Chad Treadway: So we built a tool to grade pest control websites. A five-location company, twenty-five years in business, over fifteen hundred reviews, scored a hundred thirty-six out of six hundred. A company with fifty-eight reviews whose site was built by an AI tool last year scored four hundred fourteen. After we checked both by hand, they were basically the same site. Around four hundred thirty each. So what happened?
Adam Bennett: Welcome to Pest Control Marketing That Actually Works, the podcast for pest control operators who want real growth, not empty promises. I'm Adam Bennett.
Elisabeth Pallante: And I'm Elisabeth Pallante. We're from Cube Creative Design, and for over twenty years we've helped pest control companies stop wasting money and start growing.
Adam Bennett: So today we're doing something we've never done on the show. Chad Treadway, our CMO, audited seven real pest control websites using the Cube Score, our own grading tool. And the results were not what any of us expected. This episode is a little longer than usual because the data has really earned it. So here are your three key takeaways, and they're questions, because honestly we're still sorting through the answers.
Elisabeth Pallante: First, does a website score measure your business or does it measure what a machine can read? Two very different things. Second, can an AI-built website really outscore a twenty-five-year company? And if so, what does that mean? Third, before you act on any score, what should you actually check first?
Adam Bennett: Chad, this one's yours. What did you find?
Chad Treadway: Okay, so I audited about three hundred and fifty websites at this point, all related to pest control. Most of that was for the PCT School presentation I did earlier this year. And then we built the Cube Score and I wanted to stress test some of the sites I'd looked at against ones I already knew.
Adam Bennett: Wait a minute. So you audited three hundred and fifty websites?
Chad Treadway: Yes, just for that presentation.
Adam Bennett: Wow. So are you gonna talk about all three hundred and fifty today?
Chad Treadway: No, look, you're fortunate enough you do not have to hear me go on for three hundred and fifty websites.
Adam Bennett: Okay. Whew. Okay.
Chad Treadway: So I picked seven today just to run through the Cube Score and see how they were doing. Five of the seven landed right about where I would expect, in the mid three hundreds, all within twenty points of each other. No real surprises there. And then two of them were way, way off. And I'm like, okay, what's going on here?
Adam Bennett: Wait. A hundred and thirty-six on a company with fifteen hundred reviews. That's an F. That's a failing score. How does a company with that track record score an F?
Chad Treadway: Well, that's because the crawler we use never actually got in the yard. Think of it this way. Their host runs bot protection. They're on SiteGround. It's one of the biggest managed WordPress hosts out there. Our tool pulled it up. It handed the host what it needed to prove it was a human. Then it came back with 169 bytes. Very, very small for a website. A real page is like tens of thousands. So essentially this was not a website. This was a door hanging tag.
Elisabeth Pallante: So when you say that, basically the site looked empty to the crawler, which means it looks empty to an AI search as well, right?
Chad Treadway: It's actually worse than that. The server sent back what's called a 202 status code. That's a success code. So basically it means, hey, I got your request, here's the page. Our tool thought it had the page. It took that door tag and scored it like it was a full website. It never failed. It never had a clue.
Adam Bennett: And it produced a full report.
Chad Treadway: Yes. Went through and checked every box, all six sides of the Cube Score. Wrote the advice on every section. Everything looked normal. But then we realized some of it was made up.
Adam Bennett: Yeah, and that's our tool. We're not calling out somebody else's product. This is a problem that we found in something that we built.
Chad Treadway: And I want to be clear, that company sitting on the other end of that 136, this is a well-run operation. Five locations across South Carolina. They publish pricing on four tiers, itemized. They tell you what it's based on. Their top plan is like $129 a month for a 2,500-square-foot home. That's the most complete pricing in my entire sample.
Elisabeth Pallante: Yeah, that's actually way more transparent than any of the sites that scored higher, right?
Chad Treadway: Yes, by a lot. They've got a written satisfaction guarantee with real terms behind it too. You have issues between quarterly services, they come back out, no charge. They're a chamber member. They've got a pest library, a blog, city pages, team page, careers. Probably over a hundred pages. But our tool never saw any of them.
