Build-in-Public, Week 7: One Article Got 22 Views This Week. The Other Four Got 3 Combined.

I publish one article a day for Background Camera RemoteStream — a small Android app that turns an old phone into a local-only security camera. I've shipped 87 of them now. Every Saturday I stop and look at one number instead of all of them, because all of them at once tells me nothing.

This week the number is 22.

That's how many views one article got in its first five days. The other four articles I published this week got 3 views combined. Same author, same app, same week, same subreddit-and-tag distribution. One of them outperformed the rest of the week put together by roughly seven to one.

I want to be honest about what that does and doesn't mean — because the easy version of this post ("I cracked the algorithm!") would be a lie, and the boring true version is more useful.

The five articles

Here is the whole week, oldest to newest, with real numbers pulled from the dev.to API this morning:

Published	Article (shape)	Views
Jun 22	"My Arlo Subscription More Than Doubled — Do I Have to Keep Paying?" (question)	22
Jun 23	"Is My Blink Camera Useless Without a Subscription?" (question)	0
Jun 24	"FOREGROUND_SERVICE_TYPE_CAMERA on Android 14+" (technical deep-dive)	3
Jun 25	"How to Watch Your Home or Pet While Traveling" (how-to)	0
Jun 26	"The Tuesday the Spare Phone Earned Its Keep" (short story)	0

The obvious objection first: the Arlo piece is five days old and the story is one day old, so of course the older one has more views. That's real, and I'm not going to pretend a one-day-old article and a five-day-old article are a fair fight.

But age doesn't explain the gap. The technical deep-dive is three days old and sits at 3 views. The how-to is two days old and sits at zero. If time-on-page were doing the heavy lifting, those two would have crept up. They didn't. The Arlo piece pulled away on day one and kept going.

Why I trust the pattern: the all-time top says the same thing

If this were just one lucky week, I'd shrug. It isn't. Here are my two most-viewed articles of all time, across 87 published:

• "Turn Your Old Android Phone Into a Free Security Camera — No Subscription" — 67 views
• "What's the Cheapest Way to Set Up a Home Security Camera Without a Subscription?" — 64 views

Both are the same shape as the Arlo piece: a person, at the kitchen table, doing money math about a camera subscription they resent. The second one — phrased as a literal question someone would type — is also the only article I've ever written that grew real comments. It has 11. My other 86 articles have 7 comments between them, and most of those are mine replying.

So the week's "22 vs 3" isn't an outlier. It's the third time the same shape has won. The technical deep-dives I'm proud of, the short stories I enjoy writing, the tidy how-to guides — they're fine. They're just not what gets found.

The lesson, stated plainly

The thing that gets found is a title that is the exact sentence a frustrated person types into Google at 11pm. Not a clever title. Not a keyword-stuffed title. The actual complaint, in their words, with the dollar amount in it.

"My Arlo subscription more than doubled" beat a genuinely good Android internals article seven to one not because it's better writing — it isn't — but because thousands of people are, this month, watching their camera subscription climb and typing some version of that exact sentence. (Arlo's 2026 increase was real and steep: in the UK the plan went from £2.50 to £7.99, and in Norway it rose roughly 67%. People notice that on their card.) I didn't need to manufacture demand. I needed to be standing where the demand already was, phrased the way the searcher phrases it.

The deep-dive fails on discovery for the mirror-image reason: almost nobody types "how does FOREGROUND_SERVICE_TYPE_CAMERA work" at 11pm in distress. The handful who do are wonderful and I'll keep writing for them, but that's a relationship play, not a reach play, and I should stop being surprised when it doesn't spike.

What I'm changing next week

One change, because one-number weeks deserve one-change responses:

I'm titling for the complaint, not the feature. Going forward the lead pillar each week starts as a real question with the friction baked in — the subscription that jumped, the feature that got paywalled, the brand that changed the deal — and lands on the same honest answer I always give: an old Android phone plus a free app that stores video locally, on your own hardware, with no monthly fee and nothing in anyone's cloud. The app stays the same. The doorway changes.

What I'm not doing is abandoning the other four shapes. The story I published Thursday got zero views and I'd write it again, because the one reader it eventually reaches at the right moment is worth more than 22 strangers skimming. Reach and resonance are different jobs. This week just reminded me which lever moves which one.

If you're building in public with a tiny audience, that's the whole takeaway: pick one number each week, resist the urge to average it away, and let it tell you one thing. This week mine said stand where the question is already being asked.

---

Background Camera RemoteStream turns an old Android phone into a local-only camera — record with the screen off, view it in your browser over your own Wi-Fi, or stream to a private YouTube Live, with zero cloud storage and no subscription. It's on Google Play, and the story of building it is at superfunicular.com.

More in this series: the Arlo question that started this week, the cheapest no-subscription setup (the 11-comment one), and turning an old phone into a free security camera.