Push To Talk: The strange evolution of the sponsored stream

There are currently 2,679 playable (for free!) demos for upcoming games on Steam as part of Steam Next Fest — a feast for players and a horrifying statistic if you’re hoping to launch a game next year.

Valve knows that’s entirely too many game demos, so they’ve posted a live scoreboard where you can see the most played and wishlisted demos.1

All throughout this week, two games have jostled for the #1 and #2 spots: Delta Force by Team Jade (a branch of TiMi, a subsidiary of Tencent) and SUPERVIVE by Theorycraft Games (a startup formed by a ton of former League of Legends devs). It’s a clash of the Tencent vs. Ex-Tencent titans.

Theorycraft is very well-funded by startup standards and headed by longtime LoL front man Joe Tung, who also once led production on Destiny and some of the best Halo games. So it’s not a huge surprise to see them land near the top of the heap with their first big outing.

But at the same time, it’s not obvious that having a lot of money would result in a widely played demo during Steam Next Fest. In 2024 game studios with astonishingly huge budgets are regularly failing to break out with players.

One small thing the SUPERVIVE team did was invest in Reddit ads. I’ve never seen anyone do this for a demo, but it seems to have paid off with a ton of positive engagement. Even the comments on the ads were nice.

Theorycraft Games Head of Marketing Alex Goepfert put it mildly when he told me in a Discord DM this week that “When you can run ads on Reddit, leave the comments open (which no one ever does), and not cry when you read the responses, that’s an encouraging signal.”

But the far more important signal the Theorycraft team has long been focused on has been the reaction to SUPERVIVE by content creators. They’ve been inviting creators from the League community and others to playtest SUPERVIVE for years at this point — including long before the game was revealed publicly.

“A consistent focus for us is building authentic partnerships with creators who speak directly to our target audience and genuinely enjoy the game,” Goepfert says.

The Theorycraft team playtested with creators and incorporated their notes while building SUPERVIVE. So when they sponsored some streams alongside their demo launch, many of the creators playing already knew their way around the game. The result has been a bunch of paid and organic creator endorsements for SUPERVIVE that all feel genuine:

“Caedrel going bonkers on stream after a big team fight or Tyler1 telling his audience the game is ‘actually good…’ these are the types of core memories we’re chasing as we introduce the game and studio to players,” Goepfert says.

So the SUPERVIVE campaign is probably on the cutting edge of what a good, well-funded creator marketing campaign looks like for a game launch (and certainly for a demo).

But recently I’ve been thinking in a sort of bigger picture way about the whirlwind transformation of the relationship between game devs and creators over the past ten years or so, because things have changed really fast.

Since 2014 we’ve basically gone through three distinct phases or eras in the ways that game developers and marketing teams have thought about content creators — and particularly Twitch streamers and YouTubers.

People have been posting videos about video games — and, in the early 2010s, livestreaming them — for a long time. But if I had to pick any date as the official start of the modern “gaming creator” era, I’d go with August 25, 2014, when Amazon surprised the world by purchasing Twitch for $970 million in a cash-money deal.

At the time, this valuation was seen as absurdly high by analysts. Was Amazon buying Twitch for the tech so it could better compete with Netflix and YouTube? Sure, the platform was drawing a lot of eyeballs — Twitch reported 55 million monthly active users in July of 2014 — but it wasn’t obvious what sort of business you could run with it.

It also wasn’t clear how game developers should act toward creators. Some of them were doing weird stuff and flouting trademark protections! Do you ignore them? If you’re Nintendo, do you sue them?

It was understood that streamers and YouTubers were helpful for getting new players and keeping old ones engaged, but what do you do for them? A telling factor in the game dev approach to creators in this era was the term we universally used to refer to them: influencers. This was always a cringey marketing term, but it revealed the widespread (and not always wrong) belief among marketing departments that the people watching streamers and YouTubers would believe whatever creators said and take their opinions as their own.

At this point, a lot of game devs sort of defaulted to treating creators like mini-celebrities: flying them out to events, sending them gifts, and trying — often with terrible results — to turn them into semi-official spokespeople for their games.

One funny example: Back in 2018 when I was working on PUBG, it was decided that we should partner with Shroud and Dr Disrespect (I know, I know) to put official Shroud and Doc gun skins in the game.The deal was signed, the gun skins were made, and I had personally written the official announcement post, which was scheduled to go out the following Monday. That’s when Doc went off-script and posted the following tweet:

He was pre-emptively recontextualizing a mutually agreed-upon partnership with the PUBG team (then called Bluehole) as us falling to our knees and giving in to his petulant demands. And he called us “Blueballs.”

I thought it was insane but ultimately harmless and more than a little funny.2 Not everyone agreed. We launched the gun skins anyway.

On February 4th, 2019, Respawn Entertainment revealed and surprise-launched Apex Legends. Suddenly the game was everywhere — basically every FPS streamer was on it, and hundreds of thousands of people were watching.

It was a ballsy gambit, but it worked. Exactly one month later, on March 4th, Respawn announced that the game had hit 50 million players:

This was a freakout moment for games marketers. The Apex launch violated every rule in the book (what about pre-orders? what about any kind of pre-launch marketing at all?) and broke every record.

And it seemed like streamers were the magic aid that enabled it all.