Adam Bennett: So if someone ran the Cube Score on that company right now, they'd get a hundred and thirty-six and think the site was garbage.
Chad Treadway: Right. There was nothing in the report that would tell them anything went wrong. That's the part that got me. The score didn't say we couldn't reach this site. It said here's your grade. And the grade was fake.
Adam Bennett: Okay, so the 136 was a crawl failure. What about the other end? How does a small company with fifty-eight reviews outscore everyone else in your sample?
Chad Treadway: Because the Cube Score measures structure, and an AI website builder puts out perfect structure on demand. You're going to get clean markup, 30 out of 30. Heading hierarchy, 20 out of 20. Thirty-nine geo pages covering forty towns. Pricing page right there in the main nav. Your structure score is going to come back with a ninety-six out of a hundred, which is a solid A.
Elisabeth Pallante: So the tool rewarded the things that a machine is already good at building.
Chad Treadway: And those things do matter. Schema helps search engines understand you. Heading hierarchy helps the screen readers and the AI crawlers read your content. Geo pages help you rank in the towns you actually service. None of that is fake. The structure is real.
Adam Bennett: But?
Chad Treadway: But the site is running five satellite domains. One company, fifty-eight reviews, covering forty towns across Iowa for what looks like a single operation. Five different domains all pointing back to the same business. Testimonials on the homepage are first name and town. No photos, no last names, no links. I can't verify anything. And the page metadata still points back to a cloud storage bucket from whatever AI builder spit it out.
Elisabeth Pallante: So that's pretty aggressive for a company that size. Thirty-nine geographic landing pages for forty towns is kind of exactly what you told people not to do at PCT School, right? You showed a slide called the doorway page trap.
Chad Treadway: Yep. I said in the presentation, don't sit down one weekend and crank out fifty pages and then throw them all live. Don't use a template where you swap the city name and everything else stays the same. An AI builder can do this in an afternoon. So the question isn't whether the page exists. The question is whether there's anything of value in them.
Adam Bennett: So Chad, if the structure is real and the rankings follow, does it matter that a machine built it? We tell operators to build geographic landing pages all the time. We tell them to add schema. Put pricing in the navigation. The site did all of that, right?
Chad Treadway: Right. We tell them to build geo pages with real content. Content that proves you know the area and that you're actively in it. I showed a slide at PCT School, a page that named local landmarks, local pest pressures, actual neighborhoods. That's what's going to earn you a page on the search results. Now if all thirty-nine pages say the same thing with the city swapped out, that's a doorway page. It's not a true geographic landing page. And Google has been real clear about what they want to do with those.
Elisabeth Pallante: They get rid of them, right? But the score can't tell the difference between real geographic content and templated geographic content.
Chad Treadway: Right. Not yet. Here's the honest answer. The tool counts your pages. It doesn't read them. So thirty-nine thin pages score the same as thirty-nine good ones.
Adam Bennett: Well, let's also mention something about this company because I think it's really important. This is a real small business. Veteran-owned. Legitimate five-star reviews. They're serving customers. The site is the part that's debatable. It's not the business.
Chad Treadway: Right. What my gut tells me is the owner got sold on an AI builder, or heard about it on a podcast or YouTube video, and is trying to punch above their weight online. I get it. Small business, you're gonna try everything you can. You've got fifty-eight reviews and somebody tells you you can look like the biggest company in the state. That's a tempting offer. But there's a difference between building real authority in your market and building something that only looks like authority to a machine.
Adam Bennett: So we have a strong site that scored low because the crawler bounced, and a machine-built site that scored high because the crawler got in and liked what it saw. What's the actual lesson for a person driving between jobs right now?
Chad Treadway: The lesson is really about the gate, not the score. The company's robots.txt says hey, come on in. No disallow rules, nothing blocked, not one crawler. You pull up their robots.txt and you'd swear everybody in the world can reach them. Meanwhile, their host is handing out a captcha and a noindex header to every crawler that shows up.
