Optimizing for AI Searcher Behavior

---

• Google’s removal of the “num=100” parameter impacts data visibility and analytics.
• User behavior continues shifting toward zero-click searches.
• New frameworks like GEO/AIO and Invocation & Data signal the next frontier of discoverability.
• Brands must adapt content, visibility, and measurement strategies for AI-driven search.
• Authority, trust, and brand protection remain vital in AI-generated answers.
• Tools and KPIs are evolving — visibility, prompt tracking, and engagement are key metrics.
• The AI Search Impact Rubric helps assess exposure to disruption and readiness for change.

---

Youtube: https://youtu.be/YpjglOWf-CI?si=S3BFe4kaUA075ET9

Spotify: https://open.spotify.com/episode/0GyFx9peErlysv0SuSxJge?si=LDOBZ1F5Rwyz74jxLoHihA

Apple Podcast: https://podcasts.apple.com/us/podcast/optimizing-for-ai-searcher-behavior/id1755789311?i=1000732817981

---

77% of sites lost keyword visibility after Google removed num=100: Data
https://searchengineland.com/google-num100-impact-data-462231

What Our AI Mode User Behavior Study Reveals about the Future of Search
https://www.growth-memo.com/p/what-our-ai-mode-user-behavior-study

Google users are less likely to click on links when an AI summary appears in the results
https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/

Who Is the GEO Coach?
https://signalandfriction.substack.com/p/who-is-the-geo-coach

---

Bill Scully (00:01)
Welcome to episode 38 of the International Web Effectiveness Podcast, where Bill Hunt and I, Bill Scully, focus on the strategies necessary to improve your web presence and manage your global websites.

But before we get into it, I want to make sure you follow the International Web Effectiveness Podcast on YouTube, Spotify, or Apple Podcasts so you never miss an episode. And if you’re enjoying these conversations, please leave us a quick review — it really helps motivate us to keep making more shows.

In our last episode, we examined the “AI Easy Button” myth in an energizing chat with special guests Nick Leroy and Jason Kilgore. We dug deep into a significant industry challenge — the big lie being told to leadership right now that AI is some kind of instant replacement for people and processes.

Today, we’re going to do more of a broad AI search news update. It’s been a little while since we last met, so we’ll give you our take on a few different news articles. Bill, how are you doing?

Bill Hunt (01:11)
Good. Looking forward to this — there’s a lot of stuff being pushed out into the wild these days.

Bill Scully (01:17)
Yeah, lots of news, and we’ve both been traveling. We were both at the U.S. Search Awards banquet — that was pretty cool. What a great location.

Bill Hunt (01:34)
Yeah, amazing view from that rooftop, especially when New York is all lit up. There were some great winners, and it was just nice to get out and talk to people in real life again.

Bill Scully (01:48)
Yeah. So, one of the things that popped up recently is Google’s removal of the num=100 parameter. What’s your initial thought on that — without going too deep just yet?

Bill Hunt (02:09)
I’m surprised it took them this long to get rid of it, honestly. It’s not going to stop rank tools or scrapers from scraping; it just made their lives easier.

Back when we had Dataprism, we started scraping, and that parameter made our life easier because you could get a hundred results versus the static ten. All they’ve really done now is force people to manage parameters, maintain cookies, and all that extra work that raises costs and makes things harder for tool companies.

What I didn’t really expect was some of the fallout — people not being able to get data, and all the chatter about maybe not needing to look past page one anymore. That’s been mostly directional at best anyway.

And then, I saw people talking about the corresponding decrease in impressions in Search Console data. Folks finally started realizing how much crawling was being done out there. Interestingly, they didn’t blame it so much on SEO rank-checking tools as they did on AI position-checking or answer inclusion tools. That was an interesting shift.

Bill Scully (03:48)
Yeah, I just saw a note from another agency I collaborate with — they had no idea how to explain the drop in impressions. I was surprised, given all the information out there right now.

I always thought, though I never really confirmed it, that a lot of the information shown in Search Console was a result of these tools. And this kind of proves it — once Google turned that parameter off, all those impressions disappeared.

So I think this did a couple of things. Depending on your account, you’d notice that once you start removing all the page two through ten results in Search Console, your real numbers look better. Your click-through rate goes up, your average ranking position improves. Honestly, I’m glad they took it away.

Bill Hunt (05:12)
Yeah, I think we’re still going to see some of it creep back in eventually, because people are getting smarter about how much they extract.

When we were doing it, the great thing about getting a hundred results was that if five people were tracking the same keyword, we didn’t have to do five separate calls to scrape the results — just one. That was the efficiency of it. Now they’re having to do ten calls to get the same amount of data.

It’s interesting that you mentioned how removing results beyond page one can make data look better — better averages, better CTRs. But again, that data was always directional.

I used to look at rankings from positions 11 to 20 as a way to identify opportunities to push something back onto page one. But it does show how much noise was in the data.