Later in March, Reuters reported that Respawn/Electronic Arts had paid Ninja $1 million to promote the Apex launch.3 At this point games marketers lost their collective minds.

We all started crunching numbers and scribbling math on napkins which seemed to suggest — based on cost per view and install — that $1 million for a few hours of Ninja’s time was actually an insane steal. A game like Apex could (and ultimately did) go on to make billions of dollars. If you could ensure a massive launch by spending just a few million on streamers, that’s like buying a lottery ticket that’s guaranteed to hit.

By August, execs at Microsoft took this line of thinking to its logical conclusion and offered Ninja an exclusive streaming deal on Mixer that cost them somewhere between $30 million and $50 million.

We know how that worked out for Mixer.

But back in 2019, it seemed like anything was possible, if you could just get a big streamer on your team.

The sudden and dramatic rise of Among Us in August and September of 2020 reinforced this thinking. It seemed like proof that a single streamer (for Among Us, it was Sodapoppin) could even bring dead games back to life. They were marketing miracle workers!That simplified narrative wasn’t totally accurate — Innersloth founder and CEO Forest Willard has since said that the game was actually already reaching tens of thousands of players on mobile, and its PC renaissance began in Korea — but in broad strokes the story was true. Among Us broke into the larger consciousness because of streamers.

As demand for creators’ time and attention went up, so did their prices.

As time wore on, big, quick bursts of paid streamer engagement with games became normal for just about every game launch — though success didn’t necessarily always follow for those games that shelled out. Soon, the price began to outstrip the value you could expect to get out of it.

By 2023, an industry standard for pricing had begun to settle into place — for a few hours of sponsored streamtime and a YouTube video upload, game devs could expect to pay somewhere in the neighborhood of $0.05 to $0.10 per average concurrent viewer or average video view. You want a YouTube video that’s likely to get 2 million views? That’ll be $200,000, plus fees for the agents.

If you were looking at these numbers as a marketer managing a large budget, it was all starting to look pretty familiar. Paying 5-to-10 cents per engaged view is not too far outside the range of what you’d expect to get by just paying Google to boost a YouTube video using Google Ads.4

Slowly, paid creator streams and videos were becoming just one more marketing channel. Put money in, get conversions (sales, pre-orders, whatever) out.

In an interview with the A16Z GAMES newsletter earlier this year Arturo Castro, who led marketing on the Apex Legends launch, said that he doesn’t think the its approach would necessarily work now.

“Free-to-play shooters, the rise of the influencer ecosystem in general with Fortnite, and the fact that we made creators feel like they were part of something super unique and special… all that was novel at the time,” Castro told A16Z.

But now, things have changed: “Creators know their worth, and the switching costs of moving from one game to the next is super high.”

The industry perception of the value creators can bring has evolved as well.

For one, it’s no longer clear that paid streams and videos can ever make a game “pop off.” The tactic really only works for games that are great anyway, in part because the same things that make your game sticky with players are also the things that make it appealing to streamers. It helps, for instance, to have dozens or even hundreds of hours of interesting gameplay (Deadlock, Balatro), or novel social mechanics (Among Us, Chained Together). Does creator attention make games pop off, or do games that are popping off anyway naturally draw more creator attention? The cause-effect relationship is no longer so clear.

Though it’s true that creators influence the opinions and beliefs of their viewers, audiences have gotten savvy to the way the paid-promo game works. And most creators value their relationship with their viewers more than they value any single contract with a game dev, so they’ll let slip what they really think about a game even while being paid to promote it.Check out the below exchange between a Twitch chat commenter watching CohhCarnage during a paid stream to promote Concord:

Streamers have become adept at speaking to their viewers in a way that meets their contractual obligations while still revealing their real opinions.

Part of the fun, then, of watching your favorite creator navigate this dance while doing a paid stream. Are they going to like the game? What do they really think?

Sometimes, when a game is clearly not a real fit for a creator, their audiences even make their own fun by trolling the streamer.

I can’t tell you how many times I’ve laughed at this montage of xQc viewers abusing his text-to-speech donation feature to roast Eternal Return:

It’s a difficult dance. But few creators are better at the dance than Tyler1. His entire persona is built on a WWE-style commitment to the bit, where the line between “the real” Tyler and the caricature he plays on stream is impossible to spot.

At the end of last night’s sponsored stream for SUPERVIVE, T1 subjected his viewers to a hammed-up endorsement for the game that, I am absolutely certain, Theorycraft did not tell him to say:

“ALRIGHT, THAT WAS LAST GAME,” Tyler shouted, before sighing deeply. “Again, reminder, link in the chat. Thank you SUPERVIVE for the f***ing sponsor. Game’s GOATed. Everybody should check it out. It is love, it is life. Love it. It’s awesome. Probably… questionably… I would argue… that that is going to be game of the year. Easily, by the way. When I think of fun, joy, entertainment, that game has it all. Okay?”

Does he really mean it? This is, after all, a man who has played League of Legends for 10 hours a day for the better part of a decade. He is never going to give LoL up for another game, even if he really does enjoy trying out new games like SUPERVIVE. His viewers know that.

Everybody knows how the game works, and everybody’s having fun anyway:

That’s it for this week. I’m gonna go play video games I already own in exchange for no money.

I’ll see you next Friday.