Every Friday I write down what actually happened at North Digital that week. What we built, what worked, what I got wrong, and what I'm thinking about. Clients stay anonymous. This one covers Friday October 30 through Friday November 6.
What we shipped
The operational win this week was landing page readiness. Designers finished pages for several dental locations we've been migrating, and the team lined up a DNS switch for the ones that are ready. Same theme as last week: get the pages onto faster hosting, then point traffic at them when they're actually done. No drama. Just the boring infrastructure work that makes every campaign easier to run.
We also kept building out text messaging reporting for a dental account that already has replies flowing. The ads and the inbox only matter together if the practice can see what happened after the lead landed. That report is still being shaped with the client, but the direction is clear: show the reply trail, not just the click.
On the content side, we published two blog posts. One was about a study of almost 1,900 online shops that found paid ChatGPT ad visits earn far less than Google search ad visits. My take is simple. People go to Google when they want to get something done. ChatGPT is still mostly conversational. There's a place for ChatGPT ads, but they're not replacing Google Ads for service businesses any time soon.
The other post looked at a European property operator using AI to cut costs across tenant service and admin. The lesson I care about is not the 99% cost cut headline. It's that when a business uses AI to improve margin and customer lifetime value, the budget that makes sense for paid ads goes up. Better unit economics buy you more lead volume.
Wins
In our weekly client check in, a few dental accounts came in with cost per lead numbers I'm happy to put in writing. One generated 9 leads on about $475 spent, roughly $52 a lead. Another landed at about $51 a lead. A third sat closer to $99. Several other accounts held steady through the end of October. None of that is magic. It's tight structure, patient exclusions, and someone reading the numbers every week.
Our own marketing kept booking free review calls from content. Multiple Calendly bookings came in tagged from content marketing this week. The funnel is still early, but prospects are showing up from the writing, not only from ads.
I also took a few prospect calls, including one with a company that sells marketing and payment tools to restaurants. We talked through a lean Google Ads plus landing page plan, and whether the fee could tie to converted merchants. Those conversations are still open. The useful part for us is testing whether performance style pricing fits certain verticals without turning the agency into a casino.
What I learned
1. Ship the page before you argue about the ad. We spent years optimizing campaigns pointed at slow or half finished pages. The DNS cutover only works when the page is actually ready. This week the work was design finish and switch planning, not another keyword pass.
2. Intent still beats novelty. The ChatGPT ads numbers matched what I see with clients. New ad surfaces are interesting. High intent search is still how most plumbers, dentists and clinics buy work.
3. Margin work is ad budget work. If AI (or anything else) makes each customer more profitable, you can responsibly spend more to acquire the next one. Owners who only cut costs and never reinvest the savings into demand usually stall.
4. A non marketing hot take can outperform a marketing tip. More on that below.
What we posted
On LinkedIn, I posted a hot take that Canadian telecom stocks are in for a rough decade. I use a foreign eSIM with unlimited global data for about $99 USD, which is cheaper than many Canadian plans even when I'm home, and Starlink keeps raising the pressure. The post drew a big comment thread. People argued about roaming caps, exchange rates, CRTC protection and whether the oligopoly can keep prices sticky. It was not a Google Ads tip. It was a clear opinion. That seems to be what travels.
I did not post on X this week. After a stretch of daily posts the week before, the feed went quiet while I was traveling and catching up. Better an empty week than filler.
On the blog, the two posts above went live: ChatGPT ads versus Google search ads, and the property operator AI story tied back to ad budgets.
Next week
The DNS switch for the finished landing pages should go through. We'll keep tightening the messaging report with that dental client, follow up on the restaurant tools prospect plan, and keep taking the free reviews that content is booking. A few more accounts are still waiting on rebuilds once billing details land.
If your landing pages and your ads still live in different worlds, reach out. Happy to take a look.
