Every content marketing case study starts after it worked. This one is from month 6: roughly 50 articles, 3 frameworks, 828 search impressions, and 14 clicks. I’m publishing the full Search Console export because the content marketing results timeline everyone quotes is built entirely from survivors, and this is more useful to anyone currently building a property from zero and quietly panicking.
14 clicks.
That’s what organic search sent this site between 27 February and 29 July 2026, across roughly fifty published articles and three frameworks.
I write about content strategy for a living, I’ve built content functions that contributed to 48% marketing-sourced pipeline, I publish an AEO/LLMO framework about making content discoverable, and yet my own domain, at month six, produced fourteen.
Every case study you’ve read in this field begins after the hockey stick, and mine starts here, because the useful information is in the part everyone deletes.
What did six months of Search Console data show?
Almost no clicks, a steady stream of impressions, and a site sitting mostly on page two: impressions but no clicks, in its plainest form. Straight from Search Console, 153 days, web search only:
| Metric | Six-month figure |
|---|---|
| Clicks | 14 |
| Impressions | 828 |
| CTR | 1.69% |
| Average position | 18.6 |
| Days with at least one click | 11 of 153 |
| Days with zero impressions | 42 |
| Distinct queries generating impressions | 58 |
| Queries that produced a click | 1 |
| Pages that produced a click | 6 of 91 |
The single query that produced a click was my own name.
Why did I expect more?
Because every content function I’d run before sat on a domain that already had standing. The source of that expectation is the lesson.
Each of those companies came with years of accumulated links, established crawl history, brand search volume and other people’s work compounding underneath mine. At a climate tech SaaS company I grew organic traffic 22%, and at a safety-critical software company it was 28% year on year. Those gains arrived fast enough to report quarterly, and I carried an unexamined assumption that they came from the content, when they came from the content plus a domain with standing.
The second source was the sheer volume of case studies that start at month twelve, or month eighteen, or whenever the curve turned. Read enough of those and you absorb a ramp that doesn’t exist, because you’re only shown survivors. Nobody publishes month six from a new domain, so everyone building one is calibrating against a sample that leaves them out.
How long does a new domain take to rank on Google?
Usually more than a year for page one, and new pages have it harder than they used to. Patrick Stox at Ahrefs tracked a million URLs first seen by Ahrefs’ crawler and checked how many reached Google’s top 10 within a year: only 1.74% did. Filtered to English pages with content, that rises to 6.11%, which still leaves roughly 94 in every 100 pages off page one after twelve months.
The same study found 72.9% of top-10 pages are more than three years old, up from 59% in the 2017 version, and the average #1 result is five years old. Only 13.7% of top-10 pages were under a year old, down from 22%. Stox’s summary of the whole study runs to three words:
“Older content dominates.”
Measured against that, fourteen clicks at month six is the typical outcome, and I’d been judging myself against a curve that describes the luckiest 1.74%.
A six-month-old domain has, by definition, none of what older pages have: accumulated links, crawl history and time. That’s my reading of why the age effect is so strong, and it fits every figure in my export.
Why does a site get impressions but no clicks?
Because impressions arrive before position does. Google counts an impression whenever a result appears in the current page of results a searcher is looking at, so a page ranking 18th collects an impression from anyone who reaches page two. Very few click there: Backlinko’s study of around 4 million Google results found only 0.63% of searchers clicked a result on the second page.
That’s what the export shows: 828 impressions means Google is showing this site, and an average position of 18.6 means it’s mostly showing it on page two. Of the 58 queries generating impressions, 37 sit below position 20, and those account for 196 impressions that were never going to convert however good the page was.
The pattern repeats at page level, and it’s brutal in a specific way:
| Page | Impressions | Position | Clicks |
|---|---|---|---|
| Homepage | 237 | 3.8 | 9 |
| AEO/LLMO framework | 164 | 25.5 | 0 |
| 48% pipeline article | 123 | 19.3 | 0 |
| Blog index | 90 | 8.7 | 0 |
| CV | 51 | 6.4 | 1 |
| AI citation tracking | 51 | 47.4 | 0 |
My two strongest assets, the flagship framework and the pipeline methodology piece, generated 287 impressions between them and no clicks, because they’re ranking on pages two and three. Demand is there and the content holds up; position is the constraint, and position is what takes years.
The geography confirms it: the United States produced 452 impressions and one click at an average position of 19.2. Portugal produced 12 impressions and four clicks, a 33% CTR at position 2.7, and Spain returned three clicks from eight impressions at 37.5%, so people click at healthy rates wherever this site ranks near the top.
One embarrassing finding while I’m being thorough: the query “solangerainha.com” generated 85 impressions at an average position of 8.5, with no clicks. People are searching for my domain by name and finding it eighth. That’s a technical fix I hadn’t noticed before running this export, which is itself an argument for pulling the data before theorising about why traffic is flat. Branded search is also where a lot of AI-driven discovery comes back.
Could AI Overviews explain impressions but no clicks?