Adam Bennett: Okay, so that's really technical. Can you explain it a little easier?
Chad Treadway: Yeah. Think about this. Your tech is pulling up to a house. Customer called and said, hey, the back gate's unlocked, no padlock, nothing. But there's a big huge Rottweiler, or a little chihuahua and your tech is scared of it. That guy's not going to go in the yard. And I'm not going to blame him, honestly. So he treats the front, gets back in his truck, and closes the ticket out as a full perimeter service. The gate was open the whole time. That didn't matter. Your robots.txt is the customer saying, hey, your gate's open. The host is the dog. A real person gets through because they can solve the captcha. A crawler can't. The crawler just leaves.
Elisabeth Pallante: The Cube Score has a check, an actual checkbox, called AI Crawlers Allowed. It checks whether your robots.txt file blocks crawlers like GPTBot, ClaudeBot, and PerplexityBot. This site passed that check five out of five. While the host was actively blocking every single one of those crawlers.
Chad Treadway: Right. Because the check reads what the customer told us. It never walked into the yard.
Adam Bennett: So this just generalizes. If you're on SiteGround or WP Engine or any managed host with bot protection, that could be you.
Chad Treadway: Right. Your robots.txt file is going to look clean. Your site's going to pull up fine for anybody unless they're running a crawler. But when ChatGPT or Gemini or Perplexity or Grok sends their crawler out to decide whether you're worth recommending, you might hit that same dog. And you'd never know it because there's not a callback for that one.
Elisabeth Pallante: So how do you test this?
Chad Treadway: Two ways. The quickest one is go to ChatGPT or Perplexity, start up a clean chat or a private chat. Ask it to recommend a pest control company in your town and see if you show up. If you don't, you need to figure out is your site strong enough that you should have been recommended, or is there something else going on like your hosting blocking crawlers as in our example? The second way is to email your host and ask, can GPTBot reach my site? What about ClaudeBot, PerplexityBot, GrokBot? Ask about their bot protection settings while you've got them. Most support teams either know the answer or they'll look it up and send you an email back.
Adam Bennett: I want to be careful about something here. Google's own crawler does get through on whitelisted IP ranges. So this company still shows up in regular Google search. Their Google Business Profile still works. The issue is specifically with the AI crawlers that decide who to recommend in conversational answers.
Chad Treadway: Right. Their robots.txt and their host don't agree with each other. The customer's saying the gate's open. The dog says otherwise. You're visible in regular search, but you're invisible in AI search. And that gap is going to get more expensive every month because people are asking AI for recommendations instead of scrolling a list of ten blue links.
Elisabeth Pallante: And that's really the whole argument of this show, right? Owned content and AI visibility are the compounding assets. If your host is blocking the crawlers that feed those AI answers, you're building an asset that nobody can see.
Adam Bennett: So Chad, the other five sites in your sample all clustered in the mid three hundreds after correction. What jumped out?
Chad Treadway: Four stories, four different companies, four different gaps. And every one of them shows you something the number itself won't. The first one is the easiest fix in the bunch. Small-town North Carolina company, been in business over seventy years. Number one in the local pack. Number one organically. This company's winning.
Elisabeth Pallante: So what's the problem?
Chad Treadway: Two different phone numbers on the site and two different business names. One of those numbers is an 828 area code, which is Western North Carolina. That business sits 200 miles east of there. The number is a copy and paste from another web build. It's not a typo. Somebody built the site, pulled the content in from a different project, and never cleaned it up.
Adam Bennett: So they rank number one and they have a wrong phone number on their site.
Chad Treadway: Yep. They rank number one in spite of it, which tells you how strong everything else is. But it also means somebody out there is dialing an 828 number trying to reach a company in eastern North Carolina and getting whoever actually owns that number. That's a lost customer on every wrong call. And it wrecked their NAP consistency score. Eight out of thirty.
Elisabeth Pallante: And to be clear, NAP is name, address, and phone number. That's the thing about NAP, right? It's not very exciting. Nobody gets excited about making sure their phone number matches everywhere. But when it's wrong, it's actually costing you real calls.
