Using Analytics to Create Viral Games

In 2023 I spent a month studying retention charts of 100 games, and found a handful of consistent and predictable patterns regarding mechanics that impact retention, and thus the virality of a game.
Everything in Order: • Everything in Order
What Triggers a Game to Go Viral: • What Triggers a Game t...
What Actually is Entertainment: • What ACTUALLY is enter...
How Successful Games Leverage Human Instinct: • How Successful Games L...
Retention Study Conclusions Document: docs.google.co...

[@DavidWehleGames]

This is amazing, thanks so much Gavin. I keep similar notes but yours are way more detailed, so interesting to see

[@JoeIACT]

@davidwehlegames, do you know how to get these retention graphs from any youtube video? Do you make it as well? I didn't find it anywhere.

[@thomasbrush]

I dropped everything to watch this

[@ghostcloudmusic6867]

^this guy brought me here

[@MCA-GK]

Thanks Thomas

[@packdebienvenida]

I am here because of you, Thomas. Thank you

[@X_Jay_9]

The idea of creating a game that HELPS a streamer make a video for themselves rather than relying on them to boost your game is the perfect example of a symbiotic relationship between content creators and game developers

[@AceDroo]

This is one of the reasons why Content Warning was so successful - they incorporated one of the things that made its inspiration Lethal Company so successful (how viral/shareable its moments are), and made that a core mechanic, making everyone a content creator and vastly increasing its visibility on social media

[@boad8270]

@@AceDroo i mean Lethal company still did a lot better if we go by overall popularity

[@chidorirasenganz]

@@boad8270because it was first lol

[@AceDroo]

@@boad8270 Thats very true, and honestly I think Content Warning's success was always going to be limited due to its nature as a fast follow to/parody of Lethal Company - even with that initial boost of virality that it got from being initially free on Steam and having that recording mechanic, at the end of the day I think it wasn't going to survive long term when compared to Lethal Company without having more mechanical improvements or changes to differentiate it.

[@sanatdeveloper]

I've already listened to 27 minutes of this, and it's pure DIAMOND. There are so many videos out there where you learn something useful, but this one feels like finding your first diamond in Minecraft. Thank you so much!

[@ariamohebi]

BRO, you're such a goat for sharing all of this! Thank you! ❤

[@iojourny]

First, thanks for the video. This is valuable data. * A note about End Credits - The dropoff is ok, since the alternative is just the game ending (100% dropoff). It's completely fine that some people skip it. Not to disregard your analysis, but for a developer (rather than a KZbinr choosing which game to make a video about next), there is a very important thing to remember - YT retention is not sales, and may even be harmful to sales. By far not every game *should* be designed to maximize KZbin retention. I don't have the videos saved, but I've watched more than one detailed video about this topic - sometimes (pretty often!), there is an inverse relation between a games popularity on KZbin / Twitch, and its sales. The most intuitive examples are beautiful games with shallow mechanics and linear stories (which is fine by itself, since they are just tools used to elevate the story). Actually, extreme examples here are visual novels. Why is it bad for them to be popular on KZbin? Because once you've seen someone play the game, you no longer have any reason to play them yourself! Not only would you be repeating the same linear story, you'd also be paying for the privilege, too. On the other hand, games with deep and complex mechanics may be bad for retention, but if those mechanics are appealing to the audience (as players), a certain percentage of those dropped views are actually people going on to read more about the game, and potentially buy it. I can tell from personal experience that I had bought a few games this way, while many games that retained me as a viewer never made it into my steam library. Of course, some games can have the best of both world, but that's mainly either multiplayer games, or games which rely heavily on procedural generation. Basically, KZbin retention is only good for you (as a developer) if the viewer knows (or rather, believes) that there is much more to the game than what he saw in the video. In other words, to actually buy your game after watching he video, the viewer must believe there still exist sufficient *goals* for him to achieve, which weren't covered by he video. An example for this is roguelikes, where even after watching a few runs, there is massive variation left for the player to explore (and, most importantly, the viewer knows it). If you build your game purely for KZbin views and disregard actual conversion to sales, you are gonna have a bad time.