When you bulk-exported ranking data, you’d often see old queries — sometimes five or six years old — still hanging around. That was common when people reused the same rank-checking tool for years or pulled extractions from Google Keyword Planner. You’d see stuff like “best mixers 2022.”

Who’s searching for “best mixers 2022” now? Nobody. But if that term was still stuck in your rank checker, that’s what was being tracked. You’d crack open the rank checker and, sure enough, there’d be all these old year-based queries still sitting there.

Bill Scully (06:44)
Thank you.
Yeah, yeah — rank checkers.

Bill Hunt (07:10)
And so, well, that’s where that’s coming from. And yeah, you know, it’s great — we rank number one for that. But I think it’s causing people to rethink a lot of this.

Bill Scully (07:21)
I didn’t mean to say that having that data wasn’t valuable. But when you’re just running general monthly reports, I think it’s a much cleaner report if you don’t include everything. Why give impression credit to being on the tenth page?

Bill Hunt (07:41)
Exactly. And that’s where that comes from, right? Because I remember having this conversation with someone who said, “If we rank 87th, how do we have 300 impressions when all our research and data show that most people rarely go past page three?”

So how are we being surfaced that many times if we’re ranking 87th? And then you have to say, well, it’s sort of the bots — because they’re fetching a hundred results, and you’re somewhere in that hundred.

It was always one of these things where, you know, we (well, maybe more I than you) would ding people for a lack of critical reasoning skills. And I think that’s part of it — this is forcing people to think more carefully about their data.

And you’re right — it’s not invaluable data. It’s still useful within its context. But I really like this idea of looking at things differently, because it ties into some of the other topics we’re going to talk about today — like clicks.

If you’re on page one or page five, one of the reports I always did — this was part of our “30 Minutes to Search Greatness” — was to look at what keywords we ranked in the top three or top five for, and whether we were getting a disproportionate share of clicks.

The idea was: if I’m in the top three, there are maybe just a couple of paid results above me, plus two organics — so why am I not being clicked? Especially for terms directly related to what I offer.

I think now it may push people to look at those kinds of reports more closely — to understand, “If I’m showing up, how do I get more of those people to click?”

It’s interesting because it’s forced the industry to rethink a lot on the tool side. Some said, “Hey, this is a 10x cost now because we’ve got to do ten queries.” But some of the smarter ones realized, “We can reuse the same session ID and cookie,” which obviously complicated things but helped gather consistent data.

And of course, others came in and said, “Why do you even need results two through a hundred anyway?” So again, it sparked a refreshing series of grown-up discussions across the industry.

Bill Scully (10:33)
Agreed. Anything else on that topic?

Bill Hunt (10:39)
No, I think that’s it. Just think about how you’re using that data — to your point. And maybe you can tweak your presentation layer, like Bill mentioned earlier, to show what’s happening with your page-one performance. That’s a really interesting report people can look at.

Bill Scully (11:06)
All right. The next thing we’re going to talk about is the Growth Memo that came out recently.

If people aren’t getting this, it’s an email newsletter from Kevin Indig — and actually, I didn’t realize it’s also from Amanda Johnson.

This edition is on “What Our AI Model User Behavior Study Reveals About the Future of Search.” And you brought this one up as a topic.

Bill Hunt (11:44)
I think this flows directly out of what’s happening on page one. What Kevin and the team did — and I forget which agency or company they used for the actual testing — was really interesting. Kevin and Amanda handled a lot of the analysis.

Basically, they looked at 250 unique tasks people commonly perform, and then broke those down into a series of seven broader task categories.

What stood out — and I’m just looking at their executive summary here — is that three-quarters of users’ eye movements never left the AI answer pane. They used a combination of eye-tracking and behavioral observation, and found that 88% of people didn’t even look beyond that area. They didn’t scroll down to the organic results, and they didn’t click out.

On average, users spent between 52 and 77 seconds looking at that AI-generated section. And that’s huge when you think about how things used to be. Back in the day, we used to say the average query took three to eight seconds — people would scan results, process hundreds of tiny visual cues, and then make a click decision.

Now they’re taking much longer, really reading and digesting the AI result.

And the big takeaway — which, honestly, we already suspected — is that their questions, their curiosity, or their intent (however you want to frame it) are being satisfied right there.

Unless the query is explicitly transactional — like “where to buy the best whatever” — there’s simply no click. In fact, 77.6% of the 250 use cases had zero clicks.

Another insight they shared was that user behavior is starting to resemble the early days of Google. People are skimming results — glancing quickly, forming an opinion, and moving on — but now they’re doing it entirely within the AI panel.

Bill Scully (14:11)
Yeah, exactly. So, they actually observed 37 people performing seven tasks, which created that set of 250 records.

And what really jumped out at me was that 77% of sessions had no external visits — no AI clicks or regular clicks. Which really shows how effectively Google is keeping people on the page.

Bill Hunt (14:54)
Right — and let’s take a minute to think about why that is.

