Case Study: Finding the Gap in a Market Worth 225,500 Searches a Month

Lake Louise gets about 225,500 searches a month across Canada and the United States for the place name alone. It is one of the most photographed locations in the country. A national tourism board, a national park, a ski resort, a Fairmont hotel, and a dozen travel blogs all cover it.

In theory there is no room left. There is a large gap, and data found it. Opinion did not.

This is how we found it for lakelouise.net, an independent visitor guide we are building.

What we measured

We pulled and scored 2,423 keywords containing Lake Louise, Moraine Lake, or a named trail in the area, across Canada and the United States. We then captured 18 live search result pages in full. That shows exactly who ranks, in what position, with what kind of page.

We also pulled the link profile of every competitor that appeared, because ranking is not only about the page. It is about whether the site behind it is reachable at all.

Finding 1: the main term is not winnable, and chasing it wastes a year

The search for "lake louise" returns a knowledge panel, a photo strip, a top-sights carousel, and an AI overview. The first independent guide site sits at position 19.

Its difficulty score is 51. On a new domain, that is a multi-year project with no revenue attached.

So we stopped trying to win the term and looked at what people search after they have already decided to go.

Finding 2: the incumbents are strong on authority and weak on detail

Banff and Lake Louise Tourism appears in the top ten of 22 of our captured results, backed by 6,064 referring domains. It ranks on authority.

Then we measured the pages themselves. Its Lake Agnes page is 438 words. Its canoe page is 118 words.

A page that genuinely answers the question outranks a 118-word page, even from a far weaker site. The proof was already in the data: travelbanffcanada.com ranks first for "lake louise hikes" with 432 referring domains. That is a beatable number.

Finding 3: one competitor proved the model works

thebanffblog.com earns an estimated 15,700 visits a month from Lake Louise keywords alone, across 734 ranking keywords.

That is the target shape of the business. It also told us where the opening is, because that site has no live data on it at all.

The gap: nothing that changes today has an owner

This was the finding that decided the whole project.

We looked at the search results for road conditions, webcams, and current status. Every result is a raw government tool: Alberta 511, AMA road reports, DriveBC, and Environment Canada. There is no guide site anywhere in the top ten. No People Also Ask box. No AI overview.

Nobody explains what the data means for the visitor's morning. They just publish the raw feed.

The volumes are real and the competition is soft:

• "lake louise weather", 27,100 searches a month in Canada, difficulty 3
• "lake louise parking", 8,100 a month, difficulty 10
• "lake louise paid parking", 5,400 a month, difficulty 3
• "lake louise webcam", 5,400 a month, difficulty 14, spread across five different owners with no single page collecting them

For comparison, the best independent parking page we found is 7,357 words and ranks 14th. Parks Canada holds the top two positions with a page that never says what time the car park fills.

A shorter page that leads with today's status and answers the twelve real questions in clear headings can take positions three to six. That is not a theory. It is what the gap in the data describes.

One more free advantage. Of 49 competitor pages we fetched, only two carry proper FAQ structured data. Nobody in this market has taken it.

What we built for it

Live conditions first, articles second. The site pulls weather from Environment Canada, roads and closures and cameras from Alberta 511, and avalanche danger from Avalanche Canada. A scheduled job collects each feed every ten minutes and stores it. The pages then read from our own store. They never call a rate-limited government API on a busy morning.

Two things we learned building it are the reason a competitor has not done this properly.

A dead camera returns a normal response. Fifteen of the 25 camera views near Lake Louise serve a "no live camera feed" placeholder. One served a month-old picture while its headers claimed the image was seconds old. We now fingerprint every frame, because trusting the response would mean publishing a lie about the road you are about to drive.

Not every closure near Lake Louise is on the way to Lake Louise. Alberta 511 covers the whole province. Winter road records carry no coordinates at all. Each event is graded on whether a visitor would actually drive that road.

Every tile on the site shows its own age. Stale data stays on screen with the age attached, and a dead source says so. Neither ever renders as an empty box, because an empty box is how a visitor decides a site is broken.

How we will know if it worked

We are not claiming a result yet. The site is in build and the guides are still being written.

The measurement rule was set before launch: nothing is sold on the site until analytics shows twelve consecutive weeks of growth. Advertising inventory sold against a hopeful projection is how these projects fail.

What this means for your business

The method transfers to any market, including a much smaller one.

Do not fight the head term. Measure what people search immediately before they buy. Then read the actual pages that rank, not just the site names, because authority hides a lot of thin content. The gap is almost always a question everybody in the market answers badly.

For Lake Louise, that question was "what is it like there right now". For your business it will be something equally specific, and your competitors are probably answering it in a 100-word paragraph or not at all.

I run this analysis for clients before a word of the site gets written. Email jon@delcaromedia.com.

Sources: DataForSEO live search results and keyword data, Canada and United States, English, collected 10 August 2026. Search volumes are grouped by Google across near-identical phrasings, so treat them as weights rather than as a traffic forecast.