A lot of this week pointed at the same shift: who people listen to before they decide what to watch, play, or buy. Global Citizen, whose festival was built on names like Beyoncé and Coldplay, brought on a creator partner because its headliners don't reliably pull in the audiences it's after anymore. Twitch's CEO said the quiet part out loud about why its clips matter right now. And Japan, as console-loyal as any market in the world, more than doubled its PC gaming revenue in four years on the back of games people love watching other people play.

Today: why I think "the attention era is over" undersells what actually changed, what YouTube's new reaction feature means for the creators being reacted to, and what Japan's PC boom says about how games find their players now.

PUBG's first US ComplexCon collab is built around a line every player has seen more times than they'd like, and that's why it works so well.

Krafton is bringing PUBG to ComplexCon's 10th anniversary in LA this weekend (October 3-4) with an apparel capsule made alongside Floor 13, MGK's creative studio, available at retail only to attendees. The collection is built around the number 13 and the idea of luck, under the line "Better Luck Next Time." Floor 13's creative director framed it as flipping the superstition around 13. Anyone who's played PUBG will read it a little differently, since that's more or less what the game tells you every time you don't come out on top.

A lot of the gaming collabs I've seen never get past putting a logo on a hoodie. This one pulls from a feeling the community already has, and the booth sticks with it too, built around Miramar (one of the game's signature maps) with a helicopter from MGK's live shows sitting in the middle of it all.

It's also not a one-off. PUBG did something similar at ComplexCon Hong Kong in March with Korean streetwear label Thug Club, and Krafton's head of PUBG IP experience called this the first of several experiences planned outside the game. Test in one market, learn, then show up in the next one with a sharper version. I'd like to see more publishers treat physical culture this deliberately instead of simply selling merch.

Receipt: GamesBeat

Global Citizen built its festival on celebrity headliners. Now it's bringing in creators to reach the audience those headliners stopped reaching.

The anti-poverty organization named Influential, the Publicis-owned influencer company, as its global creator partner. It's a pro bono deal that started around the World Cup, and the first big activation was built around Saturday's festival in Central Park, with roughly 20 creators lined up to post from the grounds. Influential says its opt-in network has grown from about 5 million creators when Publicis bought it in 2024 to roughly 18 million now.

The framing is what gets me. Adweek described Global Citizen as turning to creators to reach audiences its celebrity performers don't reliably pull in anymore. This is an organization whose whole brand was built on stars playing to a packed Great Lawn, deciding that a couple dozen creators are the better bet going forward. That's a big admission from anybody, and a bigger one from a cause org, since cause marketing has leaned on famous faces for as long as anyone can remember.

Cause marketing has run on famous faces for decades because fame used to be the fastest route to attention. What Global Citizen is admitting is that attention and trust have come apart, and for a cause asking people to care, trust is the one that counts.

Receipt: Adweek

YouTube now lets creators go live reacting to long-form videos by default, and it's still unclear whether the original creator sees any monetary value from it.

At Made on YouTube, the platform expanded Watch With so creators can hit "React live" on any eligible long-form video or livestream and go live reacting to it. It's on by default for creator-made long-form videos, and creators have to go into YouTube Studio to turn it off.

The creator doing the reacting keeps their ad revenue and any memberships they receive from the livestream. The creator whose video is being reacted to gets attribution and what YouTube calls "added view credit" in their analytics dashboard. As Tubefilter pointed out, it isn't clear whether that view credit pays anything at all.

Reaction is one of the biggest content categories on every platform, and a lot of it has always lived somewhere between real commentary and just rebroadcasting someone else's work. Building it into the product at least brings the attribution with it. But default-on with the opt-out tucked in settings is a pattern I've written about more than once this year, and it usually means the platform expects most people won't bother to change it.

It also fits what YouTube has been doing all year. First it started counting views from the first frame, and now it's turning every long-form video into potential raw material for someone else's livestream. The video is becoming less of a finished product and more of an input. For brands, that means a sponsored segment can end up well past the video it was bought in, into reaction streams, in front of audiences nobody chose and next to commentary nobody approved. That could be free reach or it could be your product getting picked apart live, and right now the brand has no say in which.

Receipt: Tubefilter / YouTube Blog

Two Forbes pieces this week say the creator economy's attention era is ending. I agree with them, but I think calling it an attention story misses what actually changed.

Both ran in Forbes Councils, one on what comes after the attention era and one on why influencers might spark a purchase while quality keeps people coming back. They're member-written opinion pieces rather than reporting, so I'd read them as a sign of where the industry's head is at, not as evidence. And where the industry's head is at matters because for years the working theory was simple: get in front of enough people and conversion takes care of itself.

That theory depended on authority coming with the platform. A TV spokesperson or a magazine cover. You trusted the message partly because of where it showed up.

There's also something a little funny about where these two pieces ran. Forbes Councils is a paid membership program, so the pieces carry the Forbes name through membership rather than through the newsroom. That's the old model in miniature, with the credibility coming from the masthead more than the person writing. It's not a knock on either author, and plenty of Councils pieces are sharp, but two articles about platform authority fading, published on the strength of platform authority, is a pretty good illustration of the gap they're describing.

What changed isn't really how much attention there is, but where trust lives. For most of history, the people who shaped what you bought were close to you, like family, friends, or the coworker who actually knows about cameras. That circle used to be limited by geography. Now it's sorted by interest. The person with real sway over your next camera purchase might live in another country and have 180,000 subscribers, and you trust them the way you'd trust that coworker, because you've watched them be right about this stuff for years.