They do a good job in the report of breaking down specific use cases. But it reminds me — I was at the barber the other day, and the guy in the chair next to me was talking with the barber. I didn’t catch the start of the conversation, but he said he loves AI answers because it just gives him what he needs.

He said, “I don’t need to click into one of those sites and dig through half an answer or read a bunch of nonsense. I just get the answer right away.”

Now, love it or hate it, that’s a real user perspective — a use case of one, sure, but it aligns with what this study shows. People are simply getting the answer they came for.

And this ties into a conversation we had before we went on air. Two agencies I’ve been working with — plus a prospective client — were all getting hammered over dropping lead numbers. Everyone’s asking, “How do we get our leads back?”

In a workshop I did recently, that was the number one question. And my answer was: that’s actually the easiest one to answer. You don’t. You won’t. You can’t. Because those leads are gone.

Then I explained — what kind of queries were bringing those leads in before? If it was something simple that users only had to click to get an answer, that traffic is gone for good.

I remember back when Google started eliminating sites that just told you things like, “What time is it in Stockholm, Sweden?” or ran basic currency converters.

Now Google just shows that information right on the results page. I don’t have to click through nine thousand ads and clutter just to find out what time it is in Stockholm.

Bill Scully (16:55)
Right — currency converters, exactly.

Bill Hunt (17:19)
They don’t have to click now. So what we need to think about — and I’m still trying to come up with the right wording for this — is: how do we become sticky?

Are we in the answer? Are we contributing to the answer? Is it an answer where there’s a brand association attached to it?

That’s why Kevin and his team found that when something was transactional, people often went directly to Amazon, the brand’s own site, or another vendor. And that’s creating chaos in analytics, because now that traffic is being attributed to direct instead of search.

So I asked them, when they looked at these cases, “Did you examine the search results and try to interpret why people aren’t clicking?”

Because if I’m timeanddate.org (or whatever new extension exists — .net, .ai, etc.), and someone searches “time in Stockholm,” well, that query’s answered instantly. That traffic is dead. I’m never getting that click again.

So the question becomes: Can I do anything with that? Can I add something interesting — maybe a note about daylight savings in Sweden, or a small piece of related content — that keeps users engaged somehow?

If the AI summary is sticky — meaning people are spending real time reading it — then we need to ask:

• Are we part of that entry?
• Are we additive to that entry?
• Is there paid content above it?
• What’s the secondary query a user might infer from it?
  

Because right now, none of those agencies I mentioned were thinking about that. And, not surprisingly, two of them said, “Is there a tool that can do this?”

I told them, well, there are tools that partially do it — most rank trackers and SEO suites now show whether there’s a featured snippet, a quick answer, a local pack, etc. But they don’t tell you what those features actually contain.

And of course, one person said — as we joked about earlier — “AI is the answer.” He said he could probably write a prompt to scrape and analyze those results for him. I told him, sure — but when you write that logic, make sure you think like a human. Because that prompt might come back and tell you, “Yeah, this doesn’t need a click anymore.”

So yes, maybe there’s a clever prompt or tool that can help deduce this. But at the end of the day, for your top five, top ten, top twenty queries — be the human. Go look at those results yourself. Use your brain. Decide:

• “I’ve got to concede this one,” or
  
• “I can add something here.”
  

You might even create what I call intent triggers — small prompts or hooks that lead users to take the next step, maybe click something in the right rail, or continue down a path you control.

Can we find a way to make that work for us?

Bill Scully (20:44)
Exactly. I think if you can become a bit of a ninja with that — like you said earlier — and think ahead to the next question, then you can create content that appears in results where it implies you have the next answer.

That’s how you guide people down the path you want them to take.

But yeah — big behavior shifts are happening here. People are reading and deciding within the AI mode.

Transactions are really the only guaranteed clicks left. And maybe — I don’t know if this is the answer — but I think visibility is replacing traffic as the new currency for SEOs.

That’s probably why we’re all more focused now on web efficiency — not just rankings or being in results, but understanding how we appear there.

Because when you look at all this, there are definitely impressions being made. And if you can get your brand included in those, that’s a massive win for brand awareness — even if users never visit your website.

Bill Hunt (22:12)
Exactly — and that’s why we need to look at this closely. We’ve talked a lot about brand, brand building, and awareness — and I completely agree with you.

And actually, you’ve converted me! So there it is — Bill Scully has officially turned me into a brand advocate.

Bill Hunt (22:38)
Where a brand is relevant in this context — that’s really what we need to think about. And I think this is something SEOs often get wrong. Well, really, humans do this — we like things in neat little boxes. We want one answer that fits most situations.

But here’s where I’m going with this: when I search “What time is it in Stockholm?” there’s no long, detailed story attached to that. There’s no need for background about how daylight savings time started in Sweden in 1492, or some tangent about the average coffee consumption of a Swede.

You don’t need that storytelling marketers love — the “grandma’s recipe cooked over an open fire in a tin pot when she emigrated to America” type of narrative.

