Chris Plays Games

[chris-plays-games]

Audience Support

While at VidCon, I attended a panel called “How to ask for support from your Audience.” The panel was a discussion led by Greg Miller, one of the founding members of Kinda Funny, and it was inspiring and awesome.

Greg told a compelling tale to me: The people who are giving you money -- no matter how they’re doing it -- are getting value from your content. They’re seeing something more out of what you are making than just the time engaging with it -- they’re willing to trade money for the chance to be a member. of something special. This is true for me for many creators: I get much more out than I’m putting in. Whether it’s the enjoyment of watching gameplay, or a fun chat after the stream is over; whether it’s a limited access voice channel, or just a small icon next to my name in a chat room; I feel like a part of the community when I give money, either via Fan Funding or sponsorship.

Previously, I feared that I would be seen as money-grubbing, but I think I’ve changed my perspective on that: This is not a chance for me to try to extract money from you. This is a way for you to reward me with the value that you may or may not see in my work.

If the value you see is commensurate with just stopping by the stream occasionally: Awesome. Thank you so much for being part of what I find to be an awesome community.

If the value is such that you’d like to throw an occasional dollar my way; great! That’s awesome, and it means a lot.

And if what you’d like to do is just sign up to support -- at whatever level -- every month, then I thank you for that, and I hope that I can continue to provide value that matches those expectations.

Because in the end, that’s really all this is: I’m trying to set appropriate expectations, and trying to increase my relative value to my audience to be commensurate with those expectations. I try very hard to go above and beyond in everything I do, and this won’t be any exception.

If you’re interested in more content -- like this blog, like my videos, or like the various software tools I’ve created -- or special access to, or just want to feel good about something -- consider supporting Chris Plays Games on Patreon.

[chris-plays-games]

Measuring Streaming Latency on YouTube Gaming

When live-streaming, interaction with your audience is key. I often see people asking “What’s the stream delay on YouTube Gaming?”. The answer almost always depends on what you’re trying to measure: there are multiple sources of this delay, so it’s important to make sure you’re comparing apples to apples -- and knowing where your delay comes from.

There are two common ways that people guess at the delay in YouTube Gaming streams that I’ve watched, and both of them have the same problem that makes them slightly problematic: 

Say something in the video, and ask people to type in the chat when they hear it/see it.

Type something into chat, then see how long it takes to appear in an embedded chat window in the stream video.

The biggest problem with both of these approaches is that YouTube livestream chat is not an instantaneous delivery mechanism: instead, it uses a polling infrastructure to routinely fetch messages. The YouTube Chat API makes this explicit via the pollingIntervalMillis property: “The amount of time, in milliseconds, that the client should wait before polling again for new live chat messages.” In my experience, this value can vary between 1s and 10s, depending how busy the chat is. (Even what seem like “busy” chats to me often times have polling delays of 3-5s.)

While I don’t know whether the chat interface in the website / mobile apps uses the exact same functionality, the effect seems similar from watching many streams:

For those with embedded chat, the broadcaster will sometimes respond to a message which isn’t yet included in their video, because their main browser and their in-stream browser are polling at slightly different intervals.

If you have two devices open, you may see new chat messages appear some time apart; this sometimes leads to people thinking that “mobile” or “desktop” have a higher delay -- but the reality is that it’s not that either one is always higher, but that they can simply be polling at slightly different intervals.

And from the API perspective, nightbot can sometimes seem to take a really long time to respond: in the worst case, the perceived delay can be up to 20s (10s for nightbot to get the message, plus *another* 10s for the nightbot response to make its way back to you). This isn’t the fault of nightbot in any way: It is simply the limitation of YouTube’s APIs.

When YouTube had latency of 30s, this difference didn’t matter as much -- usually you would only be affected by a few seconds of latency, which -- out of 30s -- didn’t really matter. With the Low Latency option for streaming, however, the chat delay can sometimes be almost as long as the streaming delay, making chat a poor method of measuring the actual streaming delay.

I can hear the lamented cries coming now though: “What should I do then?” you will ask me. “How can I tell what the video delay *actually* is?”

Why, my friend, I’m glad you asked.

Just this week YouTube enabled latency display in the web player “Stats for Nerds”. This is a timer from “When the data arrived at YouTube servers” until it’s actually displayed on your screen. To see this for yourself, just right click the player, and click “Stats for Nerds”. The best part about this? It tells you right up front whether the broadcaster is using low latency streaming or not.

