@astrophotocologne Channel Audit: 18,300 Subs, 404 Videos Analyzed

Looking at the raw numbers first. 404 uploads against 18,300 subs works out to roughly 45 subs per video lifetime, which sounds low — but that's the wrong way to read it. Frank's channel description states he uploads once a week, usually Mondays, and 404 weekly uploads is about 7.7 years of unbroken cadence. Channels that survive that long in a slow-growth specialty niche almost always grew the subscriber base in a long, flat curve rather than from any single breakout. The cadence itself is the story here, not the per-video average.

Honestly, the scraper handed me a weird artifact I should call out before going further — total channel views came back as 18,566, basically equal to the subscriber count, and the ten most recent uploads pulled with empty titles and zero views. That's almost certainly a German-locale or fetch hiccup rather than a real channel state, because no creator reaches 18.3K subs without lifetime view counts that dwarf the sub count. So I can't pattern-match the recent video titles directly, and I'm not going to invent any. What I can analyze is the structural positioning, the cadence, the niche choice, and the description signals — which is honestly the more interesting layer for an audit anyway.

The niche positioning is unusually narrow, and that's a compliment. Frank explicitly identifies as a Cologne-based astrophotographer focused on Deep-Sky imaging — not planetary, not solar, not landscape astro. That's a niche-within-a-niche-within-a-niche. Deep-sky is the long-exposure, narrowband, hours-of-stacking corner of astrophotography where the addressable audience is small but the watch time per viewer is enormous. People doing deep-sky will watch 30-minute processing tutorials end to end because they're stuck on a specific problem with their own data. His description lists three pillars — Technik, Bildbearbeitung, Philosophien — covering exactly the three things a deep-sky imager actually loses sleep over. That's sharp niche-vocabulary, not generic creator-speak.

The German-language choice is the most interesting strategic constraint on the whole channel. Astrophotography YouTube globally is dominated by English-language giants — AstroBackyard, Cuiv the Lazy Geek, Nebula Photos. Frank is competing in a much smaller language pool but with much lower competition for that audience. 18,300 German-language deep-sky subscribers is structurally equivalent to maybe 80-100K subscribers in the English-language market, because the addressable population is that much smaller. The flip side is defensibility — English-speaking creators in the space basically cannot erode his base because they don't speak the language. That's an underrated moat.

Where I'd want better data: upload-to-upload view variance, whether the weekly cadence has actually held the full 7.7 years or has gaps, and whether shorts adoption is zero by deliberate choice or by neglect. The content mix shows zero shorts in the last 30 uploads — all long-form. For a deep-sky creator that's defensible, because the niche rewards depth and a 60-second short on narrowband filter selection is genuinely useless to the viewer. But shorts also do the algorithmic discovery work that long-form can't carry alone, especially in a language market where the long-form audience is naturally capped earlier. A short answering a single deep-sky FAQ — "what is a flat frame and why does my image need them" — would probably feed long-form views without diluting the brand at all.

If I'm picking one forward-looking observation, the bigger lever for a channel like this in 2026 isn't shorts theater or thumbnail-optimization — it's whether the weekly upload still has a clear topic ladder behind it. After 404 videos, the catalog risks becoming a flat archive where every video is equally hard to find and there's no obvious start-here path for a new subscriber. The deep-sky audience compounds hardest when a single viewer binges five or six videos in a session, because the skill stack is genuinely sequential — mount, guiding, acquisition, calibration, stacking, processing. A playlist architecture that mirrors that learning order, with the single strongest evergreen video at each rung, would probably outperform anything tactical I could suggest at the individual-video level.