre-roll, no skippable mid-roll, no banner fighting for a viewable second. Players chose to enter a Gucci world, chose to play, chose to customize their avatars, and chose to pay the brand for the privilege. Attention was earned, not interrupted — and it reached precisely the audience that has structurally walled itself off from traditional ads.
The lesson is not 'go make a one-off metaverse stunt.' It is the Garden-to-Town arc: a viral activation is a spark, but a permanent, owned, playable property is the asset. The first earns headlines; the second compounds — 20 million visitors became 47 million, two weeks became years, and a marketing expense became a revenue stream. That is what a branded game is: media you own, audiences who opt in, and attention measured in minutes, not milliseconds.
~750xResale price vs. its own launch price
453xEngaged session vs. ad 'viewable' second
20M→47MStunt to permanent: visitor growth
Gucci proved the thesis on someone else's platform, with a borrowed audience. The next move is to build it as your own — a game that is the brand, an audience that pays you in attention and intent, and a property that appreciates every quarter it stays live. That's the difference between interrupting attention and owning the place attention chooses to go.
Gucci turned a two-week stunt into a permanent property and let its audience pay for the attention. @RELI builds branded games that do the same for you — owned media, opt-in audiences, attention measured in minutes. Let's design the destination your customers want to live in.
Hellmann's Turned Animal Crossing's Spoiled Turnips Into 100,000 Real Meals
@RELI·Field notes·9 min read
A 100-year-old mayonnaise brand built an island inside a video game, asked players to donate their rotting virtual turnips, and posted a 46.7% engagement rate — roughly 100x a banner ad. Here's the anatomy of native advertising done right, and why it changed 84% of its audience's real-world behavior.
In December 2020, a brand of mayonnaise reached 54.8 million people and changed the behavior of most of the ones it measured — without buying a single second of interruptive media on the channel that mattered. It did it by building an island.
That number is worth sitting with. Hellmann's, the 100-year-old condiment owned by Unilever, ran a campaign inside Animal Crossing: New Horizons that posted a 46.7% engagement rate. The average display ad earns a click-through rate of roughly 0.46%. That is not a typo and it is not a rounding difference. Hellmann's was, on a per-impression basis, roughly 100 times more engaging than the banner ad that funds most of the open web. The difference wasn't budget. It was the decision to stop advertising at gamers and start playing alongside them.
The thesis in one line Interrupted attention is rented and resented. Played attention is earned and remembered. Hellmann's didn't buy a slot inside Animal Crossing — it became a place players chose to fly to.
The food-waste problem that was already inside the game
To understand why this worked, you have to understand a quirk of Animal Crossing that millions of players were already living with. The game contains a feature fans call the Stalk Market: every Sunday morning a character named Daisy Mae sells white turnips in bundles of ten for 90–110 Bells each. Players hoard them, hoping to sell at a profit later in the week. The catch is brutal and unforgiving — turnips spoil the following Sunday. Miss your window and the entire investment rots into worthless, fly-swarmed garbage that the in-game shop won't touch.
So here was a game with tens of millions of players, all of whom periodically generated a pile of virtual food waste they couldn't do anything with. For most brands that's trivia. For Hellmann's — a brand whose entire modern purpose is its "Make Taste Not Waste" platform, built on the FAO's finding that roughly a third of all food produced for humans is lost or wasted, with most household waste happening at home — it was the most on-brand creative brief imaginable, and it was sitting in plain sight inside a video game. Ogilvy and Edelman didn't invent a metaphor. They found one the audience had already built.
A rotting pile of virtual turnips beside a real plate of food — the exchange rate at the heart of the campaign.
@RELI
What Hellmann's actually built
Hellmann's created Hellmann's Island — billed as the first branded 5-star Animal Crossing island, the game's highest possible rating. Players opened their game, flew their character to the island via the in-game Dodo Airlines, and dropped off their spoiled turnips. For every rotten turnip a player donated, Hellmann's converted it into a real meal for a food bank. Virtual waste in, real-world food out.