Storytelling absolutely has a place — but not in purely factual queries. If it’s “just the facts,” then there’s nothing we can really influence there.

But if the query shifts to something like “What’s the best time to visit Bhutan?” — now that’s a different story. For example, we’re going to Bhutan in December, and the first thing I searched was, “What’s the weather like in Bhutan in December?” because I wanted to know if I’d be freezing and needed to pack mountaineering gear.

That’s a question where the answer can expand. Maybe it’s the National Weather Service of Bhutan giving the factual part — but underneath that, maybe it suggests how to dress for the season.

And if I refine it further — like “What should I pack for a December climb to Tiger’s Nest?” — that’s where there’s real value. Someone could answer that completely, give practical advice, and earn authority for doing so.

Bill Scully (25:02)
Ahem.

Bill Hunt (25:09)
Right — the key gear points might include things like essential layers, good shoes, gloves — and because you answered that comprehensively, you become deserving of that visibility.

That’s the key — being complete, relevant, and helpful. That’s where storytelling contributes value. That’s where the gold is.

It’s not in “What’s the temperature in Bhutan?” — that’s just a quick factoid. But when the query becomes discovery-based, that’s where the opportunity lies.

And to your point — are we there or not? Is Bubba Joe’s Hiking Bhutan blog there, or is it an official guide or tour site? Usually, it’s the one that wrote something complete, referenced, and relevant.

We’re actually going back to the old days — writing content that’s comprehensive, factual, and cited by others.

When I come back from my Bhutan trip and write my blog post about hiking Tiger’s Nest, I’ll reference the sources I found useful — the ones that gave me gear lists or tips for picking the right guide. That’s how the ecosystem builds authority again.

So the key questions for anyone are:
A) Am I already part of the answer?
B) Could I become part of the answer?
C) Can I tweak my content to better fit the query intent?

And we’re seeing this “old-school” approach making a comeback — even gamification is hot again. People are trying tactics like adding links in their footers or manipulating “best of” lists.

Bill Scully (27:29)
Yeah — make a list of the best product companies and put yours at the top.

Bill Hunt (27:36)
Exactly — the classic listicle.

We’re seeing that again, and sure enough, we’re getting the usual emails: “Hey, I can get you to the top of that list.” Sure, it might work — for a while.

But today there was a bigger discussion I want to touch on for a second. I’ve been talking to a few people and honestly want to know — who, and at what level of arrogance, didn’t anticipate the gamification of AI results?

Bill Scully (28:13)
I’d guess the engineers working on AI outputs just haven’t set up their systems yet the way organic search results are structured.

Bill Hunt (28:29)
But why not? That’s the question. Maybe it’s one for a different podcast — but still.

Bill Scully (28:31)
I’d say, and this is just my prediction, they had to put something out fast — because ChatGPT and other tools were dominating the headlines.

Bill Hunt (28:45)
But why hasn’t Chat done it either?

Bill Scully (28:48)
They’ll never be able to — well, I shouldn’t say never.

Bill Hunt (28:52)
I think they can. If we go back to that argument I keep making — the one people say is too complex — these eligibility gates could easily be structured using something like the old PageRank model.

Something as simple as the true PageRank, not just a link count.

Bill Scully (29:12)
Yeah, you could create a subset of documents that are used for evaluation. That makes sense.

Bill Hunt (29:49)
Michael Bonfield and I were talking about this yesterday — that we should do a session on bringing back the old-school stuff. Because you know, back then, the way we used to count our link builders was by asking: Am I being cited in the body of the text, like you would in a proper research paper?

For example, “The sky is blue because Galileo or Copernicus said so — it’s the filtering of light through whatever.” Versus someone who just writes, “Yeah, the sky is blue,” and then throws in forty random references they probably never read on a bibliography page. That’s the old “Add-a-Link” page.

And this is what we started finding — when Google began clamping down on authoritative relevance and contextual relevance, that’s when a lot of the cheap link-building tactics got wiped out.

I think they can apply this again, but it’s an interesting question — and I’ll take your answer as “Google was in a rush to get this out the door.” That’s why they didn’t do it. But where it’s showing up most right now is exactly in the RAG stuff — Retrieval-Augmented Generation.

Because there’s no other way a listicle would ever get into a result set if it weren’t for RAG. What’s already baked into the system is baked in. So now, all they’ve got to do is start applying a level of rigor to RAG. Maybe that means ending up with a twenty-person quality team inside AI just to debunk this stuff.

But here’s an interesting segue — I think it was Brett Tabke or someone who wrote recently that a lot of people are starting to see penalties in normal Google results for doing some of this AI-generated stuff. So it seems like Google is using its existing quality systems to hammer the people who are participating in these schemes that affect AI overviews.

Bill Scully (32:12)
My guess is that’s more of a correlation — what’s really happening is they’re unoptimizing their pages by writing junk for AI. But yeah, it’s interesting to see that a lot of the stuff being cited in AI results doesn’t actually appear in the search results.

