From 0 to 102,472 impressions in seven months, with fewer pages ranking

In January 2026, westchesterride.com did not exist as far as Google was concerned. In July it pulled 102,472 impressions and 697 clicks. It is a local service business in Westchester County, New York, and it is mine, so the numbers come with the month attached and nothing rounded. What makes it worth publishing is not the curve. It is that between April and July the number of URLs receiving impressions went down from 192 to 154 while clicks multiplied by six.

Quick answer

• Seven months from zero to 102,472 impressions a month, with median position moving from 11.6 to 8.5.
• CTR doubled, from 0.31% to 0.68%, because the queries bringing impressions changed.
• Fewer URLs ranking, more clicks: 192 with impressions in April, 154 in July, and 61 to 90 with at least one click.
• Zero link building. The site never bought, asked for, or earned a placement on purpose.

The two series sit in separate panels rather than on a shared axis on purpose. At 102,472 against 697 the clicks line would be indistinguishable from the floor, and a second axis lets you make any two curves appear to move together, which is the trick this blog takes apart elsewhere.

What produced the growth here?

Answering situations instead of topics. The site stopped writing about ground transportation in New York and started resolving exact moments: how to get out of the stadium after a concert, what time to arrive, where to leave luggage. Each page covers one decision somebody makes on a phone, in a hurry, with a deadline.

What kind of case this is

The mental model this contradicts is that a site grows by publishing more. Under that idea the metric you watch is how many pages are indexed and receiving impressions, and both go up while the project is healthy.

Here one of them went down while the business went up. April had 192 URLs receiving impressions and July had 154, a 20% drop, with triple the total impressions and six times the clicks. The correct reading is that impressions concentrated into the pages that resolved something, and the rest stopped appearing because they never had any business in that query.

What the first page of Google says about this

I searched the query before writing, and the pattern is consistent. The first page is agency portfolios listing client wins as percentages: +312% direction requests, tripled leads, top three in sixty days, and one claiming a 1,900% revenue increase.

Percentages without a starting number are unreadable. A 1,900% increase is what you get going from 5 leads to 100, and the 5 is the part that tells you whether the case resembles yours. None of the results on that first page shows a month that went backwards, which is the tell: real seven-month curves have one, and leaving it out is a choice.

The second gap is whose site it is. These are agency pages about client work, so the numbers arrive through a filter with an incentive. This one is my own property, which is a smaller claim and a checkable one.

The starting point, with a date

January 2026. First month with Search Console data: 3,877 impressions, 12 clicks, median position per page 11.6. Of 54 URLs receiving any impression, 9 got at least one click. Six out of seven visible pages were useful to nobody.

That split is the entire diagnosis. A site with 54 pages appearing and 9 working does not have a quantity problem: it has 45 pages written for a query that was never theirs. Separating the two is what free keyword research is actually for.

What was done, in order

1. Read the queries already bringing impressions without clicks. Not the ones we wanted to win: the ones Google was already showing. That is the Performance report, and the mechanics are in how to set up Search Console. Dozens of queries came back with impressions and zero clicks, nearly all the same shape: someone asking about a specific moment and landing on a page about the topic in general.

2. Write one page per situation, not per topic. Instead of a guide to stadium transportation, a page for leaving after the concert, another for arrival timing, another for luggage. It is long-tail applied to moments rather than to keyword variants.

3. Merge and retire what competed with itself. This is the part that produced the URL decline. Several pages covered the same decision under different titles, so one stayed and the rest folded into it. It is the least performed and highest yielding operation in local SEO.

4. No links. None bought, none requested, no guest posts. Not a stance: with a site under a year old the content lever was nowhere near exhausted, and spending there would have solved the wrong problem. What the budget would have gone to instead is laid out in SEO cost for a small business.

When NOT to copy this

When the site already covers its intent well and what is missing is authority. Writing page 155 moves nothing there.

When the niche has no situations. A product searched by its own name does not support thirty pages of distinct moments.

When the competition is brand-driven. If the top ten are sites people search by name, the long tail gets in and the head does not.

When the business lives on the map pack rather than on the blue links, which is a different game covered in how to rank in Google Maps.

When you need revenue this quarter. Seven months are seven months, and money does not compress them.

The result, with its window

July 2026, complete month: 102,472 impressions, 697 clicks, median position per page 8.5. Of 154 URLs with impressions, 90 with at least one click: from one in six to six in ten.

CTR went from 0.31% to 0.68%. That number doubled because the mix changed, not because titles were rewritten: the impressions that remained came from queries the page actually answered. The three pages bringing the most clicks between July 1 and August 11 were leaving the stadium after a concert (4,686 impressions, 134 clicks, position 6.1), what time to arrive (5,844, 92, position 6.2), and where to leave luggage (3,382, 78, position 6.9). 96% of impressions came from the United States.

Mistakes that repeat

• Reading a percentage without its base. A jump from 5 to 100 and one from 5,000 to 100,000 are the same percentage and different businesses.
• Treating indexed pages as a scoreboard. Here the number that went down was the one that told the truth.
• Publishing before reading the queries you already have. The report showing what Google is already testing you for is free and it arrived before any of the writing.
• Buying links to fix a content problem. The two are not interchangeable, and one of them is expensive.
• Extrapolating a partial month. August ran 54,721 impressions in eleven days, which is not a monthly rate and is not reported as one.
• Assuming a local business needs a blog. Some do not, and the profile does the work instead, which is what Google Business Profile covers.

Data and transparency

All figures come from Search Console, pulled through the API on August 13, 2026 with a read-only credential. westchesterride.com is my own property, which is why it appears by name and with exact numbers. Client work in this portfolio gets described without naming and with rounded figures, and that difference is deliberate.

Google documents a logging error affecting this curve. Its Data Anomalies page states that a bug kept impressions from being recorded accurately between May 13, 2025 and April 27, 2026, affecting impressions, CTR and average position. This series crosses that date, so January and April were measured under one regime and July under another. Clicks were not affected, and they are 12 in January and 697 in July: the growth rests on the one column Google says is intact. Checked at the source on August 13, 2026.

The month that got worse: between February and March median position went from 10.3 to 12.6 while impressions rose from 8,700 to 28,191. That is what happens when new pages enter on worse queries. It is included because a case without a single bad month describes no real project.

Positions here are medians per page, not averages. An average moves on its own when new queries arrive, which is exactly what a growing site does.

Not measured: revenue, calls, or bookings attributed to the channel. There is no attribution set up on that site, and quoting a number without it would be inventing one.

What this changes

The useful lesson is not that a site went from 0 to 102,472 impressions. It is that the metric most reports open with pointed the wrong way during the best quarter this project had. Fewer URLs ranking read as a decline on every dashboard and was the direct result of the work that produced the clicks.

That is also why the percentages on the first page of this query are worth distrusting. Not because they are false, but because a number with no base and no window cannot be checked, and cannot be compared against your own situation, which is the only reason to read a case study in the first place. A case that can only be told in percentages is a case that does not survive being told in numbers. What a small business should actually watch instead is in local SEO for small business.

Frequently asked questions