You can also use more instantaneous communication systems to check this -- for example, instead of having people type in the chat, instead, have them hop into a voice chat on discord with you, and manage it that way.

And of course, the best way to check though is just to compare the delay from when *you* do something until *you* see it on the video.

I did a demo of this last one recently -- using the “stopwatch” built into Google Search (just search for “stopwatch”), I showed the timer in my local browser, and then watched my own stream to see how far off it was. You can check out the video at the end of this post -- what I observed is that with an 11s reported latency from YouTube, my actual latency was around 12.12 seconds. (OBS encoding time adds about a second as well; I don’t know about Xsplit.) This means that generally, the “Latency” in stats for nerds will be a pretty good sense of how long it takes from when the broadcaster says something until it gets to you.

Of course, you still have to wait for the chat to update in order to see something. But if you just want to find out how long it takes until your viewers actually hear what you say, definitely make sure you’re looking at the right thing. Whether that’s the end to end latency reported in the player, or just using a stopwatch to confirm yourself exactly how long it takes, measuring latency can be a tricky business -- and when measuring, you want to make sure that you know what you’re measuring. Using stats for nerds, or using a stopwatch and watching your own stream, are the best ways to quickly get a sense of delay.

[chris-plays-games]

How to get a million views

It’s pretty simple.

Fly a drone.

Have an animal attack it.

Have the drone fall out of the sky.

Before I started doing video game livestreams, I had a different YouTube channel for a long time. In all of that time, my top 10 videos each have at least 2000 views... but really, only one of them is worth talking about: Hawk vs. Drone, with 5.1M views.

Why does this video have so many views? The formula is right up there: technology vs. nature gets views, if nature loses. In this particular case, I got more than 1 million views in a *single day* on October 10th: this video was shown on the ABC Nightly News, the Today Show, and licensed for TV on 5 continents.

I’m not alone here:

Drone vs. Kangaroo: 2.4M views

Drone vs. Eagle: 5.9M views

Drone vs. Goat: 2.4M views

Police Eagle vs. Drone: 1.4M views (and not even in English)

The key elements that matter here are that an animal enters the picture... and the drone loses.

Does this help you at all if you are making video game videos? No, not in the slightest. And I think that if you tried to *arrange* such a video, you would quickly have every animal activist in the country up in arms (with good reason; I’d be standing right there with a pitchfork).

But if you do fly, and you happen to get that once in a lifetime image: you’ve paid for your hobby for the next year or two. (Of course, this is why I now have 7 half-assembled drones in my basement...)

[chris-plays-games]

Getting the first viewers

People often ask the question: “How do I get started with livestreaming on YouTube Gaming? I’ve been doing this for a month, and still, nobody is watching, and I do all of my SEO/tags/title/description correctly!” My answer is generally: Getting the first few viewers is something that needs to come from you, not from YouTube, search results, or anything else. Until you’ve demonstrated to a human being that your content is worth watching, you’re not going to be seen by any algorithm as being worth watching, not when there are thousands of people more talented than you (or I!) in doing what we’re doing.

The hardest viewer to get is the first consistent viewer. Having that first viewer helps a lot, because you are no longer talking to yourself: Instead, you have an audience. In the current world of YouTube Gaming, it also makes the difference between “Being in the list” and “not being in the list” in a lot of cases - e.g. the /live page on YouTube gaming only lists 100 streams, ranked by magical YouTube special sauce, but having a viewer at all tends to put you on the page. 

So in order to start getting more viewers, you have to start by having viewers.

Once you have a viewer, how do you make them a consistent viewer? It’s all about you. You're the producer, director, and (usually) only actor in a TV show that you're broadcasting. You need to be maximizing the impact of you -- otherwise, you're not going to be as interesting as the next 10 people who are doing that. This doesn’t mean it has to have a camera feed, but your goal is interaction: People don't watch a live stream because of the gameplay, not for long. They watch it because of the broadcaster. (Exceptions exist, if you are a ridiculously great gamer, or playing a brand new game, but they’re exceptions, not the rule, and are rarely going to happen when you’re playing a popular game.)

So, to that end, what makes you special? Are you funny (and funnier than the other 20, or 200, people doing the same thing)? Are you interested in game development, and able to give in-depth renditions of the specifics behind the physics engine of the game? Is there something else in your life that's interesting? Are you simply the best goddamn GTA racer on the planet? What's your hook?

