We made a livestream, got a little press and launched a successful indie game

It's Jan. 8, a Friday afternoon, and it's squeaky bum time at TinyBuild Games.

Hang on, that's maybe a little too British for the Polygon audience. What I mean to say is that it's Friday afternoon and we're closing the day up with a nail-biter that has us nearly gnawing our entire hands off.

See, 36 hours prior to this, we launched something a bit strange. We fired a Twitch livestreaming channel out into the world called Twitch Plays Punch Club, and invited people to join the livestream and all play our latest game Punch Club together by entering commands into the Twitch chat.

Remember Twitch Plays Pokémon? Our experiment was heavily inspired by that, except that there was a bit of a twist with Twitch Plays Punch Club — we told viewers that until Twitch had beaten Punch Club, the game wouldn't be released on Steam.

Twitch had managed to beat more than half of the game within 24 hours. Our in-house tests indicated it would probably take the Twitch community roughly a week to complete the game, but the real players were handily beating that estimate.

This was a problem.

Games can't launch on Steam over the weekend, so if Twitch didn't manage to beat the game before the Valve offices closed up on Friday afternoon, things could get bad. We started off feeling nervous that they were beating it too quickly, but suddenly we were terrified they wouldn’t beat it quickly enough.

Twitch ultimately managed to beat Punch Club in 36 hours, with just an hour to spare before Steam closed up for the weekend. What we didn't realize on that Friday afternoon was that we'd just accidentally set in motion one of the biggest, most jammy indie game launches in quite some time.

What the heck is Punch Club?

Punch Club is a '90s-styled boxing management tycoon game, developed by Lazy Bear Games and published by TinyBuild, that sees players training hard at the gym, finding love along the way and sometimes punching crocodiles in the face.

The idea is to build up your strength and agility as a boxer, while also going to work each day so that you can purchase food to keep yourself fit and well. Balancing your life in this way makes Punch Club feel a little like a retro version of The Sims, if The Sims were, you know, about knocking people out. The Steam launch took place last weekend, while the iOS version of the game was released today.

The game launched on Steam on Jan. 8, after Twitch managed to beat the game, and has been a rather modest success. We're not allowed to give out Steam sales figures, but you can always check SteamSpy for a rough estimate, Twitch to see the dozens of people streaming the game, or YouTube to see the hundreds of people making videos about Punch Club.

All this without any real support from the press.

How did this happen?

We came up with the idea for Twitch Plays Punch Club in early December. Lazy Bear Games set to work creating a version of the game that could be played via the Twitch live chat after we discussed the idea with both Twitch and Valve, and were presented with big thumbs-up from both.

Valve loved the idea so much that the company agreed to allow us to have pre-orders for the game go live on Jan. 7 as Twitch Plays Punch Club came into the world — Steam pre-orders being a rather rare feature for a non-AAA studio — while also featuring the game on the front page of Steam.

Twitch also featured the stream on its front page for the first 24 hours of the stream. These front-page features, combined with our own marketing push, led over 100,000 people to check out the stream, while thousands of people concurrently played the game together at its peak.

We were already ahead of the game, both literally and figuratively, well before Punch Club’s launch.

How did Twitch beat the game so quickly?

Punch Club is around 15 hours long for one playthrough, so the idea that a chaotic chat box could beat the game in just over double that time is, well, quite astounding. And yet it happened, and took us completely by surprise.

There were two main reasons behind Twitch collectively managing this incredible feat:

1. The hive mind mentality. We had wrongly assumed that the chat would become chaotic, with people randomly typing commands and all trying to do their own things. We couldn’t have been more wrong. From the moment the game started, over a thousand people in the chat all started working together to discover how the game worked, how the training regime should happen, when to go to work, which fights to enter, how often to go shopping, etc. It was quite incredible to watch, and was complemented by...
2. The 25-second vote delay. Every 25 seconds, the chat could vote on what they wanted to happen next. These votes were then tallied, and whichever command received the most votes happened on screen. We had (again, wrongly) assumed that giving Twitch one command every 25 seconds would lead to the game taking quite a while to complete.

What actually happened was the opposite — the extra breathing room between moves meant that the chat had more time to band together and figure out what they should be doing. It didn’t lead to chaos; it led to everyone having the time to take a collective breath and figure out what to do.

See, in Punch Club, you make a ton of bad choices. The game is filled with ways to mess up, and the majority of players end up going completely off the rails for a while before finally realizing that they weren't training the right stats, or picking the correct perks, or buying the correct gear. It’s part of the fun.

It's squeaky bum time at TinyBuild Games.

But the Twitch chat rarely went wrong, because they had lots of time to assess each situation and make sure they got it right. There was a bit of madness towards the end of the game, when the mechanics change quite dramatically all of sudden, yet Twitch still managed to pull themselves together and figured it all out.

It also helped that we'd sent a code for the game to a big streamer named Bikeman the day before, and he decided, quite against his own health, that he would race Twitch Plays Punch Club, and attempt to beat the game in one single sitting while streaming the entire thing.

In fact, Bikeman managed to catch up with Twitch, and eventually beat the Twitch stream to the end of the game by around 30 seconds. Both streams finishing together, with over 4,000 people concurrently watching, was the perfect way for the experiment to come to a close. It was rather astounding: The collective was able to play just as effectively as the single streamer.

The elation in the chat as the credits rolled was palpable, and I'd be lying if I said I didn't let out a little tear. Although that may have been more to do with the fact that I'd been awake and staring at the stream for 30 of the last 36 hours, rather than completely out of joy.

How did people respond to the stream?

The vast majority of Twitch viewers and Twitter users thought it was fantastic. Our mentions on Twitter were bleating consistently, and the Twitch live chat was packed with people excited with the prospect of playing the game, and helping to launch it as soon as possible.