[@ScientiaLudos]

For sure, virality just gets you attention, and on it's own doesn't do anything for sales conversions. This is why in other videos I've talked about how you need to design a game with both attention and conversions in mind. Player agency, replayability etc. are crucial for conversions, but good conversion rates with low visibility are just as worthless as a super viral game with no conversions. The more you optimize both multipliers, the better the game will perform. Edit: Note on the credits... Video ending doesn't equal a 100% dropoff, because since the video is ending, the length is shorter, and the average viewed percentage is higher, since people aren't leaving during the credits. KZbin isn't just taking into account the amount of time spent watching a video, they also factor in the percentage of the video that you watched. Because of this, whatever the end of the video is, no matter the length, it's best for people to watch until the very end, as opposed to dropping off at any point.

[@iojourny]

@@ScientiaLudos I completely agree with the fact that both multipliers should ideally be balanced, at least to the largest extent possible without damaging the eventual game too much (all of this is vague, but that's because there are too many variables depending on each specific game). I just thought it important to mention here, since with so much (well deserved) focus on virality, it's easy to lose sight of the main goal. ... regarding the video ending, are you certain? If people watching to the very end has significant weight (as opposed to total watch time), aren't basically everyone who have an outro shooting themselves in the foot?

[@ScientiaLudos]

Defo worth mentioning 👌. "aren't basically everyone who have an outro shooting themselves in the foot?" Yep they are. That's why all the largest channels that get the most views keep their outros as short as they possibly can.

[@pawelmurias]

"Hook people at the start of the video, transition them to an amazing story that they are invested in, have no dull moments, and then have a satisfying payoff at the end of the video with an abrupt ending." - leaked Mr Beast handbook. You can see how they wrap up the videos really fast after the payoff

[@dennismakesgames]

I mean this is incredible, thanks for all the effort you put into this and providing this info to everybody. Big kudos.

[@AlexFCook]

These insights are amazing. I'm grateful to have stumbled across your channel. Great value provided.

[@r8teful]

2h video of absolute gold information? This is a blessing, thank you.

[@AlexGorskov]

Thank you for the information, Gavin!

[@Hazardous]

For sure watching this whole thing tomorrow, keep up the great videos!

[@damori3604]

This is actually the best advice I've heard for indie game developers

[@sinelacico]

Great video, watched through the whole thing and wrote a lot of applicable notes for my demo's design that will help me get the design and structure a lot tighter. Thank you for these videos they are truly informative!

[@kardrasa]

Wow, amazing work. Will consider all of this when I get back to game making

[@HopperGameDevelopment]

This is good quality, comprehensive research. Impressive.

[@BzechuanInteractive]

6 minutes in and Ive paused to say this. Youre gonna get these 2 hours from me, as well as the watch time for your other 13 videos. Everything you have said has had immense value in "Indie Game Dev Mindset". And im new to the game, so I appreciate it.

[@BigSauceGames]

Locking in for the next 2 hours

[@ryanjdevlin87]

U read my mind

[@ComputeHeavy]

What an amazing video. Thank you! Someone who isn't a KZbinr and game developer would not have thought to do this analysis. Brilliant.

[@undeadpresident]

These things make sense and can help both developers and streamers. You are a good man.

[@BigSauceGames]

Very thankful for these videos Gavin! We all appreciate you sharing this info. It is interesting to think that we are not only creating games, but also tools for youtubers to use to increase their views and retention!

[@TheRealCzechmark]

This Man is a Hero Among Game Devs!

[@RobinvanTaarling-m1t]

Thank you so much for these deep analysis video's and collecting and preparing all this data. You are teaching me way more value as the game development study I just completed, and it's all for free! Can't wait to apply some of these concepts into my game.

[@Vehementus]

This is fantastic Gavin, thank you for sharing.