The mechanic ran across three flights. The original Canadian activation in August 17–22, 2020 partnered with Second Harvest and set a goal of 25,000 meals. It worked well enough that the team brought the island back twice — a Halloween edition, and a Christmas edition in December 14–18, 2020 built out with a Santa's Grotto, an orchard, and a winter spa, this time tied to UK food-redistribution charity FareShare. Across the campaigns, Hellmann's donated 100,000 real meals.
Note what is absent here. There is no pre-roll. No logo slapped on a loading screen. No "sponsored by" overlay. The advertising is the gameplay loop — fly somewhere, drop something off, get a reward. A player donating turnips on Hellmann's Island was doing the exact same set of actions they'd do on any friend's island. The brand didn't interrupt the game; it added a destination to it.
The timing: a 26-million-player audience nobody could reach with TV
This landed at a singular moment. Animal Crossing: New Horizons launched March 20, 2020 — the same week much of the world locked down. It sold 11.77 million copies in 12 days, hit 22.4 million units by June 30, 2020, and reached 26.04 million by September. SuperData clocked five million digital copies in March alone, the biggest single-month digital console launch in history at the time.
Animal Crossing: New Horizons cumulative worldwide sales in 2020 — the audience Hellmann's stepped into.Source: Nintendo earnings via Nintendo Life & VGC, 2020
That audience skewed exactly toward the cohort traditional media struggles to buy: younger, mobile-first, ad-blocking, social-cause-attentive consumers who, as Marketing Dive noted at the time, "tend to favor brands that support social movements." You cannot reliably reach these people with a 30-second spot. They had migrated into a game. Hellmann's followed them in and spoke their language instead of dubbing its own over the top.
Why this is native advertising in the truest sense Native advertising isn't a banner that matches the font of the page. It's a brand presence that's indistinguishable from the thing the audience already came to do. Hellmann's Island passed the only test that matters: players would have flown there even if they didn't know a brand was behind it.
The numbers: what 'earned' attention looks like
The campaign's results read like a category error when you line them up against paid media. 66.1 million impressions. 54.8 million reach. 6.8 million engagements. A 46.7% engagement rate. On Twitter, players didn't just see the campaign — they messaged the brand asking for a tour: over 5,000 DMs requesting access, with 1,100 arriving in the first hour. More than 15,000 people voluntarily tagged Hellmann's and its hashtag to show off their visit. The campaign hit Reddit's front page and landed on Twitter's #BestofTweets2020.
Hellmann's Island engagement rate vs. typical digital ad click-through rates. Native, played attention versus rented, interrupted attention.Source: AToMiC Awards 2021 (Hellmann's); Smart Insights / CXL display benchmarks
Read that chart slowly. The thing brands spend billions on — display — converts attention at a fraction of one percent. The thing Hellmann's built converted nearly half of the people it touched into active participants. When your audience is DMing you 1,100 times an hour to be let in, you have inverted the entire economics of advertising. You are no longer paying to intrude; people are lining up to be reached.
Hellmann's vs. Virtual Food Waste — verified results — Reported campaign results for Hellmann's vs. Virtual Food Waste.
Reach and engagement are the easy wins. The hard one — the one almost no campaign can credibly claim — is changing what people actually do. Hellmann's reported that 84% of the target audience said the experience motivated them to reduce food waste in their own homes. Not "recalled the ad." Not "viewed favorably." Motivated to change real behavior, off-screen, in their own kitchens.
Self-reported attitudinal shift: share of the target audience who said Hellmann's Island motivated them to cut food waste at home.Source: AToMiC Awards 2021 (Hellmann's vs. Virtual Food Waste)
This is the payoff of played attention. A player who spends ten minutes flying to your island, walking around, and personally dropping off turnips to fund a meal has done something. They've performed the brand's purpose with their own hands. Behavioral science is blunt about this: action changes belief far more reliably than messaging does. Hellmann's didn't tell people food waste matters; it let them experience the conversion of waste into food and then donate it. The 84% isn't a marketing miracle — it's the predictable result of turning an audience from viewers into actors.