Bill Hunt (32:23)
Exactly.

Right — they may cite it and they may reference it. And here’s another interesting thing — there was a study that said changing the date of a research paper is often nothing more than changing the date. I forget how big their sample size was, but it’s wild.

Bill Scully (32:42)
Yep, yep.

Bill Hunt (32:54)
And — what’s a politically correct way to say this? — how many hastily thrown-together, angle-serving “authoritative” research papers are being pushed into the market right now?

I’ve seen these — they look like official studies. There are maybe five people behind them, probably computer science grad students at some university.

The first big one was the so-called Bible of Geo — “Factors Influencing Geo Search.” It looked academic, but it was basically written by five people from an SEO company. Then another came out, which I wrote an article about — though I didn’t post it because, honestly, I wasn’t in the mood to get bitch-slapped online that day.

Bill Scully (33:47)
You—

Bill Hunt (33:47)
Right — where that paper apparently validated schema as a critical factor in AI results.

But the issue is this: SEOs tend to conflate structured data with schema markup. Anytime “structured data” is mentioned, they assume it means schema.Bill Scully (34:20)
Yeah — that’s a misnomer, right?

Bill Hunt (34:24)
Exactly. And that’s what happened here. Someone took one statement and immediately equated A to B — but the study didn’t actually test it.

Now, I’ve seen a few people say they’ve tested it themselves and it seems to work — maybe it does in AI overviews or in Bing. So yeah, do schema. In fact, my next call is about schema, within that same context.

Anyway, my point is that people are getting really creative. It’s a very exciting time.

And where we were going with this — Google recently removed the “100 results” parameter that we used to tell people to use. We’d say, “Make sure you’re checking what’s ranking on page one,” because page one now — with AI — looks completely different.

According to data from Kevin (and supported by Pew Research), people really aren’t clicking anymore. Why? Because they don’t need to.

Profound even held a “Zero Click” event in New York — I didn’t attend, but I read the notes. A few friends were there.

One big takeaway was that we need to focus again on building community — not just in search, but across brand ecosystems.

And this ties back to your earlier point on brand building — what the folks at .dashMeredith are doing. They’re taking their sites and trying to create community, real connections, and collaborative networks.

It’s kind of a return to the old-school days, when people would meet at PubCon over a beer and share knowledge. Or like the Enterprise Tech Council, which started with five of the biggest B2B tech competitors sitting down at dinner saying, “How are you selling this idea to your boss?”

We need more of that — more collaboration, more real conversations.

And the last thing — something I’ve become a big believer in, thanks to my good friend Jeremy Sanchez — is this idea of invocation.

I actually want to leapfrog this whole “getting into AI results by skewing content” game. Forget gamification — just do what China did. They didn’t bother laying telephone lines — they said, “Screw it, let’s go straight to mobile.”

They leapfrogged a whole generation of tech. That’s where my head’s at. Why waste time—

Bill Scully (37:33)
Mm-hmm.

Bill Hunt (37:42)
Doing this — and I mean, Jeremy sold me on it with one simple query he did. It was something like, “Tesla Model whatever, in blue, near me.”

There was no dealer evangelizing, no brand affinity — he didn’t care about the brand story. He just wanted to know: Where can I find one?

And that type of result — nothing could produce it. I mean Google tried, and the other bots tried, but they had to go out to all these sites and pull it in. The premise was simple: What if Tesla, or the dealer, or whoever, had this data in their ERP backend and fed it directly?

We’re seeing this now with what Perplexity is doing — and ChatGPT too — both have browsers now. These guys want to pull everything in. Look at Walmart, for example.

I remember a project — this might take us off-topic a bit, so bring me back if it does — but I think it was around 1998. I worked with a guy who made synthetic oil for race cars and wanted to adapt it for consumer use. In doing that, he created a product kind of like Fix-a-Flat — though maybe not that exact brand — but this one was for bicycles.

He wanted to reach people who did off-road and street biking, and eventually expand to China and Japan because those countries had the most bicycles per capita, next to Amsterdam. That was his plan — he wanted to sell it internationally.

Bill Scully (39:21)
Mm-hmm.

Bill Hunt (39:36)
And he said, “By the way, I’d love to get into Walmart.”

I told him, “Well, then you’ve got to go to Arkansas and pitch it.”

He did — and long story short, we used the web to find people who didn’t even know they needed this product. But when he finally got into Walmart, they made him change everything: pricing, descriptions, packaging — the whole deal.

And if you think about it, that was the precursor to product feeds — all the structured data Walmart needed to put his item into their database. They wanted to know: What’s the price? What’s the market price? What’s the price differential?

I went through something similar with Pernod Ricard years later — working with e-retailers like Total Wine. We had to help them clean their product feeds.

I even wrote about it — we once had “P. Noir” in our feed for Pinot Noir.

Bill Scully (40:10)
Mm.

Bill Hunt (40:32)
And I asked the wine team, “Was Pinot really that long of a word that we had to abbreviate it?”