For established sites, partly, but not here. AI summaries do cut clicks: Pew Research Center tracked the browsing of 900 US adults in March 2025 and found people clicked a traditional result in 8% of visits when an AI summary appeared, against 15% when one didn’t. Seer Interactive’s data across 3,119 search terms from 42 client organisations shows organic CTR on queries with AI Overviews falling from 1.76% to 0.61% between June 2024 and September 2025, and dropping on queries without them too.
That’s a serious headwind, and it’s why the wider organic traffic decline hits some pages far harder than others. It doesn’t explain this export, though: the homepage at position 3.8 took nine clicks from 237 impressions, and the pages that don’t get clicked are the ones that haven’t reached page one.
Is the curve bending?
Yes, though one month isn’t a trend. I went into this export expecting a flat line, and it isn’t flat.
| Month | Clicks | Impressions |
|---|---|---|
| March | 2 | 29 |
| April | 0 | 24 |
| May | 3 | 131 |
| June | 1 | 196 |
| July | 8 | 448 |
July alone produced 54% of all impressions in the period and 8 of the 14 clicks, and impressions grew roughly fifteen-fold between March and July.
I want to be careful here, because this is exactly the point where a piece like this starts turning into a success story. 448 impressions is still tiny, and I’ve written before about not confusing a nice-looking metric with a working one. July could be a fluke, or an artefact of publishing more that month.
It’s consistent with the mechanism, though: impressions come first, positions improve slowly after that, and clicks are the last to move. Impressions but no clicks is what that sequence looks like from the inside, and a site at month six with rising impressions and an average position of 18.6 is somewhere in the middle of the curve, well off the bottom of it.
What should you measure instead of organic clicks?
Whatever the site was built to do, picked before you see the data. Optimistic measurement here would be celebrating 828 impressions as reach, and impressions are the classic vanity substitute in exactly this situation. Measuring properly means asking what this site exists for, and whether it’s doing that.
Organic traffic was never the point. This site exists to convince a small, specific audience: hiring managers deciding whether to interview me, and people checking whether the frameworks hold up. That audience doesn’t arrive by searching content strategy topics; someone sends them a link, or they read my name somewhere and type it in.
On that measure it has worked. The strongest opportunities I’ve had in this period came as inbound approaches from people who’d read the work before making contact, which is the owned-audience argument doing what it claims, and the reason a site like this pays off before a hiring process starts more than inside one.
The caveat applies to my own good news too, because that’s a small sample as well. What I can say is that the metric matching this site’s purpose is moving, while the metric that looks most embarrassing measures a job the site was never built for.
What would I sequence differently?
- Start the domain earlier and publish less at the start. The binding constraint is calendar time more than volume, so ten articles published eighteen months ago would beat fifty published six months ago.
- Make the navigational query the first fix. Ranking 8.5 for my own domain name is the cheapest win going, and I only found it by exporting the data, so check branded and navigational queries before writing anything new.
- Front-load the pages built to be linked to. The frameworks are what people reference and share, and they build the authority everything else depends on, so they should have come before the pages built to be found.
- Write expectations down at the start. An unwritten expectation can’t be tested, and I’d have caught the domain-authority assumption in week one instead of month five.
Why publish a bad result?
Because it’s more useful, harder to fake, and it makes any later good result mean more. The obvious move is to sit on this until the figure improves, then write the version with the hockey stick in it.
Someone at month four with forty articles and single-digit clicks is deciding right now whether they’re failing, and the data says they probably aren’t, but almost nobody publishes the evidence. The survivorship problem feeds itself: only the wins get written up, so the published picture of a normal ramp drifts further from the median every year.
It’s also harder to fake. Anyone can publish a good result; publishing fourteen clicks next to a framework about discoverability, with my name on both, is a bet that being checkable beats being impressive. I applied the same logic when I scored my own site against my own audit framework and published the failures. And if this site performs in eighteen months, that case study will be stronger for having a documented month six with an unflattering figure attached.
Figures are the complete Google Search Console export for solangerainha.com, web search, 27 February to 29 July 2026. I’ll publish the next reading regardless of what it says.
Frequently asked questions
Longer than most case studies suggest, because they’re written by survivors. Ahrefs found only 1.74% of newly published pages reach Google’s top 10 within a year, 72.9% of top-10 pages are over three years old, and the average #1 result is five years old. On a new domain with no accumulated links or crawl history, single-digit or low double-digit organic clicks at six months is a normal outcome, not a sign of failure.
Because impressions arrive before position does. Google counts an impression whenever a result appears on the page of results someone is viewing, so a site ranking on page two collects impressions from the few people who get that far, and only 0.63% of searchers click a second-page result in Backlinko’s data. On this site, 828 impressions at an average position of 18.6 produced 14 clicks, with 37 of 58 queries below position 20.
They contribute on established sites: Pew Research Center found people clicked a traditional result in 8% of visits with an AI summary present, against 15% without. On a new domain, position is usually the bigger cause. Check your page-one pages first: if they get clicked at healthy rates and your page-two pages don’t, the problem is ranking, and AI summaries are a secondary headwind.
Measure impressions and average position, because they move first and tell you whether you have a demand problem or a position problem. Then measure against what the site was built to do: for a site whose purpose is credibility with a small audience, direct visits, referral traffic and inbound approaches count for more than search clicks. Pick the metric that matches your goal before you see the data.