For each spoiled turnip players dropped off on Hellmann's Island, Hellmann's donated a real meal to those in need. — Campaign mechanic, Hellmann's vs. Virtual Food Waste (Ogilvy / Edelman, 2020)
A player's avatar handing over turnips at the donation point — the moment a brand purpose becomes a player action.
@RELI
Why it didn't feel like an ad — and why that's the whole point
Strip the campaign down and you find a repeatable blueprint, not a one-off stunt. Hellmann's did three things every brand-in-game effort should and most don't.
It found an emergent player behavior instead of inventing one. Turnip-hoarding and turnip-rot already existed. The brand attached itself to a habit millions had, rather than asking players to learn a new one.
It built a destination, not an impression. An island is a place you go, explore, and bring friends to — not a frame you scroll past. Destinations earn dwell time; impressions beg for it.
It tied the mechanic to a real-world stake. The turnip-for-meal exchange made participation feel like a public good rather than a pitch, which is precisely why 15,000 people chose to broadcast it for free.
The result is advertising that is, functionally, indistinguishable from the product experience. And the business didn't suffer for being subtle: the agency reported Hellmann's saw underlying sales growth in 2020 at 4x its 2019 rate across the period spanning these activations, and the broader "Make Taste Not Waste" purpose has since been credited with inspiring over 200 million people across the US, Canada and UK to waste less food. Purpose that's playable scales. Purpose that's preached gets skipped.
46.7%Engagement rate — ~100x a typical display ad
54.8MPeople reached without paid in-game media
84%Said it changed their food-waste behavior
What this means for your brand
Hellmann's didn't have a gaming budget the size of a Super Bowl spot. It had an insight — that the audience it wanted had moved into a game, and that the game already contained a perfect metaphor for the brand's purpose — and the discipline to build inside the player's world instead of shouting at it from outside. That's not a mayonnaise trick. It's a model any brand can run, and the platforms have only multiplied since 2020.
The catch is that you can't retrofit this onto a media plan. A 46.7% engagement rate and an 84% behavior shift don't come from buying placements; they come from building something players genuinely want to play. That requires a studio that thinks like game designers and marketers at the same time — that can find your version of the spoiled turnip and turn it into a destination. That is exactly what @RELI does.
The takeaway Hellmann's proved the most valuable real estate in marketing isn't a screen you interrupt — it's a world your audience chose to enter. Build inside it, and you don't pay for attention. You earn it, and they thank you for the chance to give it.
Your brand has a spoiled-turnip moment. We'll find it.
Hellmann's didn't buy attention inside Animal Crossing — it built a place players flew to and walked away changed. @RELI builds branded games and native in-game experiences that earn played attention instead of renting interrupted attention. If you want a 46.7% engagement rate instead of a 0.46% one, let's build the game your audience actually wants to play.
Burger King Sponsored a Tiny League Two Club to Hijack FIFA for £50K
@RELI·Field notes·9 min read
For roughly £50,000 — the price of a fourth-tier shirt deal — Burger King put its logo on Messi, Ronaldo and Neymar inside FIFA 20, then turned 10 million players into its media channel. The output: 1.25 billion impressions, a 43-year-first shirt sellout, and three Cannes Grands Prix. The clearest proof yet that played attention beats interrupted attention.
Fifty thousand pounds. That is what it cost Burger King to put its logo on the shirts of Lionel Messi, Cristiano Ronaldo and Neymar — not in a stadium, not on a billboard, but inside the most-played football video game on Earth. The fast-food chain never bought a single second of in-game advertising. It bought a fourth-tier English football club's shirt sponsorship, and let EA Sports' own licensing engine do the rest.
The one-line version Burger King spent roughly £50,000 to sponsor Stevenage FC, a League Two club. Because every licensed kit is rendered inside FIFA 20, the BK logo appeared on shirts in front of more than 10 million players in the game's first weeks — as native game content, not an ad. Players then volunteered as the media channel. Output: 25,000+ user-generated clips, 1.25 billion earned impressions, a club's first-ever shirt sellout in 43 years, and three Cannes Lions Grands Prix.
