New: YouTube Storytelling: Beyond Retention - Course Now Available

Strategist answers

Our finance channel is declining at 460K subs, yet one long video takes off while a similar one barely moves and Shorts vary too; how do I work out what's behind those swings?

For 300K-500K Personal finance / investing channels; small team (2-5); long-form + Shorts; declining.

86% of this answer is quoted word for word from 6 YouTube strategists. Each quote links to its source.

For a 460K finance channel, start with how a video earns wider distribution, and why the signals shift as it spreads.

“If the video performs well in the data here it will be shown to more of this group. The viral moment comes when YouTube starts to show this video to a wider audience and it still performs really well. This tells you that the video appears to a lot of people, not only fans of Matt Rober, so they start to recommend it to more and more people who they feel will enjoy the content.”
Paddy Galloway (@PaddyG96), "How Mark Rober is beating the YouTube Algorithm (Genius Strategy)"; transcribed from a machine transcript
“If a video does 10m views +, on some channels I work with CTR can easily drop by 5-10%. AVD drops as well with how much colder the viewer becomes naturally. After working with 500+ channels I can tell you it really varies which element impacts what. There’s no one size fits all”
Leroy ter Braak (@LeroyterBraak), X post, 2026-10-09

With a small team judging every upload, it helps to know how uneven results are even on strong channels.

“Most people fail on YouTube because they think its a game of always getting more views on every video you make. Yet when I analysed 40 world-class channels, only 3 in 10 videos performed above their average.”
Ed Lawrence (Film Booth) (@Ed_FilmBooth), X post, 2026-06-29
“The algorithm is just audience behavior at scale. Improve the moments where viewers lose interest, but more importantly, study why some ideas work and some don’t.”
Leroy ter Braak (@LeroyterBraak), X post, 2025-11-20

Dabi on reading the analytics behind two similar long videos:

“What should you use instead of subscribers for channel analytics? Well, there’s a lot of metrics that even advanced YouTubers easily get wrong, like AVD and CTR. I don’t blame people for looking at these, they’re freaking hard metrics to interpret correctly and there’s not really a good evidence-based source of truth for how to do so yet.”
Dabi (@dabidoYT), X post, 2026-05-14
“Understand that a retention curve is the total duration divided by 100 points. So for example, a 100 minute long video means each point is a full minute! (in 2026) That is important for retention analyses.”
Dabi (@dabidoYT), X post, 2026-10-21

On the Shorts side, Daniel Faraday and Paddy Galloway both keep the two formats apart:

“both groups have winners and losers. shorts share doesn't sort them. the two biggest risers do have something in common: long, dense training for one kind of buyer. haynes posts hour long meta ads training. gordon moved to operator interviews and full sales courses. so if your long videos are falling, shorts won't rescue them. fix the long videos.”
Daniel Faraday (@FaradayYouTube), X post, 2026-11-02
“Using short form to inform long form. They are two different beasts, just because something is viral on short form doesn’t mean that topic will convert. To get great at long form you need to study what is doing great in… long form!”
Paddy Galloway (@PaddyG96), X post, 2026-02-03

Not every swing comes from the video itself; Mario Joos points to the feed, while Daniel Faraday and Ed Lawrence point to broad hits fading.

“In many cases, now only 2 long-form recommendations are shown, with the second row already taking up slots for short-form recommendations (see second image). Because of this, the ratio of long-form versus short-form often skews toward 2 long-form to 5 short-form recommendations on PC. This means that we're seeing a choice reduction of up to 80% long-form recommendation compared to previous home feeds.”
Mario Joos (@MarioJoos), X post, 2025-12-21
“this doesn't prove drift causes decline. hormozi is broad and held flat. but a broad hit can fade as fast as it came. for an expert channel, the topic that fits your offer is your floor. keep most uploads on it while you test anything wider.”
Daniel Faraday (@FaradayYouTube), X post, 2026-10-01
“The truth is very few people manage to maintain the exact level of reach and attention that originally blew them up for more than just a few years. And this is incredibly hard to avoid because of something called hedonic adaptation. What that means is people quickly normalize what once felt special to them.”
Ed Lawrence (Film Booth) (@Ed_FilmBooth), "Why Growing A Personal Brand Is An AWFUL Idea"; transcribed from a machine transcript

Read The YouTube Algorithm: Explained to work through those inconsistent views. Dabi's guide explains CTR, AVD and satisfaction signals, helping you replace competing algorithm myths with mechanics. Algorithm guide