Once you figure that out: talk yourself up in shared social spaces that value those traits.

My hook is that I play retro games. From the NES and Genesis era back to Atari, I play games that predate some of my audience members being alive. (Literally. A 19 year old was watching me play Doom for a while; Doom came out 21 years ago.) I've found an audience by chatting with people about their favorite games from growing up... and discussing these things with people, finding their favorites means I can often say "Oh, I've played that on stream."

I've pulled audience from friends from "real life" -- High school friends, friends from past jobs, etc. I've pulled audience from interacting in chats with other broadcasters. I've pulled audience from participating in the YouTube Gaming subreddit, where I talk about streaming, subscriber growth, etc. I might have even picked up one or two subscribers from this very Tumblr.

I also am technically knowledgable about streaming, and the internet, and other things like that, so I talk about those things. I tell stories about my travels -- for conferences and personal travel. I tell stories about my work. I tell stories about my family: growing up in the Midwest, moving to the Northeast, becoming a parent to two step-kids at 19. These are stories that keep people coming back.

I started streaming on a schedule on November 9th. From November 9th to January 18th, I streamed once a week -- every Monday at 9:15PM EST, with 3 additional 'surprise' streams where I had extra time, plus a week of almost constant streaming (6 days out of 7) around New Years Day; a total of around 20 streams. During those two months, I grew from 73 subscribers to 450; and I grew from an average of 6 people watching me for the duration of that first stream (peak 11) to an average of 26 people watching me for my recent Monday stream (peak 54).

Lastly, and probably most importantly long term: Schedule. If I look at your channel, and I have no idea when you stream -- and it doesn't seem consistent -- I will probably not stop by to watch you, and if I don’t watch you, I probably won’t subscribe. If you want to build an audience, stream on a schedule. It doesn't need to be every day, though that is obviously the fastest way to grow: as I said, I started with every Monday, and I was able to grow without a problem, at least to the moderate size I am now. But you need to stick to a schedule: Same time, on a weekly basis, that people know they can check you out. This lets you get those first viewers coming back -- and that's the most important part.

So:

Figure out what's special about you. Emphasize that.

Find people who care, and get them to watch you.

Schedule exactly when your livestreams are so that everyone knows when they can see more of you.

Do these things, and you’ll start to see your audience grow.

[chris-plays-games]

Subscriber Growth

I began streaming on a regular schedule on November 9th, 2015. Prior to that, I had occasionally streamed, but never on a schedule. I also participated in Extra Life, live-streaming for 24 hours over the course of Nov 7th and 8th. At the time I began streaming on a weekly schedule, I had 71 subscribers.

As of today, I have 430 subscribers.  My 28 day rolling average of daily sub growth goes up by about 2 every month: In my first month, I was gaining 2 subs per day; in my second month, 4 subs per day, and I’m currently gaining an average of 6-7 subs per day. My most successful single day was 11/23/2015, where I gained 33 subs; in part this was because of a raid from another popular broadcaster. However, in general, many of my subs come when I’m *not* streaming:  - I have established myself as part of a community of broadcasters on YouTube Gaming, and regularly get and give shoutouts to other broadcasters. This basically means every time I say “hi” in a livestream chat, I have a chance to gain a subscriber, because the broadcaster will usually give a brief explanation of what I do (”retro games on Monday nights”).  - Typically, I raid another channel when completing my stream. Usually my concurrent viewcount is around 15 at the end of an evening, and about half that many follow me and say “#cpgraid”. This typically results in someone asking what I do, etc. and leads to more subscribers. If I were to stream daily, I would likely grow faster: during a week when I streamed 6 days shortly after Christmas, I received 70 subscribers (an average of 10 per day, rather than my more typical 6). During a typical stream, I will typically gain approximately 3-5 subs; typically a raid at the end will pick me up another 3-5. Of course, these numbers are not valuable on their own: subscribers are not interesting if they aren’t showing up to your streams. So in another blog post, I’ll summarize what my viewership looks like, which is a far more interesting metric to me.

[chris-plays-games]

YouTube Gaming

I’m a broadcaster on YouTube Gaming, streaming retro games on Mondays at 9:15PM EST. Part of the reason I started doing this is to understand all of the components that go into streaming video games -- technical, social, etc.

A lot of times people ask questions about this process, so I figure I can use this as a way to share some of my thoughts, etc. on running a small YouTube Gaming channel.