What a 42x view gap says about your brand voice

The numbers

A streaming service wanted to know how to keep young men around between seasons. Arcus read men 18 to 24 in their own posts and comments on TikTok, Reddit, YouTube, X and Instagram, and counted what they did. On TikTok, that meant 293 posts from eight formats the audience runs themselves, like the boys’ group chat, the gym, the fantasy league and the boys’ trip. Each post was watched and sorted by hand. Three kinds matter here.

The kinds are simple. A bit is acted out and has a punchline. A sincere post is someone talking straight to camera about advice, motivation or a feeling. A brand post is a company account trying the same register; one in the set, a post selling an app, did 2,242 views, close to the brand median.

91,850 divided by 2,171 is 42. That’s the gap. Against a brand post it’s 55. The order held in seven of the eight formats. The biggest bit in the set, the boys’ group chat played as a hostage scene, did 22.5 million views. See the audience report.

Same like rate, 42 times the reach

Here’s the part most teams miss. Per viewer, a sincere post is liked as often as a bit: 107 likes per 1,000 views for sincere, 106 for the bit. The people who saw the sincere posts liked them. Far fewer people saw them.

So viewers didn’t rate the sincere posts lower. The difference is how many people each kind reached. And under a bit, the comments become a second act. A quarter of the most-liked comments across the study played along with a bit: extending it, quoting it back, or pitching it as a film. The most-liked of those, under a group-chat sketch, read “could be a amazing movie if it was a real movie” and took 464,928 likes.

The top comment under a bit carried a median of 242 likes. Under a sincere post, 3. Under a brand post, 1. When a brand posts, the room has little to join.

The sincere post that traveled

One sincere post in the 293 broke out. A bricklayer posted a clip of his 19-year-old apprentice asleep after a 90-day outage: 5,007,997 views. Nothing about it was a bit. What it had was a real person in the frame, someone the viewer could recognize.

That’s the useful exception. Sincere content can work for this audience when someone real is on screen and the brand isn’t the one talking. It doesn’t work as a caption from a company account telling people how to feel.

What it means for your brand voice

The report turned the counts into four rules. None of them asks a brand to be funnier than its audience.

1. Play along. Reply to the bit. Don’t start one. Be the callback, the riff, the pitch. A line the report offered as a model: “rewatch starts tonight. season 1, episode 1. the group chat has been warned.”
2. Tell, don’t ask. State a time and when it comes back. Don’t ask the room a question. Questions were 3.9% of the most-liked comments. The model is a line from one of their own threads, 44 points: “It should not even be a question. We tee off at X o’clock.”
3. Hand over the mic. Let a fan’s line be the caption. One of the service’s regional accounts posted a short line in the fans’ own voice for a men’s title and did 12.1 million views. The main US account’s biggest men’s post did 524,100.
4. No glossary. No slang in the caption. One word in 62 they write is slang, and they use those words to judge things. See how young audiences write.

The voice already existed inside the company, on its X account and its regional TikTok accounts. A short, flat line from the X account naming a season and episode, with a two-word sign-off, took more than 1,500 likes. The job was to move that voice to the main account. Most brands are in the same spot: someone on the team already writes like this.

There’s one more signal worth knowing. The audience grades the marketing. A comment giving the marketing team advice on how to bill the new season took 54,000 likes under one of the service’s trailers. When the marketing talked like a fan, the replies said so. They notice either way.

Run it on your own account

1. Pull your last 30 posts on one platform and the views for each.
2. Label each one: a bit, a sincere post, a product post, other. Watch them; don’t guess from the caption.
3. Find the median views for each label. Divide the biggest by the smallest.
4. Read the top comment under each post and note its likes. Where is the room joining in?
5. Find the accounts your audience already follows for the bit you want. Could you reply there instead of posting?
6. Rewrite your next five captions with the four rules. No questions to the room. No slang. A time, if there is one.
7. Compare medians after a month.

Questions worth asking your team: Which of our posts are bits, and how did they do against the rest? Who on our side already writes in the audience’s voice? What could we reply to this week instead of posting?

Common mistakes, and what this doesn’t prove

• Making the brand the comedian. Brand posts did worst of the three kinds. The rules are about joining, not performing.
• Writing off sincerity. The like rate was the same. Put a real person in the frame.
• Copying the format without the room. A bit works because the audience already runs it. Find theirs first.
• Reading averages. One 22.5 million view post would swamp an average. Medians keep the comparison fair.

The counts come from one audience, men 18 to 24, on public TikTok posts in eight formats, sorted by hand in September 2026. They show what gets watched and joined. They don’t show that a bit will move sign-ups or viewing for your brand. That needs your own numbers lined up against the posts, and a test. Listening that counts is covered in this guide.

To run this count on your own audience, try it free with a question like “What does our audience reward that we never post?” We line up what people say online with your own numbers and walk you through the answer in 30 minutes.