They said, “No, we just had a fixed number of spaces — that’s all.”

But what happened was, every retailer that took our feed started using “P. Noir” — and we saw it showing up in search results everywhere. When we asked them to fix it, they said, “That’s not our job — it’s yours.”

So here’s the long story short: what was important 25 years ago for that guy trying to get into Walmart… what was important 8 years ago with Pernod Ricard and cleaning product feeds… it’s still important today.

I’ve seen people agonize over their Google product feeds, but here’s what I’m curious about — how many people are getting pressure from Amazon and Walmart to fix theirs?

Because these companies are going fully agentic — they both now call themselves “Google killers.” They want people searching directly on their platforms. They’ve got search suggestion tools, product finders — and all of it depends on one thing: clean, optimized feeds.

And here’s the funny part — those same feeds feed into schema.

It’s fascinating to me how many people don’t make that connection.

Bill Scully (42:04)
Yeah.

Bill Hunt (42:13)
And yet, they’ll say, “Oh, we don’t have the data to populate schema tags through JSON.”

But those same companies will pay someone a million dollars to optimize their Google product feed!

Bill Scully (42:20)
That’s mind-boggling.

You don’t know what your product is? You don’t know the weight? The name? You don’t have a good description? Come on.

Bill Hunt (42:34)
Exactly — it’s mind-boggling.

Jeremy and I sat down for dinner — what was supposed to be an hour and a half turned into four hours. We were trying to break it all down: How do we crack this? How do we get people to think differently?

And I think the change has to come from inside.

It was good to see Profound talking about invocation too.

And one last thing — I did a workshop on AI recently, for CMOs. Before the session, I ran a survey asking, “What do you want to know?”

The open-ended question was, If there’s one thing you absolutely need to leave this session knowing, what is it?

When I charted the results, the number one answer — at 22% — was strategy and readiness. People wanted to know, How do I prepare for this?

This was right after I’d already covered “What’s the impact of what’s happening” in two to three slides. Then they asked, “Can you come up with a set of questions to help us figure out if we’re ripe for disruption?”

I think maybe on our next podcast, we can dig into that. Because it boiled down to about five or six key questions.

Like: Have your customers’ behaviors shifted?

I saw a stat — I think it was from Gartner — that said 90% of B2B procurement agents now use some form of AI, typically ChatGPT, to source products.

Bill Scully (44:38)
Yeah, I mean, that’s a great application. You basically write a detailed prompt, and ten minutes later it comes back — deep research, tons of variations, every meta angle you could imagine.

And like you said, you can even ask it, “Why am I not ranking?” It can evaluate competitors too — tell you why you should or shouldn’t talk to them.

Yeah, yeah.

Bill Hunt (45:10)
Exactly. That’s what I told him.

I gave him some prompts — told him, start with something simple like, “I’m a [your role], looking for who makes which blue widgets,” or whatever your product is.

And if you don’t show up? Ask it why.

Just like we did when we ran that test…

Bill Scully (45:22)
Yeah — I’m a purchasing agent.

Bill Hunt (45:39)
Right. We posted that article about what we found when we searched for “Absolut Vodka.”

We asked it, “Why doesn’t Absolut show up?” — and it actually told us why.

Bill Scully (45:51)
Yeah, it was a specific search — “best vodka for cocktails.”

First we did “best vodka,” and then we asked it, “How did you define best vodka?” That’s what excluded it.

Then we changed it to “best vodka for cocktails,” and suddenly it ranked much better.

Bill Hunt (45:54)
Exactly — “best vodka,” then “best vodka for cocktails.”

That’s what converted me into being a brand marketing advocate, because in that result — especially for the cocktail query — it explained why Absolut didn’t show up.

It said the bartenders who participated in the test — basically a blind taste-off — didn’t rate it as highly. There were comments about taste, and it didn’t make the top lists from places like Epicurious.

So, back to the point — if you’re not in those authentic, authoritative sources, you just don’t make the cut.

And honestly, if you’ve got a crappy product, it’s still a crappy product — even if you rank number one.

Search used to be able to elevate bad products, but now? Bad products stay bad because people won’t talk about them favorably.

Bill Scully (46:49)

 Yeah — brand citations.

Bill Hunt (47:03)
Exactly.

And that’s the key — asking those potentially uncomfortable questions: Why weren’t we in the results set?

You might be surprised. Then you can go out and fix it.

Because one of the big questions people asked me recently was, “How do we deal with negative results?”

Well, what’s actually negative about it? What did they say?

If it’s “crappy customer service,” then fix your customer service.

If it’s “grainy product with an aftertaste,” fix your product.

If it’s legitimate criticism, address it. If it’s incorrect, trace the source and see if you can correct it.

It all goes back to reputation building.

So if we step back and recap — the big story lately is: search as we knew it is over.

We need to rethink it.

That’s why I like what Mike King and others are saying — we need to reimagine our job.

