A candid conversation with Phil Wiseman of Analytics at Profit on real SEO, bad AI habits, and why your metrics dashboard is probably lying to you.
Phil Wiseman didn't set out to become an SEO expert. He was running a calibration business, frustrated that his marketing agency kept handing him vanity metrics impressions, reach, likes when all he wanted to know was: how many RFQs did I get? That frustration sparked a journey that eventually led him to build Analytics at Profit, a consultancy focused on cutting through noise and delivering the numbers that actually move a business forward.
We sat down with Phil for a no-holds-barred conversation about the state of SEO, the AI content crisis, and what it really takes to show up and get clicked in today's search landscape.
Phil's origin story is more common than you'd think. Twelve years ago, he hired an agency, got a shiny report full of charts, and realized he had no idea what any of it meant or how it connected to actual revenue. Sound familiar?
Today, he says the problem hasn't gone away. Agencies still lead with reach and impressions. Clients still get excited. And the fundamental question "did this bring in customers? " still goes unanswered.
Phil's remedy is ruthless simplicity. His dashboards, built through Databox, are stripped down to what matters:
"I don't like Google Studio," he admits bluntly. "It tells people things they don't want to hear in a way they can't use. I want to know: did the plumber get an appointment? Did the woodworker get an order? That's the dashboard I build."
Phil doesn't mince words about what's happening to the internet right now: "People are producing really shoddy content with AI, ruining their websites, and then coming to me to fix it."
His core objection isn't with AI itself he uses it constantly. It's with lazy AI: tossing a vague prompt at ChatGPT and publishing whatever comes out. The result? Websites full of content that ranks for nothing, answers no real question, and sounds like it was written by a very confident robot who has never spoken to a customer.
Phil's content strategy sounds almost embarrassingly simple until you realize almost nobody actually does it. He calls it "ask and answer."
He talks to the service desk. He asks: what weird questions did customers call about this week? One week it was toilets making a hissing sound during a water main break. He wrote a blog about exactly that. It ranked. It got clicks. It solved a real problem.
Phil's analytical process is where his background in data really shines. He pulls 90 days of Google Search Console and Analytics data, drops it into Perplexity, and looks for two specific patterns:
You showed up but your title or meta description isn't compelling. Time to rewrite the entry point the page title and meta description need work before anything else.
You've found a niche that resonates. Expand it. Research other ways people search this topic and write more content around it.
His sweet spot? 500 to 1,000 impressions with a 5%+ click-through rate. Not viral, not flashy just consistent, quality traffic from people who actually wanted what you wrote.
Here's where Phil pushes back on conventional wisdom. Everyone wants to "rank in AI overviews." Phil says that's the wrong goal.
"Showing up in AI overview is brand awareness. What you want are clicks." He points to a client who does business valuations this client gets roughly 20 click-throughs per month from ChatGPT and Perplexity. That's not because he gamed anything. It's because he writes deeply expert content that AI tools actually trust as a source.
Phil also notes a simple structural trick: content with bulleted lists and FAQ sections tends to get picked up by AI overviews faster. Not because of any hack but because AI systems, like human readers, appreciate content that's easy to scan and extract value from.
Not inherently but lazy AI content is. Phil uses AI extensively in his workflow, but always with detailed prompts, human oversight, and quality checks. The problem isn't the tool, it's skipping the research and brand voice work that makes content genuinely useful.
Phil believes it's the opposite, traditional SEO fundamentals (great content, proper structure, schema, clean site) are precisely what gets you into AI overviews. You can't skip the foundation and expect AI systems to trust your site.
ChatGPT is designed to always provide an answer even if that means hallucinating one. Perplexity cites its sources, which makes it far more reliable for research and training a custom brand voice. Phil has also found it better at staying accurate when given updated information to work from.
It may be a structure problem, not a quality problem. AI systems tend to favor content that's formatted for easy extraction think bullet lists, clear heading hierarchy, and FAQ sections. If your top-ranking content is dense prose without those elements, try adding a structured FAQ section at the bottom and see what changes.
Schema is structured data you add to a webpage to help search engines understand what your content is about. Phil calls it a critical and widely overlooked part of SEO. He writes it with AI assistance, but always runs it through a schema validator to catch errors before publishing. Even geo-specific schema, he says, makes a measurable difference in local rankings.
Google's own autocomplete. Start typing any query and watch the suggestions. Those are real searches, ranked by popularity, updated every few weeks. Phil teaches this at local library classes because it costs nothing and reflects exactly how humans are searching right now.
Phil is the founder of Analytics at Profit, a consultancy helping small businesses particularly in skilled trades understand their data and build SEO strategies grounded in real human behaviour. He also teaches digital marketing at local libraries and holds what he claims is the record for most LinkedIn account restrictions for telling the truth about shady marketing practices.