The cheapest seat in football was also the loudest
In the summer of 2019, Burger King became the front-of-shirt sponsor of Stevenage FC for the 2019/20 season. Stevenage — nicknamed 'The Boro' — sit in League Two, the fourth tier of English football, and on a normal matchday they play in front of a few thousand people. By the logic of traditional sponsorship, this was a rounding error of a deal: a reported £50,000 for the shirt rights of a club most football fans had never heard of.
But the agency behind it, DAVID Madrid/Miami, was not buying Stevenage's stadium. It was buying Stevenage's data layer. EA Sports holds licenses for hundreds of real clubs, and a licensed club's real kit — sponsor logo and all — is faithfully rendered inside FIFA. The moment Burger King's logo went on the physical shirt, it was contractually destined to appear on the virtual one too. As a Burger King spokesperson put it, even at the bottom of the fourth division, Stevenage 'would appear in the same video game as expensive players.'
~£50,000Reported sponsorship spend
4th tierLeague Two — bottom of English pro football
2019/20Season the deal went live
A League Two kit, rendered on millions of screens worldwide — the BK logo as native game content, not an overlay
@RELI
FIFA 20 was the perfect Trojan horse
Timing was everything. FIFA 20 launched in late September 2019, and within the early-access window EA Sports announced it had already crossed 10 million players. Those players had completed more than 450 million matches and scored 1.2 billion goals — all in a matter of weeks. This was not a niche title. It was one of the largest concurrent attention pools in entertainment, refreshing every single year.
Crucially, FIFA's Career Mode lets players transfer any superstar to any club. So a gamer could sign Messi, Ronaldo and Neymar to humble Stevenage — and every one of those world-famous players would now run out wearing the Burger King logo. The brand wasn't interrupting the game. The brand was the game, threaded into the exact object players stare at for hours: the kit.
Why this is native advertising, not product placement Product placement is passive — a logo sits in a scene and hopes to be noticed. Here, the logo lived inside the game's functional data and the player chose to wear it, again and again, in matches they initiated. Attention wasn't bought or interrupted. It was earned through play — the exact thing ad-blockers, skip buttons and banner blindness were invented to defeat.
The Stevenage Challenge: turning players into the media buy
Buying the kit got the logo in the game. The genius was what came next. Burger King launched the #StevenageChallenge: pick Stevenage in FIFA 20, score spectacular goals, post the clip, and Burger King would send you real food — free Whoppers and chips. Overnight, a fourth-tier club became the most desirable team in the game.
The economics flipped. Instead of paying media owners to distribute its message, Burger King paid in burgers and let the audience distribute it for free. Every player chasing a Whopper produced a branded highlight reel and pushed it across Twitter, YouTube and Twitch. The campaign ran for roughly a week of peak intensity and generated more than 25,000 goal clips shared online — a voluntary content engine, paid for in fast food.
There's more to advertising than just rubbing your brand's logo in your customer's face.— Fernando Machado, then Global CMO, Burger King
Machado, the architect of Burger King's most awarded era, framed the idea as bridging two worlds: 'We always thought that we could do something powerful by putting together the world of gaming and real sports — almost bridging the digital and analogical worlds with an idea.' Stevenage became the most-used team in FIFA 20's Career Mode and the most-played team on Twitch, with gamers stacking the squad with Messi, Ronaldo and Neymar.
25,000+ player-made goal clips — a global media channel paid for in Whoppers, not impressions
@RELI
The output dwarfed the input
This is the number every CMO should tape to their monitor. A reported £50,000 sponsorship generated 1.25 billion earned media impressions by the most-cited estimate — and the campaign's own Cannes/Warc submission documented 227 million media impacts across 150+ placements. Either figure represents a leverage ratio that no paid media plan can approach.
£50K turned into a billion impressions
A ~£50,000 shirt sponsorship of a fourth-tier club generated 227M documented media impacts and an estimated 1.25B earned impressions inside FIFA.Source: Spend: Adweek. Impressions: industry estimate (1.25B) and Begin Fusion / Warc Cannes submission (227M).
campaign_inputs_outputs — The full ledger of the Stevenage Challenge — every figure traced to a named source.
