Live viewers come from the text, not the stream: how to get more viewers on a live stream

Sunny Ray · Jun 21, 2026 · 6 min read

Live viewers come from the text around the stream, not the stream. I figured that out by accident in June, when one stream did over a hundred live viewers and the next day's did close to 200.

On June 17 I interviewed a Bitcoin guy who had sold his real estate for Bitcoin. That one popped at around 120 live viewers. On June 18 I had a guest who had been in the room in Ethereum's earliest days and has spent years doing legal structuring for crypto companies. Close to 200 live. On our team call later that day someone said what I was thinking: even for an influencer, 200 live is a lot.

So how do you get more viewers on a live stream? Give the platform something to index while you are live. Here is exactly what I changed.

What I used to do, out of desperation

For months my growth tactic on X was embarrassing. While streaming, I would retweet my own live post, undo it, retweet it again. Five times an hour, hoping the algorithm would show it to someone new.

It did not work, and a teammate pointed out the obvious: platforms do not love that kind of behavior. It is the same family as follow-unfollow games. Low effort, a little spammy, and it teaches the algorithm nothing about what your stream is about.

Threading the conversation as it happens

A few days before those two streams I switched. I always prepare about 20 questions for a guest. Now, as I ask each one on air, I paste it as a reply under the pinned live post. One question every minute or two. By the end of the stream there are 20 questions threaded under the video.

My theory for why it works: X is a text-heavy place. A video with no words around it is invisible to anyone scrolling past or searching. A thread of sharp questions is text that people find, read and want answered. Someone sees a question they want answered and clicks into the live video to hear the answer.

Each new question also gives the post a fresh bit of activity. I could watch the count tick up a few viewers after each one. I do not know the algorithm's internals and I will not pretend to. I know the count moved when the text moved.

Pick a guest people already follow

The second variable is the guest, and I will not pretend it was all my tactic. A guest with history and a following brings their own audience, especially when they share and retweet the stream. I now ask guests before we go live to share it, and most are happy to, because it is their interview.

A fair caution on reading these numbers. Two days is not a trend, and the June 18 guest was a bigger name than the day before. But the threading was the only thing I did differently across both, and both beat anything I had done before.

Why a B2B stream doesn't need an audience

Here is the part that keeps this in perspective. Our streams do not exist for viewers. On a June 16 call we talked about running five streams a day, and the reason was never reach. We do them because the guests are customers or prospects we are trying to close. Earlier that month the number was three to ten streams a day, with deals closing directly from guests.

So a stream with three live viewers can still be the best sales call of the week. Viewers are a bonus. But a bonus that compounds: the guest sees a hundred-plus people watching their interview, feels good about saying yes, and shares it. The next invitation lands better because the last stream looked alive.

AI-produced Shorts: what the numbers looked like

The other text-and-clips experiment that month was Amir's. He built a fully automated YouTube channel to test a workflow, on a topic he barely thought about. He asks for ten ideas, picks one, and the system researches it, scripts it, records a voiceover, makes the visuals, assembles the video, uploads it and sets the thumbnail, title and description.

His summary on our June 18 call: "the shorts are killing it." Some were sitting at 1,000 to 1,500 views, fully AI generated, within days of launching.

The lesson for a business show is not to fire your editor. It is that short-form distribution is now cheap enough to run on every stream. A long interview becomes a handful of clips, each with its own title and text, each another door back to the guest and the show.

Open with energy, not with logistics

One more change that month came from a teammate's note on my videos: introduce the guest and get people excited before diving in, instead of opening cold with a quick hello and the first question. So now the first minute is a short, energetic setup of who this person is and why the next hour matters.

It sounds small. It is the minute that decides whether someone who just clicked in from a threaded question stays. It also forced me to talk to the camera every day, which I told the team was my holy grail: being able to switch the camera on and just go. I was not there yet in June. Doing it daily is how you get there.

Turn the stream into text afterwards

The stream ends. The text should not. After every episode:

• Keep the question thread live under the video so late searchers find it.
• Pull the transcript and publish it somewhere crawlable.
• Cut clips from the three best answers and title each with the question it answers.
• Send the guest the clips the same day so they post them too.

A live stream is an hour. The text around it can work for months.

If you want to see how one interview turns into a stream, clips, posts and a follow-up, take the tour of the show system.

Frequently asked questions