I Published 31 SEO Blog Posts in Six Months. Google Sent Me 10 Clicks.

In March I launched PageDrop, a one-person service that builds one-page websites for local businesses for $297, once. I'm a developer with 20 years behind me, I knew SEO mattered, and I did what every guide says: pick low-competition keywords, publish consistently, make the pages fast.

Six months later I have 31 blog posts, each checked for 95+ on mobile PageSpeed when it went out, structured data on all of them, a sitemap Google reads within hours of each deploy, and 35 of 36 URLs indexed. The 36 are the 31 posts plus the homepage, the blog index, a Romanian version of the homepage, a landing page and the privacy page.

Google has sent 10 clicks. Every one I can attribute came from someone typing some spelling of the brand name. No blog post has ever received a click.

The short answer: publishing volume does not create rankings on a domain nobody links to. Thirty-one technically clean, well-researched posts produced 2,555 impressions, 10 brand clicks and zero non-brand clicks in six months. Keyword difficulty scores, page speed and schema were all beside the point. Links were the point.

None of that is rare. Ahrefs found that 96.55% of pages get no traffic from Google at all, across about 14 billion pages. What that study cannot tell you is what the failure looks like from inside one small site, query by query and day by day: which pages Google tries and then drops, which of your best queries are robots, and which numbers in your own dashboard are lying to you. That is what the rest of this is.

This is the post I wish I had read in March. All numbers come straight from the Search Console API, with my own country excluded, through 22 September 2026.

The Numbers

2,555 impressions and 10 clicks across six months, spread over 171 distinct search queries. Seven of the ten clicks are attributed to a query, and all seven are brand spellings: "pages drop" three, "page drop" two, "pagedrop" one, "pagesdrop" one. The other three are in the anonymised bucket Search Console does not break down.

Of the 31 posts, 17 have never been shown to anyone at all, and 22 have never been shown more than five times in total, ever. All 31 together hold 753 impressions, and nine of them account for 745 of those. The homepage alone has 908 - more than every blog post on the site combined.

On the business side: two paying clients since launch, both through people I know and a listing on Upwork. Zero from Google. The site that sells search-friendly websites has sold nothing through search.

Lesson 1: "Easy" Keyword Difficulty Means Easy for Someone Else

My content plan came out of a free keyword difficulty tool. The finding felt like a cheat code: add a trade to the query and the difficulty collapses. "Restaurant website design", "salon website design", "plumber website design" - all marked Easy, all with real volume. I wrote a pillar post for each trade, then a cost post, then a checklist. The painter cluster is complete: three posts, internally linked, answering everything a painter could ask about a website.

Those three posts have had zero impressions. So has every plumber post and every handyman post. The whole electrician cluster has had exactly one impression in six months.

Keyword difficulty scores estimate how many referring domains the pages on page one have, which is the same thing the Domain Authority and Domain Rating numbers are built on. "Easy" means the pages ranking there don't have many. It does not mean a domain with two referring domains, both of them friends, can join them. Difficulty is measured against the pages already ranking, never against your site.

Lesson 2: Google Tries Your Page, Then Decides

The June spike in the table is one post: a breakdown of what Wix really costs over three years. It went live on 21 May, and from 29 May it was showing up for 101 different Wix pricing searches in the US and UK - "wix transaction fees", "wix cost per month", "wix domain renewal price", variations I would never have thought to target. It collected 396 impressions in thirteen days.

It was never good placement. The best position any of those queries reached was 47; the median was 73. On 10 June it went to zero and has stayed there for over three months. The page is still indexed. Google crawled it again this month. It simply stopped showing it.

A restaurant website checklist did the same thing more slowly. It held position 10 to 20 for "restaurant website checklist" on seventeen days between 30 May and 27 July - the closest anything on the site has come to page one - and then vanished.

The first weeks of impressions for a new post are not a trend. They are a trial. If the domain has no authority to back the page up, the trial ends.

Lesson 3: Your Own Searches Poison Your Data

The first time I looked at Search Console, several pages sat at an average position of 1 to 3. They were my own searches: I am in Romania, I searched for my own site to check it was there, and Google averaged my position-1 results in with everyone else's position-60 ones.

Excluding one country moved some page averages by 40 places. Every number in this post has Romania filtered out for that reason. Filter out your own country, or your own city if you sell locally, before you believe a single position figure.

Lesson 4: Some of Your "Best" Queries Are Robots

My two biggest non-brand queries were phrasings of the same question: "why are wix websites hard to rank on google" at 49 impressions, and a near-identical version ending in a question mark at 44. Ninety-three impressions between them, ahead of every other non-brand query on the site. Real demand, I thought, and I wrote a post answering why Wix websites are hard to rank on Google.

Then I looked at both by country, device and day. The exact phrase: Australia, desktop only, one to four impressions on 31 separate days between 1 August and 17 September. The question-mark variant: New Zealand, desktop only, one or two impressions on 40 days since 3 June. Never a phone, never a third country, and never a spelling anywhere between the two.