The data here is older than people think, and it's been pointing in the same direction for some time. Nielsen's 2021 Trust in Advertising study found that 71% of consumers trust advertising, opinions, and product placements from influencers. Across nearly 200 campaigns Nielsen measured, influencer ads drove a 9-point lift in purchase intent over people who didn't see them. I'd be upfront about the limits, though. Those are trust and intent numbers, not sales, and a lot of this still gets measured with the wrong instruments (the same problem I dug into with the Affinity 100 a couple of weeks ago).

For brands, that changes what the job is. With reach, the work was mostly buying the biggest audience you could afford. With trust, it looks more like mapping it all out. Who in your category do people actually listen to right before they decide to buy, not just while they're scrolling? That list is usually shorter and a lot more specific than the one an agency sends over, and it takes time inside the community to build.

It rewards patience in a way reach never did. A one-off integration starts from zero every time. A creator who's been talking about your category for a year has already done half the persuading before your next campaign starts, because the audience has watched the relationship hold up.

The catch (and I think it's the real reason this keeps losing budget fights) is that it shows up in a dashboard more slowly than a reach number does. Reach is a cost line, and cost lines are easy to defend in a quarterly review. Trust behaves more like an asset, but it doesn't sit on a wrap report or metric screenshot, so it tends to lose that argument to whatever's easier to put on a slide. And when it goes wrong, a community that feels used talks about it, and it remembers.

Receipt: Forbes Councils: What Comes After the Creator Economy's Attention Era / Forbes Councils: Why Influencers Might Inspire a Purchase But Quality Keeps Consumers Coming Back / Nielsen

Twitch is building a Recaps feed that turns every stream into a highlight reel, and it takes a swing at live streaming's oldest problem, which is that if you missed it, you missed it.

CEO Dan Clancy told GamesBeat at Tokyo Game Show that creators will be able to pick five to eight clips at the end of a stream and get a highlight reel generated automatically. The reels stay up permanently, and Twitch can algorithmically recommend them to people who don't already follow the creator. Clancy said it should roll out in the coming weeks.

Live has always had a strange relationship with its own audience because the urgency is the whole appeal. You had to be there, and being there is what makes a moment feel like it belongs to you. But that same urgency caps the audience at whoever managed to tune in. Clancy put a rough number on it. At any given point in a stream, maybe 10% of a creator's community is watching, and everyone else hears about it secondhand, if they hear about it at all.

Until now, the fix for that mostly lived off Twitch. Bigger streamers hired editors to cut recap videos, and clippers pushed the best moments out to YouTube, Instagram, and TikTok. That works, but it means the second life of a stream happens somewhere else, and mostly for creators who can afford the support team to make it. Recaps brings some of that on-platform and offers access to every creator, including the ones who could never justify paying an editor.

Recaps could also make the FOMO work in Twitch's favor. Someone who catches a recap of a stream they missed gets a taste of what happened and a reason to show up live next time, which is exactly the loop Twitch spent years resisting back when it treated short-form as competition. Done right, the recap works like a teaser trailer for the next stream.

Sponsorships ride along with that too. Clancy estimated a sponsor read might reach around 15% of a creator's community during the stream itself, and anyone who's bought Twitch integrations knows the feeling of paying for a community and getting whoever tuned in during that time. A sponsored moment that makes the recap keeps working well after the stream ends, and Twitch says it's working on making clips more useful for sponsored content specifically. It cuts the other way too, though, since the streamer is the one picking the clips. Five to eight slots is tight, and a creator choosing what best represents the stream is more than likely going to reach for the funniest or most chaotic moments before an ad read, so there's a good chance the sponsored segment never makes the cut. Unless deals start writing the recap feature into the terms, the part of the stream a brand paid for may be the part most likely to get left out of its second life.

Whether five to eight clips can carry what makes a stream worth watching is still an open question. A lot of live's appeal is the slow build and the chaos between the big moments, and a highlight reel is built to cut exactly that out. If the recap ends up being enough, some people will simply stop showing up live. Clancy's betting on the opposite, and we'll find out fairly quickly once it ships.

Receipt: GamesBeat

Japan's domestic PC gaming revenue in 2025, according to a preliminary report from Japan's Computer Entertainment Supplier's Association (CESA), up from ¥132.1 billion in 2021. That's more than double in four years and the first time it's cleared ¥300 billion. Japan has been about as console-loyal as any major market in the world, with Nintendo and Sony on top and even Microsoft struggling to make a dent. So a jump like that is a big shift there.

Per Kotaku, most of it came out of streaming culture, especially YouTube in Japan, and global games that are built to be watched as much as played. The three biggest PC games in Japan by revenue were Counter-Strike 2, PUBG: Battlegrounds, and Apex Legends. Of the top twelve, only one, Capcom's Monster Hunter Wilds, came from a Japanese studio.

Set that next to the rest of the market and it gets much sharper. Mobile is still 71% of Japan's game revenue but has grown just 4.5% since 2021. Console bounced around until the Switch 2 lifted it. PC went from 6% of the market to 12%, and it's driving almost all of the overall growth.

My read is that streamers picked these games up first, and at some point their viewers went from watching to playing. That would also explain why the list is so heavy on competitive shooters. They make for good streams, with a fresh round every few minutes and plenty worth clipping. For Western publishers, it suggests a game's biggest asset in Japan might be how it plays on someone else's channel, which is very different from what usually goes into simple localization.

Kotaku raises one caveat. Hardware prices are climbing, partly because of AI-driven demand for key components, and that could slow this down.

Receipt: Kotaku / GamesRadar

Forward this to whoever on your team still reports creator campaigns by follower count. The question that matters is who the audience actually listens to before they buy, and follower count can't answer it.

Tapped in. In case you aren't.