[@ccavatax]

Thanks for the great insight, I'll save that for the future

[@adrianweber2958]

great info, so usefull . Thanks!

[@robinshengames]

Wow, I am sitting here, watching and actually taking notes. Thank you for this!

[@valkoinendev]

Love content like this! thanks

[@Dash_Dominion]

kudos ! This gives some external Motivatioon.. Thanks sir !

[@user-uq2ko4jq2v]

Interesting topic. Thanks for doing this man. Hope we can pay it forward once we have found some form of success.

[@valkoinendev]

Saw the text document and knew this was going to be some of the best advice ever!

[@fuseboxjackson2789]

You should make a video or a series of videos analyzing already viral games to point out mechanics that made the games viral

[@nullx2368]

Its similar on tips around good presentation talks. Its all about the promise at the start and how its delivered by the end. The world really acts the same in most situations

[@thomasbrush]

29:50 lol Gavin begins to panick

[@kthuludev]

your channel is a gold mine but i have to point out you look like you dont get any sleep!! cant wait to see the day when you have healthy looking eyes. keep up the great work!!

[@neenaw]

i shall sleep with gavins game dev affirmations blasting into my ears

[@ryanjdevlin87]

Legit masterclass

[@MCA_Critic]

I’m not at work anymore but I’ll listen to it at work tomorrow 🙏🏻

[@damanimuham3601]

Thanks for going thu this.

[@almostalwaysadam]

I would love to see the retention data for this video. Great video!

[@UsmanDev]

Huge W masterclass, do you have any information regarding what causes a viewer to purchase the game after seeing gameplay, there must be some factors that incentivize the purchase instead of just watching youtubers play it

[@valson_officiel]

This dude is doing a great job please don't delete these beautiful videos i will comeback to them and take some notes NiceVideo

[@SamGarcia]

What I'm thinking is putting a streamer mode in a game that cuts out all the retention drop stuff for recording and streaming but the normal mode is when one isn't streaming.

[@tedbendixson]

Watching this, it's now obvious to me why you were so successful with Choo Choo Charles.

[@seamusoblainn]

Good stuff. Watched it over 3 hours in another device and took notes of baout 6 pages. Interesting where games and videos overlap and where they don't.

[@seamusoblainn]

Edit: 13

[@samuelalvarez7631]

I'm with you on a lot of the things you're saying, but you have to keep in mind that some people use Alpha Beta Gamer's videos to help them complete a puzzle or a part of the game. That's why I think in some games, when finishing a level, there's a significant drop in retention, even when the transition between levels is really quick.

[@AaronGwynaireGameDev]

Loving that music at 24:06! :D

[@jenniferspengler4688]

I just started watching and can't wait to learn. And I've never seen this channel before. (Just subbed!) But I'm so distracted by that door that I already know I'm gonna have to rewind a few minutes because I zoned out on it. Is that part of the retention strategy?! 😅😅😂

[@SenkaZver]

5:30 thats a really good analogy. Content creators are looking for new content, gold mines. And we can provide the gold mine and pickaxes.

[@sirbumbo]

Keep in mind avoid over thinking all of this stuff. Don't sacrifice actual mechanics in-game that are more enjoyable for your players just because they don't translate well to a viewing format. Gavin said himself in earlier videos these aren't meant for beginners, this is only if you already know what you're doing and you REALLY want to minmax. Especially the bit with the credits well yeah... it's the end of the video. Credits roll and the youtuber was about to end it anyway. Doesn't matter at that point.

[@ComfortAmbience]

cool channel and video, nice insights

[@centneriss]

Please start a discord community for this channel. I think there is a lot of interesting people that get pulled by this type of content. I want to meet them. I'm sure some of us here would be happy to help you run it, so it doesn't take too much of your time.

[@AndrzejGieraltCreative]

Great stuff, awesome that you crunched so much data. Do you have any info on click through rates and how that affects virality?

