The Smart Marketer Podcast

There are a handful of genuinely useful data points - North Beam aggregate stats, portfolio-level Q1/Q2 comparisons, audience network quality issues, and Meta's LTV optimization shift - but they are surrounded by substantial filler including motivational tangents, AI boosterism, and mindset coaching that adds no operator value. The insight-to-minute ratio is low.

“conversion rate went from an average 1.4% in January 2025 to.7. So it halved”

“Meta has been for some reason pushing quite a lot of traffic towards the audience network...your cost per click might be really low. Um, but it's generally, it's generally junk traffic”

The overwhelming majority of takes are standard industry consensus - 'focus on fundamentals,' 'diversify channels,' 'trust the algorithm or don't,' and 'mindset matters.' The AI-enabled vs. AI-native framing and the Meta LTV optimization observation are modestly fresh but are barely developed before the conversation pivots to motivational content.

“we are now going from an AI enabled place to or AI enabled businesses to becoming AI native businesses where AI is actually starting to do a lot of the work”

“having the right mindset about this is absolutely crucial”

Pep Houfen is a genuine practitioner running a real agency with a real client portfolio, which produces the episode's most credible data points. However, he is an internal business partner of the host at the same company, which limits independence and prevents any outside perspective; much of his airtime is crowded out by the host's extended monologues.

“we work with so many clients, there are always brands that are scaling, there are always brands that are Going to tough spots. So it was fun to basically look at all of the brands and what are we seeing overall”

“top line revenue is down in Q2 now with about 10%”

The episode's best moments are genuinely specific - North Beam's named report with exact conversion rate and CPC figures, plus the agency's own portfolio aggregates - but most other claims remain vague ('some clients,' 'a nine figure brand owner,' unnamed accounts), and no client names, campaign structures, or dollar outcomes are ever disclosed.

“cost per click went from around $2.70 in January 2025 to $1.33 around now. Right. So, like May 2026”

“North Beam published, uh, basically a brief report, and they compared conversion rates and cost per click to last year”

The host routinely answers her own questions, cuts into the guest's responses with multi-paragraph monologues, and never challenges a single claim Pep makes; the result is a two-person lecture with interjections rather than a rigorous conversation. Questions are largely leading or rhetorical, and the extended mindset and AI sections read as brand positioning rather than inquiry.

“So, Pep, I know you've put together, uh, sort of a pulse of meta advertising so far in 2026. I'm, uh, excited to dive in and see what you found”

“What do you think this is? I have my own hypothesis...It's a really good question to think about what the true answer is. The true answer is from a conversion rate. I don't really know”