Working notes on building SyncSpark: workflow design, agent orchestration, and what I'm noticing as the small-business operating model gets rebuilt around AI.
Review volume is not a vanity metric. It is the signal Google uses to decide whether your rating is real.
Most small business owners treat their Google rating like a report card. Once it hits 4.8 or 5 stars, they stop thinking about it. But the rating is only half the story. The other half is how many reviews produced it.
Read note →The generalist trap is not about capability. It is about what you give up when you say yes to everything.
When I started SyncSpark, I took whatever came in. Websites, SEO, social media, ads, email strategy. The logic was sound: more services meant more clients. The reality looked different on the inside.
Read note →Google and AI engines do not read your site the way your customers do. Here is what they are actually looking for.
Every small business owner I talk to has thought about their website from the customer's perspective. Almost none have thought about it from Google's. The gap between those two perspectives is where most search visibility is lost.
Read note →Everyone is asking if FAQ schema is dead. They are asking the wrong question.
On May 7, 2026, Google stopped showing FAQ rich results for most websites. Within a week, every SEO forum was asking the same question: is FAQ schema dead? The answer is no. But the reason matters more than the answer.
Read note →Most businesses run ads before the foundation exists. This is why that fails.
The standard advice is: if growth stalls, run ads. The problem is that ads do not fix a broken foundation. They amplify it. A site that converts at one percent converts at one percent with paid traffic too, just at a much higher cost.
Read note →Good work and search visibility are disconnected systems. Here is what that gap actually looks like.
He has been in business for over a decade. Twenty Google reviews, every one of them five stars. Ask ChatGPT who the best electricians in Vancouver are. His name does not come up. This is not unusual. It is the default.
Read note →Everyone treats Lovable and Bolt as competition. The data says otherwise.
When a founder builds their own site with an AI tool, the conventional read is that I lost a project. My actual experience: those same founders show up three to six months later with a specific, urgent problem. The discovery call is easy. The problem is obvious. The path forward is clear.
Read note →