[@pixelshaketv]

Obviously not all the way through the video yet, but at around 17:30 you note the 13% dip after a death, and I don't believe that data point has been correctly attributed. It looks to me based on the graph that the real dip is actually almost 0% difference but rather, the general bleed is being masked by the peak created by people RE-WATCHING the actual death. loving the video mate, just wanted to point out something that maybe sheds a new light on something.

[@ScientiaLudos]

Certainly could be, would make sense with this being a horror game. Probably should have shown more non-horror examples for player deaths, but oh well

[@pixelshaketv]

@@ScientiaLudos It's all good, honestly I 100% agree with your conclusion on deaths. I know I skip forward on deaths most of the time! Getting access to these analytics is a public service so thank you.

[@zZMazeZz]

"gooners" really got me haha

[@terminus8497]

great video thank you . note that this data is sourced from a niche KZbin channel's viewers and may not represent the broader or more diverse video game market

[@TurtleBoyoIsCool]

I would be so interested if you did like one or two videos just on normal youtube analytic stuff

[@Bot-mv9qb]

Thank you so much for sharing your research! Would you be willing to share the Google sheet document as well? 🙏 I would love to dive into the data as well and see if maybe there's more stuff that could be learned from it. Regardless, thanks again for sharing this! :)

[@LoganJones-us2uf]

The Google sheet would be incredibly useful for re-enforcing the concepts through examples!!! I agree!.

[@LmAnubis]

Great vid

[@gryzlaw]

Amazing video with invaluable insights. Just one comment, this isn't virality. Virality is how much viewers engage with content through commenting, sharing, reposting, and stitching. Which I would argue would probably have some different takeaways from just audience retention

[@5naf_plushies482]

Awesome! Subbed to watch all your videos later!

[@truonghainam]

Great inspection, if you able to share the excel file as well that'd be awesome! Thank again!

[@phdost3267]

Do you have any information on the age range of the viewers? There are multiple paths to virality and I think "lets play" viewership is good to reach a younger audience but then you have games like vampire survivors or balatro that have relatively horrible views but still go viral from a different path.. probably reaching a more mature audience. EDIT: so both VS and Balatro have also strongly benefited from lets plays (splattercatgaming and northernlion) though it's still true that these games would fit awkwardly with your model which could be shown by the lack of views they both have. I still think it's an age thing where games that target older gamers need less views to get the ball rolling and have a better conversion rate.

[@heyheythrowaway]

A game doesn't need to be packed with boss fights, jump scares, and chase scenes to benefit from being streamed. Regardless of what kind of game you make, it's still imperative in marketing to reach out to streamers. He's just giving some analysis on what common features he'd use to make the most viral game possible. While a good thing to consider, I think that can't be your only concern when designing a game. As he said, these are just guidelines.

[@phdost3267]

@@heyheythrowaway I'm not disagreeing. I'm just suggesting that his analysis has a younger market bias and the techniques you'd use to go viral with an older audience may have similarities but are ultimately different.

[@ScientiaLudos]

Def a good point. Mass virality is much rarer and more difficult without a younger target demographic. More mature target demographics typically require better word of mouth, or use of the Steam algorithm.

[@grape8667]

is using comic sans for Cuffbust an attention or an action multiplier

[@RickyWiild]

14:50 Think i nailed that one personally

[@bore_inc]

Damn! We living in the best timeline where this video is free.

[@Kilakilic]

my retention is dipping

[@vlad_churukanov]

24:41 well the player get the idea of core loop and leave, so what? How does the fact that the viewer watches to the end correlate with the player's interest/disinterest? I subscribe to alpha beta gamer and when I watch such videos, I fast forward a couple of clicks to the middle and then fast forward a couple of clicks to the end.

[@packdebienvenida]

what's behind that door with so many security to close it? Zombies?

[@neververver]

who is bro hiding from

[@TheRayll]

W

[@FR3NKD]

Hey Gavin, how can we connect?