In fact, according to Grammarly, last week I wrote 387,000 words worth of stuff. One week!

I just need to start sharing it.

But the reason is simple — I’ve been obsessed with web effectiveness. That’s what we’re advocating now:
If search as a channel is disrupted, then what’s next?

Bill Scully (48:47)
Woo! Yeah.

Bill Hunt (49:01)
Because the people who are clicking — they’re 10x, 12x, even 100x more likely to convert.

They’ve already done their homework. They’re vetted. They got their questions answered.

So, are our funnels — to use your term — too leaky?

Bill Scully (49:13)
Yeah. They’re motivated.

Bill Hunt (49:31)
Exactly. These are the most qualified visitors — and yet, are we converting them?

That’s what I’m digging into for my new book on web effectiveness — this idea that we shouldn’t accept our “perfect” funnel diagrams at face value.

Because while we’re staring at the right side — the people who did convert — we’re not looking at all the holes on the left side where people are dropping off.

Bill Hunt (50:00)
So I tell this story about a senior VP at IBM who once said to me, “I don’t get out of bed for twelve million.”

And, you know, it ended up being one of those career-limiting moves because I looked at him and said, “Dude, that’s your problem, not mine.”

You’ve got all this junk — gates, speed bumps, friction, pain — in front of your customers. I’m bringing you two million people, and only six of them make it through your gateway. That’s not on me. That’s on you.

After he got over the shock of some low-level guy calling him out, he said, “You know what? You’re right. We need to look at that.”

He’s actually the one who came up with the metaphor: we need to look at the holes on the left side of the waterfall — or maybe the bucket’s a better metaphor.

And that’s exactly what people need to do: look at why these hyper-qualified people aren’t moving forward.

You’ve probably got a great Looker report for this, but ask yourself — what’s happening on your page that’s stopping that flow? Because these aren’t casual visitors anymore.

They’re hyper-qualified.

So, what are we doing — or not doing — that’s blocking them?

You work with a lot of lead-generation teams, right? Off the top of your head, what are a couple of things you see people doing in that lead pathway that they could fix right now?

Bill Scully (51:45)
Yeah, totally agree.

Off the top of my head? The number of fields on a form, for starters. Sometimes just finding the form is a pain.

Then there’s intent alignment — are they even on the right page? If a user lands on a page that doesn’t answer their question, they’ll hit the back button in two seconds.

And if your page doesn’t help lead them to their next question, then they’ve got no reason to stay.

Speed, too — slow pages kill conversions.

Those are the big ones I see. Got any more?

Bill Hunt (52:43)
Yeah — friction.

Speed, load times — are they waiting around?

And don’t even get me started on forms with forty qualification questions. Come on.

Bill Scully (52:53)
Right? And then they ask, “Do you want our newsletter?”

I’m like, if I fill out this form, I know I’m getting spammed.

Bill Hunt (53:00)
Exactly.

You fill it out, and suddenly you’re getting 900 emails and 40 sales calls a day.

Bill Scully (53:08)
That’s the biggest friction point. Nobody wants to talk to a salesperson until they need to talk to a salesperson.

So put the information on your website! Give them the answers upfront.

Understand the purchase decision process — meet them where they are.

Bill Hunt (53:18)
Right, right.

So there’s your homework, listeners:

Go old-school. Run a query for something where you’ve lost leads, and go through your own process.

And — quick story — I actually love getting negative comments now because they motivate me to write more.

Bill Scully (53:37)
(laughs) Yeah?

Bill Hunt (54:02)
Yeah. So here’s an example.

We live in Connecticut, and every fall the sprinkler company has to come out to blow out the water lines. They sent me this email saying, “You either have to prepay or be home when the crew arrives so you can pay by credit card.”

All for my convenience, of course.

So I’m reading this email, and it says: “To prepay, go to our website, click the ‘Billing’ tab, and call us to give your credit card.”

Okay — sounds fine. But what were the two things missing from this urgent email to prepay?

Can you guess?

Bill Scully (54:44)
(laughs) I mean, I’d assume there’s a Venmo link or something, right? Just click and pay?

Bill Hunt (55:02)
Nope. Not even close.

Two things missing:

First — they said, “Go to our website and click the Billing tab.” But what should have been in that line?

A link.

There was no link. None.

And the second? “Call us.”

No phone number. Not even a click-to-call link.

Bill Scully (55:17)
Oh my god. No link? No phone number? That’s insane.

Bill Hunt (55:35)
Exactly!

No click-to-call. No convenience at all. And, come on — we’ve all got smartphones. If you put the number there, I can just tap it and call you instantly.

They did have their address in the footer, but no phone number.

So I go to the website, dig around, and try to find “Billing.” Of course, I forget that’s even the term, because they bill and I pay — totally different mental model.

Bill Scully (55:44)
That’s crazy. Just have a big Pay Now button at the top!

Bill Hunt (56:05)
Exactly! A big “Pay Now” button.