Chad Treadway: So the second story. A Western North Carolina company, family-owned, started in 2020. Best About page in the whole set. A perfect thirty out of thirty. The owner is named and photographed, sixteen team members with real bios, no filler. Founding story, state registrations. These folks did the hard human work.
Adam Bennett: And?
Chad Treadway: Transparency came back at eighteen out of a hundred. No pricing anywhere on the site. No guarantee, no warranty, no cancellation terms, no refund policy. And their messaging leans on "people first, you're not just a number," which is fine, but I've also seen that exact language on a few hundred competitor sites, so it doesn't necessarily separate you.
Elisabeth Pallante: Right, the action does. Everybody likes to say they put people first and then you see it in the action. That's the piece we keep coming back to on this show. Building a real team page with photos and bios, that's the hard thing most companies skip. Putting a starting price on the page, that's technically easy. And this owner did the hard thing and skipped the easy thing.
Chad Treadway: A great About page is not going to save you if you skip the easy part. And it shows up in the score. You can have the best About page in the state and still not clear four hundred because you won't answer the first question everybody has.
Adam Bennett: Yeah, and just a side note, if your About page has the picture of your tech that's going there, I bet Fifi would recognize them and say come on in, it's safe back here. But we covered this in Episode 26. The transparency problem. Ninety-two percent of pest control sites are hiding their pricing. And that stat came directly from our own research across all those pest control clients that you did, Chad.
Chad Treadway: And this company is in that ninety-two percent, which really bugs me, no pun intended, because they clearly care about their customer. They just never carried that over to the one piece of information everybody wants before they'll pick up the phone. Third story here, and this is one of my favorites. A Coastal Carolina company, in business since 1973. Over five hundred sixty Google reviews at a 4.9. Highest review count in the set, more than three times over. And a perfect twenty out of twenty score on AI citations.
Adam Bennett: So Chad, what does twenty out of twenty on AI citations actually look like?
Chad Treadway: Twenty mentions within the top three source placement across Google AI Overviews or ChatGPT. Their blog posts are on fire ants versus red ants, termite swarm season, cockroach treatment. Somebody asks ChatGPT about local pest control in the area, this company's going to come up. The AI is pulling their content as the primary source.
Adam Bennett: That's the finish line we've been telling people to run toward.
Chad Treadway: It really is. And they got there. The content strategy worked. The AI knows who they are. But here's the gap. Thirteen staff members with photos on the site, but no bios. Not thin bios. Zip. Zilch. Nada. And the guarantee says "guaranteed results." No terms. No explanation. What does that mean? What happens if I'm not satisfied? Nothing.
Elisabeth Pallante: Yeah, it's not very clear. So AI is sending people to the site and then the site doesn't finish the job, right?
Chad Treadway: Right. AI citations are your new top of funnel. Getting cited is step one. Converting the person who actually lands on your site is still step two. Right now this company is winning step one and dropping step two.
Elisabeth Pallante: To be fair though, five hundred sixty reviews at 4.9 is enormous social proof. That might be enough to convert a lot of people even without detailed bios and a written guarantee. The reviews are doing a lot of trust-building work that the site isn't.
Chad Treadway: And that's fair. They're obviously a successful company with over fifty years in business. But my question is, how much more would they close if the site matched the reputation? Five hundred sixty reviews gets them in the door, but I'd like to see the site finish it.
Elisabeth Pallante: Yeah, I get that. It's true. It's also possible that the reviews are finishing the job and the site just needs to not get in the way, right? But ideally, I think you're right. It'd be best if both pieces were well-oiled machines with no hiccups.
Chad Treadway: Last one. A Triangle-area North Carolina company. Had the strongest content library in the sample. Authority score of ninety out of a hundred, and that's about as close to perfect as I've seen. Blog posts, educational hub, content freshness, all maxed out. 162 reviews at a flat 5.0. They published pricing, three tiers: twenty-six, sixty-four, and seventy-nine dollars a month. They named a guarantee, they call it the Pest-Free Guarantee, and it's got real terms behind it. Free re-services between visits. That's exactly what we've been asking companies to do for thirty-eight episodes.