Real people don't search like that. Rank tracking tools do. Somebody, probably an agency tracking Wix rankings, is checking those phrases from two countries, and every check lands in my Search Console looking like demand. Real demand looks like the Wix pricing post above: 101 messy variations in a fortnight. A phrase that turns up almost daily from one country, on one device type, at a tiny fixed volume and with no variants, is a tracker until proven otherwise.

The punchline: the post I wrote to serve that "demand" has had 11 impressions since it went live. Six from New Zealand, five from Australia. I wrote a page for two robots, and the robots are its entire audience.

Lesson 5: Check Your Name Before You Buy the Domain

"pagedrop" ranks around position 16 for its own name. The top results belong to pagedrop.ai, pagedrop.io, page-drop.com and a GitHub repository. Most of the brand impressions I do get arrive through spellings like "page drop" and "pages drop", which is also where the clicks came from.

A name that is only yours in the .pro namespace is not only yours.

Lesson 6: Speed Is Table Stakes, Not a Ranking Strategy

This one hurt, because speed is my whole pitch. Every post is checked on mobile Lighthouse before it goes out: 95 or above, zero layout shift, and a paragraph of text as the largest element rather than an image. None of that got a single post onto page one.

Mobile PageSpeed and Core Web Vitals are a tie-breaker between pages that already deserve to rank. On a domain nobody links to, there is nothing to break a tie with.

What I Got Wrong, in One Sentence

I treated SEO as a writing problem when it was a trust problem. Thirty-one posts on a domain with two referring domains is thirty-one pages Google has no reason to believe.

One caveat I cannot argue my way out of: this data cannot separate "nobody links to it" from "not good enough to be linked to". Search Console reports impressions, clicks and position, not whether anyone who landed on a page found it useful. The restaurant checklist showed at position 10 to 20 on seventeen days and earned no clicks, and a better page might have earned some.

What I'm Doing Now

I stopped writing trade posts. The remaining ones in the queue are paused. I only write a new post when Search Console or autocomplete shows people actually typing the query, and when the current top ten is weak: forum threads, outdated pages, results that don't answer the question.

I'm working on links instead. Not bought ones. The places I can honestly put a link:

• my open-source libraries' documentation sites, some of which get real traffic (world_countries has 1,600 stars). One thing I learned doing this: links in a GitHub README carry rel="nofollow", so they pass nothing. The separate documentation sites on github.io don't have that restriction.
• a credit line on client sites, where the client agrees
• posts like this one, which is the kind of thing people link to when it's useful

I measure with free tools only. The Search Console API gives everything the web UI does, with filters the UI makes painful. For keyword ideas I use Google and Bing autocomplete, expanded with a small script that sends each seed with a-z and question prefixes and counts how often and how high each suggestion comes back. It is not search volume, but it is real queries, and it cost nothing.

If You're Starting a Small Business Site Today

1. Get your first five links before your fifth post.
2. Check what your brand name ranks for before you buy the domain.
3. Exclude your own country in Search Console on day one.
4. Treat the first month of impressions on any post as a trial, not a result.
5. Look at your top query by country, device and day before you write a post for it.
6. Write for queries you can see people typing, not for scores in a keyword tool.

None of this means the pages were wasted. They are the reason a visitor who does land on the site finds answers instead of a pricing page, and they cost nothing to keep. They are just not a traffic strategy, and I sold them to myself as one.

I'll post an update in three months with the same table, whatever it says.

Frequently Asked Questions

How long does it take for a new website to rank on Google?

Longer than six months if nothing links to it. This site was indexed within days and never held a position better than 10 for any non-brand query. Indexing is fast; earning enough trust to rank - what people call the sandbox effect - is the slow part, and more pages do not fix it.

Why is my blog getting impressions but no clicks?

Almost always because the impressions are at positions nobody scrolls to. Mine averaged position 60 to 75. At that depth a page is technically "showing in search" and will never be seen. Check the average position on your impressions before treating them as an audience.

Does keyword difficulty matter for a brand new site?

Not in the way the score implies. Difficulty estimates how many referring domains the current page-one results have. A low score means those pages are weak relative to an average competitor, not that a domain with no links can displace them. For a new site, every keyword is hard.

Why did my new post get impressions and then stop?

Google tested it. New pages often get a short burst of impressions across many queries, then fall out when the data does not justify keeping them there. One of my posts collected 396 impressions across 101 queries in thirteen days, then went to zero for three months while staying indexed.

Can a query in Search Console be fake?

It can be a rank tracking tool rather than a person. The signature is one device, one country, a tiny count on most days, and no spelling variants. Two of my top non-brand queries were daily automated checks from Australia and New Zealand. Real demand arrives in messy variations.

Does page speed on its own get a new site to page one?

No. It breaks ties between pages that already compete. Every post here scores 95 or above on mobile with zero layout shift, and not one of them reached page one. Core Web Vitals are a ranking factor, but a weak one next to links and relevance, and they cannot substitute for either.

How many blog posts do you need before SEO works?

There is no count that works on its own. Thirty-one posts produced zero non-brand clicks here. The same effort spent earning five genuine links would have done more, because links are what lets any of the pages hold a position once Google tests them.