The Steam forums were a different story. Many people were pre-ordering the game and questioning why they had to wait until Twitch beat it in order to play it.

Of course, this is literally how pre-orders work — you pay money for a game and then receive it at a later date — but many Steam users were uneasy about the fact we hadn't set a specific date, and instead had stated a vague "when Twitch beats it." Some players seemed to feel as if we were trying to keep the game from them.

It didn't help that on the Steam page, above where it had the "pre-purchase Punch Club" button, there was a green banner with the word "Available" on it — a banner that was simply part of the Steam UI, and which we couldn't remove. It wasn’t optimal.

We later added a clarification on Steam stating that if Twitch never beat the game we'd launch it on Jan. 25. This seemed to satisfy many complaints. There was no chance the stream was ever going to take until Jan. 25, but having a set date on the books made the players a bit more comfortable.

Meanwhile, the video game press had no idea what to do with the situation. As the stream gained traction, a lot of big news sites realized that they should probably cover it, and a number just reported the facts and left it well alone.

A small number seemed weirdly irritated by it, like it was some sort of marketing abomination — the irony being, of course, that they were now reporting on the stream to their readers, where they wouldn't have before were it not so strange and outlandish. It was much more effective than a press release!

Out of everything that TinyBuild has ever done in the last five years, this was by far the biggest response from the press that we've seen, and continues to be nearly a week later.

The aftermath

I said at the beginning that with Twitch Plays Punch Club, we accidentally spawned one of the biggest indie game launches in many months. It wasn't the stream that was the accident — it was the timing of its conclusion.

When Twitch beat Punch Club an hour before Steam was closing up, and the game went live as the last thing that happened on Friday afternoon, this entirely random launch time/date just happened to be the most perfect we could have hoped for. I wish I could say this was all part of an elaborate plan, but as this timeline shows, it was mostly a good starting idea mixed with dumb luck.

This is how it went down:

Friday afternoon: Twitch beats Punch Club in the middle of a stream that had gone viral, and Punch Club launches on Steam. The video game press has gone home, so we receive very little press for the launch. However, since Steam is now closed for the weekend, the four front-page features we receive all stay there for the entire weekend.

Friday evening: The entire Twitch streamer base comes online, finds codes for the game in their inboxes, and starts streaming the game. Over 50 people were streaming the game on the evening of the launch, pushing the game onto the front page of Twitch. More people were streaming Punch Club than Minecraft, Destiny, Grand Theft Auto 5 and many other giant AAA titles, for at least a short period of time.

Saturday-Sunday: It's the weekend, so every Twitch streamer who normally is at work during the day now has the entire weekend to play Punch Club. There were always around 60 people streaming the game on the service and, once again, it was always on the front page of Twitch. As a result, the game never left the top 15 bestselling games on Steam, and at its peak reached the No. 5 slot, just above XCOM 2. That's some good company in which to find yourself.

Monday: The press comes online, sees that a game called Punch Club has been at the top of the Steam charts all weekend, and starts looking into it. YouTubers have also been playing the game over the weekend, and are now putting videos out in droves, with hundreds uploaded and millions of views over the space of 48 hours.

Tuesday: More YouTubers and Twitch streamers are discovering the game, thanks to new press and our constant appearance in the Top Selling chart on Steam.

And so, nearly a week after launch, the Punch Club train rolls on. At one point over the weekend, there were over 4,200 people concurrently playing the game on Steam, pushing it into the list of top 100 played games on the entirety of Steam.

So what's your point?

This was a hell of a lot of words, so let me try to whittle all of this down into a conclusion:

Twitch is super important for video games right now

Regardless of whether you like watching game livestreams, Twitch is by far the largest influencer of which indie games the general populace is playing — even more so than YouTube.

I'm not just using this one data point to come to this conclusion. TinyBuild has been a Twitch-focused company for the last year now, and it's safe to say we've kinda exploded as a result. While we couldn't have planned the timing, having Twitch be an important part of our strategy was never in question.

Party Hard, which featured Twitch chat integration and was announced on the Twitch stage at PAX East, was a huge commercial success for us as a direct result of its Twitch integration. Our other titles, like SpeedRunners and No Time to Explain, have grown from strength to strength thanks to special Twitch promotions we've put in place.

Game developers should be considering how Twitch can work with their games. Twitch integration has not proven to be an overly time-consuming feature for any of our projects, and has paid for itself in spades come launch.

For non-AAA games, the video game press isn't really all that useful

Up until Tuesday afternoon, Punch Club didn't have a single review on Metacritic, yet it was already by far one of the bestselling indie titles of the last several months.

That's not to say that the games press is completely useless for developers — we've seen a new batch of YouTubers and Twitch streamers playing the game since the press began reporting on the game on Monday — but it does mean that, where game developers were once told that it was important to first focus on getting press for their games in previous years, that's really not the case anymore.

Getting press is now more of an afterthought for us, whereas getting streamers and YouTubers to play our games is always at the forefront of our minds.

As for our future with Twitch, we’re currently putting together a new Twitch-based idea for the launch of Dungelot: Shattered Lands very soon, and at PAX South in a couple of weeks, Twitch has given us the entire first hour of its PAX South livestream to do gameplay reveals for four brand-new titles that we're announcing this week.

To any game developers out there who are considering dabbling with Twitch integration, and courting Twitch streaming? Stop dabbling. Go do it.

Update: As was pointed out on Twitter, the game was heavily promoted to the press, although it failed to gain much traction. We’ve since edited the post to address this fact.

Mike Rose is the loud British guy from developer and publisher tinyBuild Games. He previously wrote about video games for a decade, at places like Gamasutra, Kotaku and Pocket Gamer. Now he signs games, shouts a lot, and probably tweets too much.