Metric
Figure
Source
Sponsorship spend
~£50,000
Adweek
FIFA 20 players (first weeks)
10,000,000+
EA Sports
User-generated goal clips
25,000+
Begin Fusion (Warc/Cannes)
Documented media impacts
227,000,000+ (150+ placements)
Begin Fusion (Warc/Cannes)
Earned media impressions (est.)
1.25 billion
Industry estimate (The Football Week)
Replica shirt sales lift
+300% (first sellout in 43 years)
Begin Fusion / Campaign
UK app awareness lift
+8 percentage points
Begin Fusion
Source: Adweek, EA Sports, Begin Fusion (Warc/Cannes submission), The Football Week, Campaign
The real-world spillover: a 43-year first
Branded games are often dismissed as vanity reach that never touches the business. This one did. Stevenage's replica shirt sold out for the first time in the club's 43-year history, with sales up 300%. Demand came from people who had never set foot in Hertfordshire; they wanted the jersey because they'd played as the team in FIFA. Virtual affinity converted into physical merchandise revenue for a tiny club that had never seen anything like it.
For Burger King the brand effect was equally concrete: a documented +8 percentage point lift in app awareness in the UK, the channel the chain most wanted to grow. The campaign reached a young, male, ad-averse, console-and-streaming audience — precisely the demographic that pays to skip pre-roll and runs ad-blockers — and did it through gameplay they actively chose.
+300%Replica shirt sales lift
43 yrsFirst-ever shirt sellout in club history
+8 ptsUK app awareness lift
Three Grands Prix — and a new playbook
At Cannes Lions 2021, the Stevenage Challenge won Grands Prix in both Direct and Social & Influencer, plus a Titanium Lion — the festival's award reserved for work that breaks new ground and points to the future of the industry. Few campaigns in any year sweep three of Cannes' top prizes. The Direct jury called it a unanimous benchmark for the category.
awards_ledger — The Stevenage Challenge's Cannes Lions haul, agency DAVID Madrid/Miami for Burger King.
Award
Category
Festival
Grand Prix
Direct
Cannes Lions 2021
Grand Prix
Social & Influencer
Cannes Lions 2021
Titanium Lion
Titanium (breakthrough work)
Cannes Lions 2021
Gold + Silver
Direct
Cannes Lions 2021
Silver
Social & Influencer
Cannes Lions 2021
Source: Adweek, Campaign, SHOOTonline
The reason the industry rewarded it so heavily is that it rewrote a rule. You do not need to build a custom game, rent a virtual stadium, or buy a media package to live inside a game world. You can embed in an existing AAA title's data layer for a fraction of the cost — and if you design a mechanic players want to participate in, they become your distribution. That is the whole thesis of native, played attention in one £50,000 case study.
The contrarian takeaway Interrupted attention is a tax you pay over and over — every impression costs money and earns resentment. Played attention compounds: the audience does the work, the reach is earned, and the people you reach are the ones who pay specifically to avoid ads. Burger King didn't outspend anyone. It out-positioned everyone.
What this means if you're a brand today
FIFA is now EA Sports FC, and the gaming attention pool has only grown — across football titles, Roblox, Fortnite and the wider creator-game economy. The Stevenage lesson is not 'go sponsor a small club.' It's that the highest-leverage marketing of the next decade lives inside play, where attention is voluntary, native and immune to the ad-blocking and skip-button behavior that's gutting traditional media. The brands that win won't be the ones with the biggest media budgets. They'll be the ones who become something worth playing.
That is exactly what @RELI builds. We don't buy you an ad slot in someone else's game — we build the branded game, design the mechanic that makes players want to share, and embed your brand into the experience itself, the way Burger King embedded into the kit. Earned attention, owned by you, engineered from the first pixel. The Stevenage Challenge proved the model on a borrowed data layer for £50,000. Imagine what it does when the whole game is yours.
Burger King hijacked someone else's game for £50K and walked away with 1.25 billion impressions and three Cannes Grands Prix. @RELI builds the branded game you own outright — the experience, the mechanic, the share loop — so the attention is native, earned, and yours. Let's build the game your audience plays instead of the ad they skip.