Adam Bennett: That's really the poster child. That's what I want to see when I search for a pest control company. How much is it going to cost? And if it's guaranteed, what does that actually mean? Simple.
Chad Treadway: Right. But here's the kicker. There's no About page. Not a thin one. There just isn't one. It's like they forgot about it. I went and checked the live navigation myself. Services, Service Area, Resources, Contact. That's the whole menu. You can't find out who owns the company. That's a zero out of thirty on the About page check. And it earned every point of that zero.
Elisabeth Pallante: Wait. Best content, published pricing, a named guarantee, and no About page. How does that actually happen?
Chad Treadway: I think they built the business side of the site and forgot the human side. They've got the best content library in the set. They publish pricing most of their competitors won't touch. They've got a guarantee with real terms on it. Then they skip the one thing that tells the buyer who is coming to their house.
Adam Bennett: Can they still clear four hundred out of six hundred points with everything else they've got?
Chad Treadway: No. The verified score comes in around three sixty. That About page is a thirty-point ceiling they can't get through. The best content in the state doesn't matter if nobody can find out who runs the company.
Adam Bennett: So we've been pretty open about the fact that our tool got things wrong. What are we actually doing about it?
Chad Treadway: We found nineteen defects across seven runs and every one of them pushes the score down. Not one inflates it. So if the Cube Score tells you your site is strong, that's probably right. If it tells you your site is weak, that might be on us and not really yours. Although it does point to an issue you may see with other AI crawlers as well.
Elisabeth Pallante: That's an important asymmetry.
Chad Treadway: It is. And the fixes come down to three things. First, crawl detection. The tool has to know it never reached the site and say so instead of handing you a score. Second, it needs to read more of the page. Right now it's missing footers, image alt text, and any service page that isn't the homepage. Third, the checks that use AI to grade your content need input validation. If there's nothing there to grade, the score should say we couldn't check this, not that you passed.
Adam Bennett: And what's great is that with all of these scores that we've tested, we've actually fixed the tool. So now it's going to tell you and be an even stronger tool for you and for us to check out everybody's site. So one thing to fix this week, Chad. What's the move?
Chad Treadway: Go ask AI about yourself. It's the same thing as when we used to Google ourselves. Open ChatGPT or Perplexity, type in best pest control company in your town, and see if you come up. If you don't, that's your starting point. Either your content isn't strong enough to get cited or your host has a dog in the yard that nobody wants to go through. That's your fix. And it's only going to take a couple of minutes.
Elisabeth Pallante: Right, that's how you'll find out what the fix is.
Elisabeth Pallante: And download the Cube Score Self-Audit Worksheet. It walks you through all six sides of the score so you can spot your own gaps. The link is in the show notes.
Adam Bennett: If that test just made you a little nervous, subscribe so you don't miss next week. Hit follow on Apple Podcasts, Spotify, or YouTube. We're here every Tuesday. So let's do a quick gut check before we go. Did we make the case on these three?
Elisabeth Pallante: Number one: a website score measures what a machine can read, not what your business is worth. Number two: an AI-built site can outscore a twenty-five-year operation because it produces perfect machine-readable scaffolding. Form is not the same as substance. Number three: before you act on any website score, ask what the crawler actually retrieved. Then check whether your own hosting is blocking the crawlers you want.
Adam Bennett: If you want help figuring out where your site actually stands, visit marketingthatactuallyworks.ai for a free pest control marketing audit. We'll show you what's working and what's costing you money.
Elisabeth Pallante: And if you want the precise score with AI citation analysis and competitive comparisons, run your Cube Score at thecubescore.com. It takes sixty seconds.
Adam Bennett: If you got value today, leave us a five-star review. It really helps other pest control operators find the show. Next Tuesday, we're talking about upselling and cross-selling to your current customers.
Elisabeth Pallante: Thanks for listening to Pest Control Marketing That Actually Works. We'll see you next Tuesday.