Instead, there’s a menu for “Get a Quote,” “Winterize Your System,” “New Sprinkler Installs” — everything but billing.

And here’s the kicker: when I finally click “Billing,” the form looks like it’s straight out of 2004.

It says “Pay with Credit Card,” with an empty “Amount” field and an “Invoice” box.

No idea what the amount should be — no lookup, no auto-fill.

The email said, “Type ‘Winterization’ in the invoice field,” so I did that, thinking it might look up the price. Nope.

Didn’t work.

So I just typed in “$100” and clicked submit. That’s when I finally found the invoice field.

Bill Scully (56:46)
(laughing) You actually had to guess the amount? Unreal.

Bill Hunt (57:04)
You know what someone said to me on LinkedIn? They said, “Well, they’re not web people.”

And that’s exactly my point. In my article, I wrote that it’s their fault for not following their own process and spotting the friction. It’s also the fault of whoever set up that email. And it’s the fault of the agency that built it.

Two out of those three can absolutely be held accountable for the dysfunction — because they’re the ones managing the tech.

The owner? Brilliant guy. Total artist when it comes to sprinklers and landscaping. But he doesn’t know jack about the web — nor should he. The people he hired, though, should know better. This should never have happened.

Anyway, going back to our main point — go out, do that query, see what shows up.

Bill Scully (57:45)
Right, exactly. He shouldn’t have to know that stuff.

Bill Hunt (58:00)
Right — see if you can affect what shows up. Visit your own website, look at the experience the way Bill was saying — maybe it’s out of context, maybe the AI pulled snippets from all over your site, but the link it serves up isn’t actually helpful.

See if you can reorganize or re-architect that.

A lot of tools out there will now scrape the top results and magically, through the “power of AI,” write you a page that includes what they have and you don’t. Maybe that’s useful — if it makes sense.

Then follow the conversion process all the way through. See if people can actually hand over their hard-earned money to you.

That’s the assignment for today: find out how much friction, pain, and suffering you’re putting your customers through.

Bill Scully (58:54)
Yeah. And hey — based on this conversation, with the end of the year coming up, people should really start thinking about how to pitch for budget.

Make sure you’re building out your web presence management — even if people aren’t actually visiting your site. Don’t let “no traffic” be the reason you don’t get funded.

Bill Hunt (59:27)
Exactly! I actually wrote an article about that on my Substack, Signal and Friction.

It’s called When Your SEO Currency Goes Away.

For years, traffic was the currency we bartered with — “Do this for me, and I’ll bring you this much traffic.” That was how we justified investment. But now that model’s changing.

So I wrote a few examples around that, and I’ve got a new four-part series coming out while I’m traveling. It’s called The Chicken or the Egg.

Bill Scully (59:30)
(laughs) Okay, I’m already intrigued.

Bill Hunt (59:58)
Yeah — so the series started because someone at OpenAI said it feels unnatural to go to a website these days.

That caused a total firestorm — everyone debating whether websites are even necessary anymore.

And then, the very next day, a developer asked me, “Do we even need websites anymore?”

So I wrote three parts:
Part one tackles the OpenAI comment.
Part two is literally The Chicken or the Egg — reminding people that your website isn’t the only source of truth, but making the argument for why you still need one.
And part three explains how to adapt your site for feeds and structured data so it stays relevant.

Then there’s a possible fourth part — a more MBA-style, data-driven approach, almost like a CapEx validation, for proving your website’s necessity.

It’ll all be up on Signal and Friction in about two weeks, once I finish my Web Effectiveness case study.

And — oh, I can’t wait for the hate mail — I’ve got a post coming called Should an MBA Run Your Web Effectiveness Program?

SEOs will immediately assume I’m saying an MBA should run SEO. But I actually put a disclaimer at the top: Read before commenting!

Because what I’m really saying is, you don’t need an MBA — you need MBA-like thinking.

So that’s a three-part series that just dropped, followed by Do We Still Need a Website?

Bill Scully (1:02:10)
I totally agree. It’s an important discussion.

Bill Hunt (1:02:22)
Yeah — and it’s going to make for some great reading.

Bill Scully (1:02:24)
All right, we’ll have links to all of that — plus everything we talked about today — in the show notes.

You can find them in your favorite podcast app, or on our website at internationalwebmastery.com.

I actually really liked today’s episode — talking through the news and what it means for us.

Any final thoughts?

Bill Hunt (1:02:52)
Just one: let us know what you want to hear. Seriously.

And if there are guests you’d like us to talk to — like last time, that episode got such great feedback — tell us! We’ll try to bring them on.

Bill Scully (1:03:18)
Perfect. Well, we hope you’re finding value in these conversations.

If you’re enjoying the show, please take a moment to rate and review the podcast on your favorite platform.

You can find us on all the social media sites — that’s a great way to give us feedback too.

Thanks, Bill. Talk to you soon. Bye!

Bill Hunt (1:03:39)
Cool — thanks. See you next time.