[@supapaw]

As an 8 year old I wanted to get into making games so I could explore my creativity, make art and express myself. Finding that making games nowadays is the opposite of that is just depressing.

[@ScientiaLudos]

I feel the exact same way, it really sucks

[@thiagod47]

Don't get discouraged man, most of art has always been made with constraints. That being said, you can still just make the game you want, as long as you treat it like a hobby and dont care so much about making buckets of money.

[@sinelacico]

Games have always been like this, to a stronger extent but games have always been a product-- you can personally choose to not view them that way and as art, which they also are, and make what you want to make, it just becomes less financially feasible that way unless you can support yourself and the game through other means. At the end of the day we all gotta eat and have a place to sleep.

[@tedbendixson]

You could pick and choose which of these things you want to pay attention to. Most of them don't seem to affect the core gameplay that much. If anything, most of the advice is about stopping interruptions that make people bail, and you would want that in most games anyway. It's the difference between Mega Man X vs. Mighty No. 9. Mega Man X just throws you into the action, and it's awesome. Mighthy No. 9 starts with all this exposition and I lose all interest.

[@packdebienvenida]

welcome into adulthood Brother

[@uncl33]

oh boy, 2 hours, let me guess, it's your favorite topic!

[@Jouunaas]

how is the retention on this channel then :_D

[@OriginRow]

This is Quest

[@StoryFurnace]

The fact this is free is insane.

[@JustinWellsDev]

First hour club here we goooooo

[@alexleonardkrea]

Dude, what the hell are you trying to keep out of that room 😂?

[@vurkanseker]

bro, read your business e-mails. you should allow KZbin Join. This content worth a gold nugget for an indie dev like me. I want to pay back somehow.

[@NoCodeFilmmaker]

Basically design a game for someone with ADHD lol By the way, your door has NYC vibes

[@SandorClegane-TheHound]

Bro why you got so many locks on that door? And your eyes so red.. Just a couple asks prior to me watching this...

[@gandordev]

Very intelligent, I bet you have high IQ.

[@ЮлианМацкевич]

For a guy who's all about pulling all the stops, you are awfuly hesitant about introducing a bit of booba to the games

[@adamcolon]

Yeah, this 2 hour video of watching a human talk isn't gonna let the ADHD crowd get any value from this... just sayin

[@AustinJD84]

You have made exactly squat

[@andreash7920]

Why have you locked yourself up like that. is your neighbourhood that dengerous?

[@balefirefury]

Why should I trust you on this topic when you can't condense a 10 page paper into something more digestible than 2+ hours? You did a horrible job of retaining me here..

[@lukaspaver]

Not everything can be explained simply.

[@AdobadoFantastico]

Because he's not trying to make money off of you. He doesn't need your retention here. This is for your benefit if you have the patience to listen. Do you want the info now, or 5 years from now as a retrospective when he finds the time to edit?

[@balefirefury]

@@AdobadoFantastico I got the info by reading his document in less than 15 minutes. The video didn't need to be 2 hours long, period. It's an unedited ramble. I just thought the irony of the situation was funny but let's not pretend this is some gold star transcendent information here. It's about optimizing games for the most schizoid outcomes. I'm a gamer and a game developer. I want to play and make games that are fun. Saying 'this x and y feature caused a 3% drop in viewer retention' is some psychotic capitalist brain rot dogshit. It's interesting to look at but it's not changing my approach to making games. Nor should we venerate this guy to god status just because he was lucky and had a single game go viral. That doesn't make it good, nor should it be your goal as a dev. I'm not surprised though, most people will look towards whatever they think will give them success. I just happen to hard disagree that these parameters are going to be what gets you there. None of that to even mention the obvious that if this guy was really a master of virality he wouldn't be averaging 5k views a video.

[@chilenitonuma]

@balefirefury, this is just a side channel specifically aimed at developing. His entertainment channel averages closer to 200k per video with some going past 5 million views.

[@balefirefury]

@@chilenitonuma doesn't change anything I said except the last sentence