Update, November 2026: display ads went live on the site on 3 November 2026, and traffic has since passed 25,000 pageviews a month.
Socionics Insight is now doing roughly 20,000 pageviews a month. In April it was around 5,000.
That's the headline, and headlines are the least interesting part. What follows is the version I'd have wanted to read a year ago: what actually moved, what I got wrong, and why growth and revenue turn out to be two separate problems.
The growth
Nothing clever happened. No viral post, no backlink campaign, no growth hack.
What happened was consistent publishing on a subject almost nobody covers properly in English. Socionics has a serious literature in Russian and Ukrainian, a scattered presence on a handful of forums, and very little that a curious English speaker can land on and trust. That gap has been the whole strategy.
The search numbers tell the story more precisely than the pageview count. Average position moved from 9.34 to 6.25 over the same period. Click-through rate roughly doubled, from about 1% to 2.17%. Those two things compound: better positions show you to more people, and better titles convert more of them.
The counterintuitive bit: I deleted a third of the site
At the end of July, Socionics Insight went from 372 indexed pages to 206.
I want to be careful about how I describe why, because I first described it wrongly to myself. My first instinct was that the removed pages were "thin content", using low traffic as the measure. That's a bad measure. Content quality and search demand are different things. A page can be excellent and simply answer a question nobody is asking.
The better way to see it is as a ratio: how much of what Google has indexed is actually earning anything. Pages that were crawled and declined, near-duplicate comparison pages, and individual pages that belonged behind a hub rather than in the index. Those went, with redirects where a redirect made sense.
The site got smaller. The traffic didn't.
The part nobody warns you about: geography
Here's where the tidy growth story runs into economics.
Display advertising isn't priced on how many people read you. It's priced on where they are. Most premium ad networks want a majority of your traffic to come from the US, UK, Canada and Australia. Fifty per cent is a common minimum.
Socionics Insight sits at about 27%.
That isn't a failure. It follows directly from what the site is. Socionics has its deepest roots in Eastern Europe and its most active new interest across South and Southeast Asia and Latin America. An English-language socionics resource is going to be read globally. The audience is exactly the one I wanted. It just happens to be one that premium ad networks pay very little for.
So one network was ruled out on those grounds alone. That was a useful thing to learn early rather than after a rejection email.
Where it stands
I'm currently working with Newor Media, a Google Certified Publishing Partner, to get AdSense set up properly. The site-level review is with Google as I write this. Approved isn't the same as pending, and I'm not going to describe it as approved until it is.
What I would say is that this is deliberately a baseline, not a business. Ads on a site with this traffic mix are a modest passive floor. The real revenue lines are the typing service, the books and Socion, the matching app. Advertising just stops the hosting bill being a pure cost.
What I'm watching next
Three things, all unresolved.
Desktop click-through rate is well below mobile on the same queries at the same positions. The most likely explanation is AI Overviews absorbing US desktop clicks before anyone reaches a link. If that's right, it's structural rather than fixable, and the response is to be the place people go after the summary rather than instead of it.
All sixteen type pages share the same click-through problem: 0.68% to 1.22% at positions six and seven. Sixteen pages failing in the same way isn't sixteen problems. It's one template problem, and it's next on the list.
And the articles section is where most of the crawled-but-not-indexed pages live. That's the clearest remaining sign that something in there isn't earning its place.
The thing underneath all of this
The moat is accuracy. Socionics attracts a technically literate, sceptical audience who will catch an overclaim immediately and never quite trust you again. That applies to the theory, and it applies to writing like this too. Which is why this issue says "in review" rather than "approved", and why the pruning explanation is the one I ended up with rather than the one I started with.
Growth is easy to narrate after the fact. Most of it was showing up.