Real results
We Rebuilt Their Sites.
Here Is What Moved.
Straight out of Google Analytics. From sites we launched this year.
2.4x
the visitors who stayed and read instead of leaving
A local roofer
Half
of everyone who arrived came from a search, not from paid media
A local emergency vet
64%
of people arriving from Google stayed instead of bouncing back to the results
A local window and door company
A local roofer
Two and a Half Times as Many People Actually Read the Page.
36 days after launch, against the 36 days before
2.4x
the visitors who stayed and read instead of leaving
Engaged sessions, 1,339 to 3,260 (+143.5%)
Someone has a ceiling stain and a bad feeling. They are on their phone, they have three roofers open in three tabs, and they are giving each one about five seconds to convince them.
Two and a half times as many people stayed on the page instead of bouncing straight back out
Engaged sessions 1,339 → 3,260 +143.5%
Roughly 1 in 3 visitors stuck around, where it used to be 1 in 5
Engagement rate 20.78% → 29.31% +41.1%
A five second glance turned into an actual read
Average engagement time per session 5 sec → 10 sec +98.7%
About 4,700 more people landed on the site over the same stretch
Sessions 6,445 → 11,121 +72.6%
More of them arrived from a plain Google search, with nothing paid behind it
Organic search sessions 367 → 448 +22.1%
People clicked, tapped and scrolled far more once they were on it
Event count 21,177 → 41,492 +95.9%
Same company, same crew, same trucks, same spend. The only thing that changed was the website, and the homeowner with the leak stopped bouncing off it.
People asking an AI assistant for a roofer started landing on them too: 12 visits to 17 over the same window. Small numbers, and we are not dressing them up as a percentage. The point is that the channel exists at all now.
A local emergency vet
More Than Half of Them Found It Without an Ad.
First six weeks after launch
Half
of everyone who arrived came from a search, not from paid media
1,287 of 2,386 sessions were organic (53.94%)
It is eleven at night and somebody's dog is in trouble. They are frightened, they are typing one-handed, and they are going to call whoever answers the question fastest.
More than half of everyone who showed up found them through a search, with no ad spend behind it
Organic share of all traffic 53.94%
Half of those searchers stayed and read instead of going back to the results
Organic engagement rate 50.12%
645 frightened pet owners got far enough into the page to find what they needed
Engaged organic sessions 645
Six weeks from launch, with no paid media, this is what the site pulled on its own
Total sessions 2,386
For an emergency clinic the search result IS the front door. Half the people coming through it stayed long enough to read, which is the whole job at eleven at night.
A local window and door company
Two Out of Three Searchers Stayed and Read.
August 1 to September 13, 2026
64%
of people arriving from Google stayed instead of bouncing back to the results
67 of 104 organic sessions engaged (64.42%)
It is a considered purchase. Nobody buys eleven windows off a five second visit. They are comparing, they are reading, and the company that explains itself properly is the one that gets the call.
On a big-ticket job the win is not the click, it is the minute after it. A minute on the page is somebody working out whether to trust you.
Nearly two in three people from a search stayed on the site instead of clicking back and trying a competitor
Organic engagement rate 64.42%
Close to a full minute per visit, which is reading, not skimming
Average engagement time per session, organic 59 sec
A local pest control company
The AI Assistants Are Already Sending People.
After launch
70%
of the people arriving from AI assistants stayed and read
AI assistant visitor engagement
Somebody heard scratching in the wall last night. In 2026 a fair number of them do not open Google at all. They ask ChatGPT who to call.
Not yet confirmed against a capture
Seven in ten people who arrived from an AI assistant stayed on the page
AI visitor engagement 70%
They stayed over a minute, so the tool sent people who actually wanted the service
Time on site, AI visitors 60+ sec
Every site in this set now shows AI assistants as their own channel in Analytics. That traffic is arriving today, and it only arrives if the site is built so those tools can read it.
The Proof Is in the Pudding
Our Own Numbers
We don't just build these sites. We run our own business on one. So we tested it on ourselves.
We killed 100% of our Facebook ads and 95% of our Google ads. All of that paid traffic, gone. Traffic dropped, exactly like you'd expect when you pull the spend.
Then it came back. Within a few months, Google and AI search were bringing in enough on their own to replace the paid ads. Same traffic, none of the cost. We're saving thousands and thousands of dollars a month.
Our real Google Analytics data is below. After all, the proof is in the pudding.
That's what it means for a site to earn its own traffic.
Sessions, down 4.63%
9,672 before, 9,224 after. That is what pulling 100% of the Facebook ads and 95% of the Google ads cost us.
Organic Search sessions
Before241
After910
Up +277.59%
Organic Search engaged sessions
Before141
After568
Up +302.84%
Engagement rate
Before11.71%
After26.71%
Up +128.04%
Engaged sessions
Before1,133
After2,464
Up +117.48%
Avg engagement time per session
Before8s
After12s
Up +47.21%
Events per session
Before3.95
After4.84
Up +22.7%
Google Analytics, smallbusiness-seo.com. Jun 15 - Aug 13, 2026 next to Apr 16 - Jun 14, 2026.
That's the power of the 100K site.
The Pattern
Across every one of these launches the strongest signal is not more traffic on its own. Traffic is easy to buy and easy to fake.
It is what people did after they arrived. More engaged sessions. Higher engagement rates. Longer visits. Organic search climbing with no spend behind it. And a channel that did not exist a couple of years ago sending real people.
Same company, same crew, same market. The website was the variable.
Straight Answers
- Why are the businesses not named?
- These are real clients and their Analytics is their business, not ours. You get the industry and the numbers, which is the part that tells you anything. Every figure comes from Google Analytics on a site we built.
- How long does it take to see a change?
- The roofing results above cover 36 days after launch against the 36 days before. The emergency vet numbers are the first six weeks after launch. Weeks, not years.
- Is more traffic the point?
- No. Engaged sessions rose 143.5% for the roofer while sessions rose 72.6%, so the bigger move was in what people did after they arrived. Traffic you can buy. Engagement you have to earn with the site itself.
- Are AI assistants really sending visitors yet?
- Yes, and it shows up as its own channel in Google Analytics. One site went from 12 to 17 visits from AI assistants over 36 days. Small numbers today, but the channel is live now, not coming.
Want to See Yours Before You Pay?
We build it first. You drive it, poke at it, and decide after you have seen the thing working. Not a mockup. The real site.
