Local Directory SEO Case Study: 58,000 Clicks in Seven Months

Between January and August 2026 we grew a 1,000-page local directory from roughly 150 organic clicks a day to peaks above 440, and booked 57,908 clicks from 2.58 million impressions along the way. This is a local directory SEO case study with the real Search Console numbers, including the parts that did not work the way most SEO advice says they should.

The site is one of our own properties, a national directory of bricks-and-mortar retailers organised by state, then city, then individual business. We are not naming it because the niche is beside the point, but owning it is precisely why we can publish the raw Search Console export instead of the tidied summary a client would let us show. What makes it a useful case is that the site had nothing going for it at the start: no brand recognition, no press, no link-buying budget. Every visitor arrived because one page answered a specific question better than the alternatives. If you run a game store, a nightclub, a dispensary or a roofing company, anything where customers search by location, the mechanics below transfer directly.

The Headline Numbers

The trailing thirty-day average rose 88% across the window. Measured from the early floor near 150 clicks a day to the best days above 440, the chart shows roughly a threefold climb. Both numbers are in the table because they answer different questions: the average tells you what a normal day looks like now, the peak tells you where the ceiling sits when a good week lands.

Lesson One: The Head Term Is a Trap

Every client who asks us about local SEO wants to rank for the one obvious phrase. In this category that phrase was "[category] near me". We do rank for it. Here is what it actually delivered:

Twenty-two thousand impressions on the biggest term in the category produced 290 clicks. A 1.32% click-through rate, because position ten on a "near me" search sits underneath a map pack, a row of ads and nine other results. Those impressions look impressive in a monthly report and they buy almost nothing.

Now compare three individual business listing pages from the same site:

One eighth of the impressions, nearly three times the clicks. A page answering "is this specific business still open and where is it" converts at fifteen to twenty percent because the searcher already knows what they want. The generic term attracts people who are still deciding, and Google hands the top of that page to Maps.

The practical takeaway: stop grading a local campaign on its rank for the category term. Grade it on how many specific pages earn a double-digit click-through rate.

Lesson Two: The Traffic Lives in the Tail

Search Console reported 999 pages earning clicks in this window. The single best page took 887 clicks, which is 1.5% of the total. The top eleven pages combined accounted for about 10%. The other 90% came from hundreds of pages each contributing a few dozen visits a month.

The lower chart is the one worth studying. That is the keyword count climbing from almost nothing to more than 20,000 ranking terms, with the pale band at the bottom showing positions 51-100 and the darker bands above it showing pages working their way up. Growth in a project like this looks like a rising tide of thousands of small terms, not one keyword breaking into the top three.

That distribution is the whole strategy. You do not win a category like this with ten excellent pages. You win it with a genuinely useful page for every city and every business in it, each one indexable, fast and actually distinct from its neighbours. Thin duplicate pages get filtered and drag the rest down with them. Pages carrying real, differentiated information about one location get to compete on terms nobody else has bothered with.

Most businesses get this wrong in the opposite direction. They publish one "our locations" page listing forty branches and wonder why none of them rank. Forty pages, each with its own hours, photos, directions and reason to exist, will outperform that page every time.

Lesson Three: This Is a Phone Product, and the Data Is Not Subtle

Mobile delivered 89% of all clicks. The position column is the interesting part: the same site averages position 8.6 on mobile and 20.7 on desktop. Those are effectively two different search engines with two different result sets, and the gap is not an accident. Local intent skews to phones, Google knows it, and it reorders accordingly.

If your Core Web Vitals look fine on a laptop and mediocre on a mid-range Android, you are optimising the version of your site that nine out of ten visitors will never see.

Lesson Four: There Was No Brand to Lean On

Across 1,002 tracked queries, searches containing the brand name produced zero clicks. Not a small number. Zero.

That matters whenever you read a traffic report, including your own. An established name shows a large slice of traffic from people typing that name into Google, and that traffic was already yours. Every one of these 57,908 clicks came from somebody who had never heard of the site and found it because a page matched what they needed. Harder to produce, and a more honest number to report.

Why the Third-Party Tools Said a Fraction of This

On the same date, a well-known third-party SEO tool estimated this site's organic traffic at 2.7K a month and falling. Search Console says it is doing roughly 10,000 clicks a month and rising.

Both are correct about what they measure. Third-party estimates are models built from a sampled keyword database and a click-curve assumption, and they systematically miss long-tail traffic, which is precisely what this site is made of. Twenty thousand ranking keywords, almost all of them low volume, is the exact profile those tools underestimate.

We report results from Search Console and analytics, never from a third-party estimate, and we would encourage you to hold any agency to the same standard. If someone shows you a growth chart from a tool that has no access to your data, ask to see the Search Console export instead. The gap here ran a factor of three or four in the site's favour, but it can just as easily run the other way.

What Generative AI Changed, and What It Did Not

Search Console now reports impressions inside Google's generative AI features separately. Over roughly ten weeks this site logged 11.4K impressions in AI Overviews and AI Mode, trending up. Third-party tracking separately shows 234 of its pages cited in AI search, 225 of them in ChatGPT.

Two things worth saying plainly, because there is a lot of noise on this topic.

Nothing special was done to earn it. No trick, no schema hack, no "AI optimisation" package. The pages getting cited are the same pages that were already ranking: specific, factual, well structured, answering one clear question. Structured data and clean headings help a machine parse a page and we use both, but they are table stakes rather than a lever.

An AI impression is not a visit. A user who gets their answer inside the overview may never click through. That is a real risk for thin informational content and much less of one for pages like these, because knowing a business exists is not the same as knowing how to get there and when it opens. The businesses most exposed are the ones whose pages exist only to answer a question. The least exposed are the ones whose pages end in an action: directions, a booking, a phone number, a price.

Our position for clients in 2026 is unchanged in substance. Build pages that deserve to be the answer, make them trivially machine-readable, and make sure the reason to click is something a summary cannot hand over on your behalf.

What This Would Look Like for Your Business

The site in this case study is a directory, but none of the mechanics are niche. Any business with locations, service areas, product categories or events is sitting on the same opportunity:

• Build one real page per thing you want to be found for. One per location, one per service area, one per product line. Not a single page listing them all.
• Make each page genuinely different. Real hours, real photos, real local detail. Templated pages with the city name swapped get filtered out.
• Judge the campaign on specific pages, not the category term. A page at position five with a 15% click-through beats a category term at position ten with 1.3%, every time.
• Test on a mid-range phone. That is where nine out of ten visitors are, and where Google is ranking you differently than you think.
• Give every page a reason to be clicked. Directions, availability, a price, a booking. Something a summary cannot deliver for you.
• Measure in Search Console. Not in a tool that is guessing at your numbers from the outside.

This took seven months of consistent work rather than one clever fix. No viral post, no link buying. Structure, coverage, page quality and patience, measured every week against first-party data.

If you want to know what that would look like for your site, start with an SEO audit or read how we approach local SEO. We are based in Las Vegas and we work with businesses that need customers to physically walk through a door.