Nearly seven years on YouTube with a full-time job: what actually changed at 1k, 10k and 100k

Short answer, because you searched for it: with a full-time job, a tutorial channel took me nearly seven years to reach 100,000 subscribers. The first video went up in 2020; the Silver Play Button arrived in September 2026. The first 1,000 took the longest. Growth was never smooth, and the thing that changed most along the way was not the algorithm, the thumbnails or the camera. It was what happened under the videos.

This is the honest version, with numbers from my own YouTube Studio.

Who this is from

I run a channel about DevOps and cloud: Terraform, Kubernetes, AWS, Google Cloud, lately AI tooling. I have a full-time engineering job, a family, and I live in Sweden. Videos get made in the evenings and on weekends. No editor for most of these years and no team. The channel started, in my own words at the time, as me explaining Kubernetes to an audience of exactly nobody.

If you are reading this at 300 subscribers wondering whether it is worth continuing, this is written for you.

0 to 1,000: the years nobody watches

The first 1,000 subscribers took 14 months, the longest stretch of any milestone. I was uploading tutorials on topics I was learning at work, and most of them got a few hundred views from search over months, not days.

What worked, in hindsight, was one boring decision: answer a specific question completely. "How to install Terraform on Ubuntu" beats "Terraform tips" every time in a tutorial niche, because the person arriving has a problem, not curiosity. My most-viewed early videos were all like that.

What did not work: everything I tried because someone on YouTube said it worked. Posting schedules, hashtags, trend-chasing. In a technical niche, search intent does the heavy lifting and the video has to deliver.

Comments in this period: a handful a week, and I answered every one within hours. They were the best part. Someone in another country writing "this got me through my interview" is a strange and wonderful thing to read at 11pm.

1,000 to 10,000: series changed everything

The jump to 10,000 came from series. Instead of one-off videos I started numbering them: Part 1, Part 2, and so on, on one topic. When someone lands on Part 5 of the AWS VPC series, they go back and watch 1 to 4. Watch time per viewer went up, and the algorithm started recommending the series to people who had watched one part.

Comments in this period: a few dozen a week, and now a real share were technical questions about my videos. "Step 4 fails with a 403." "Does this work with the new version?" I still answered nearly all of them, but I remember the first weekend I opened Studio and felt behind.

10,000 to 100,000: the channel became a second job

I wish I could say one thing caused this stretch, but it was consistency over years, plus a few videos that did much better than average and pulled the rest up. The AWS networking series, the Helm chart tutorial, the Google Cloud series, and later anything about AI tooling for engineers. Three videos from 2023 still sit at the top of the channel by views, each above 230,000.

By this point a video was doing well if it pulled a few thousand views in the first weeks, and the back catalogue was doing the rest. Today the channel gets around 250,000 views a month, and lifetime it has passed 7.4 million views and 770,000 hours of watch time. Most of that comes from search, on videos I made a long time ago, and most of it from people who have not subscribed.

Here is what changed that nobody had warned me about. Comments scaled with the back catalogue, not with the new uploads. Every old video keeps collecting questions. At 100,000 subscribers I receive around 200 comments a week, and because it is a technical channel, most of them need a real answer.

I did the arithmetic one evening. A proper answer to a technical question takes me about 3 minutes. 200 a week is 10 hours. I have a job. So I did what I suspect most creators do: I answered the newest ones, felt guilty about the rest, and let them sink. Studio can filter to unanswered comments, but with a few hundred videos that is one endless scroll with no count, and answering each one from a blank box is the part that takes the hours.

What did not change in all those years

• Thumbnails matter less in tutorial niches than in entertainment. Clear text beats a surprised face.
• The best video ideas came from comments. The question that repeats three times in a week is the next video. Always was.
• You cannot outsource your voice. Editors, yes. Thumbnails, yes. Replies to viewers, no. A viewer can tell.

What I would tell myself at 1,000 subscribers

1. Pick a niche where people search with a problem. Serve the problem.
2. Number your videos. Series are the closest thing to a growth hack I have found.
3. Decide early how you will handle comments, because the volume arrives suddenly and the guilt arrives with it.
4. Keep the job. The channel is better for the fact that it never had to pay rent.

About the comments, since that is the part that surprised me

The comment problem is the one I ended up solving with code, because I am an engineer and that is how we cope. I wrote a small tool that pulls every comment across the whole channel that I have not replied to into one list, drafts a reply for each in my own words, and lets me edit and approve them one at a time. Nothing posts on its own, ever. I was strict about that: a reply I have not read is not really from me, and auto-reply bots get channels flagged.

That tool became ReplyDial. It took my comment time from about 10 hours a week to about 2 and a half, and I finally answered the old questions too. I mention it because it is the honest end of this story, not because this post is about it.

This week, without any tool: open one video from two years ago and read its comments. You will find a question. Answer it. That is the habit, and everything else is scale.

If you want to see your own number first: the free unanswered-comments checker shows how many comments on your last 20 videos never got a reply. No sign-in. Mine was uncomfortable.