EDGE of the Web

The episode delivers a handful of genuinely tactical ideas - feeding GSC impressions-with-no-clicks into an LLM for content gap analysis, creating a GA4 channel via regex for AI referral traffic, and a methodology for calculating AI traffic ROI in dollar terms. However, the signal is diluted by long host monologues, two sponsor reads, and several passages of standard SEO commentary that add no new ideas.

“I'm going to go into Google Search Console and I'm going to look at the queries for this particular page that had impressions but no clicks. So as in they were probably pretty far down in the, you know, in the lower positions. And then I'm going to feed that. Usually I'll take those queries and feed it into an LLM”

“you can add a channel in GA4 and you can also take this regex and apply it to any other tracking tool at all. Fathom, Matomo, doesn't matter. And create a channel for AI tools. So people coming to your website from ChatGPT, Claude, Perplexity, et cetera. It takes five minutes to do”

Most frameworks are standard SEO/content strategy fare recycled competently, but there are two sharper framings: the 'selling against inertia' concept (credited to April Dunford, not original to the guest) and the 'ChatGPT is Google in a trench coat' point about local AI queries being essentially map-pack lookups. Neither is a first-principles breakthrough, but both are more concrete than typical podcast takes.

“ChatGPT is just Google in a trench coat and it's just going out and getting whatever happens to be in the map pack and then presenting it back to you”

“you're not selling necessarily against your competitors, you're also selling against. What if we did nothing?”

Dana DiTomaso is a genuine practitioner - agency president, long-time conference speaker at MozCon and SMX, course creator with real client data - who speaks from actual implementation experience rather than thought-leadership abstraction. She is solidly credible but operates at the agency-SMB level, not at a scale that would produce truly exceptional data or stakes.

“I think I first talked about this@mozcon 2014 and a talk called Prove youe Value. It's on YouTube if you want to watch it. And, uh, yeah, we're still bad at it, so let's get better at it”

“I just did, um, a consent review for a client and I was looking at their consent setup, which was incorrect, which is, I would say, about half the time. That's what I'm dealing with now is fixing up bad consent configurations”

The episode earns decent marks for naming real tools (SparkToro, Gumshoe, Fathom, Matomo), real people (Celeste Gonzalez of Rickety Roo, Brittany Muller, April Dunford), and producing an actual dollar figure and conversion rate delta from her own business. The weakness is that most client data is anonymized or described at a very high level, and the headline $84k figure comes from a small course business rather than a scaled operation.

“a one and a half percentage higher conversion rate from AI tools versus just general organic search. And with that, based on our average order value, that results in an extra $84,000 of revenue from AI in a course of a year”

“Celeste Gonzalez of Rickety Roo just had a great post on how to do this. She has a tool, I think, built in Colab, maybe that will go out and grab reviews”

The host frequently delivers multi-sentence monologues before arriving at a vague question ('give me a lay of the land'), rarely follows up on specific claims, and never challenges the guest. The best moment is the open-ended 'What am I not asking right now?' prompt, but the overall dynamic is a promotional conversation rather than a rigorous interview.

“So. So, uh, give me a lay of the land here. Where should we go in looking at our own content now?”

“What am I not asking right now that we should pack into this episode?”