Mike Downer: Hi, everybody. I’m Mike Downer, your host. Welcome to The Unlearning Lab with JAR Consulting. I’m here with Kevin Wosmansky, President, Founder, CEO, and Chief Warehouse Officer. How are we doing today, Kevin?

Kevin Wosmansky: Good, buddy. Good. I love it. You know, I like how I get a new title every time. “Warehouse Officer” is a new one, considering we don’t even own a warehouse. But we do seem to have a lot of stuff in our business, like cameras and all kinds of equipment, so I’ll give you that one.

Mike Downer: Already. Okay, that actually is fair. So, Kevin, last time we were talking about mistakes. Here’s what we’re going to do today: we need to talk about the top five mistakes businesses make when hiring a marketing agency in the new AI world. I’m going to let you take it over and tell me what those five things are.

Kevin Wosmansky: I love this question. This was fun. We kind of stumbled across this during our last recording.

Mike, the top five mistakes businesses make when hiring a marketing agency in the new AI world are: keyword stuffing; GEO listed as a service with no definable tech stack; siloed web development; lack of third-party brand presence; and measuring clicks and conversions only, without tracking citations and conversions.

So those are the five top mistakes that businesses make when hiring a marketing agency in this new AI search world.

Mike Downer: I’m sure there are a lot of people like me out there who are not informed on any of those words, and that all kind of sounded French to me. So if you could start breaking these down one by one, I’m sure people of my caliber would really appreciate it.

What exactly is keyword stuffing?

Kevin Wosmansky: Keyword stuffing is just taking tons of keywords about your business and shoving them into your website or into blogs. This is kind of the dinosaur mistake.

If you’re hiring an agency that’s still pitching high-density keyword strategies, long-form fluff, 3,000-word blog posts that take forever to get to the point, or traditional metric promises like ranking number one on blue links, here’s the thing: large language models don’t care about keywords and keyword density. They care about semantic clarity and intent matching.

Those are the things that matter regarding content.

For example, if you have an agency that wants to promise to write a massive wall of text just so they can stuff 50 of your keywords in every single time, AI engines don’t like that. AI and large language models scan for direct answers, and the sooner you can get to the point, the better.

That’s an old strategy. Ten or fifteen years ago, you could write tons of fluff and it would yield results. That strategy is dead.

Mike Downer: Okay, so I remember when we were doing this as kids, you know, seven years ago when I started with you. That was a big thing to do. So now when you say that, it’s like, “No, don’t do that.”

Kevin Wosmansky: Well, hang on. We didn’t do keyword stuffing. We would write a lot of content, and you would try to create as much content as possible.

I’ve seen a lot of examples from prospects and businesses that came to us. We would look at their websites, and there are tools today that run through your website and tell you, “You’re using this keyword too many times. Google is going to penalize you for that.”

Mike Downer: No, it makes sense. It’s just getting smarter. You’re helping everyone keep up with it.

The second thing you talked about, which I don’t understand, maybe you can clarify for me, is GEO listed as a service with no definable tech stack. What does that mean?

Kevin Wosmansky: All right, this is a fun one. I was going to make this number one, but I thought we’d have an easier lead-in.

Generative engine optimization, AI search optimization, universal search optimization, large language model sentiment training — all of these brand-new, shiny buzzwords. I’m seeing agencies add AI search optimization and all these different things.

The problem is this: if you do not have a sophisticated tech stack to measure and track this, then it is nothing more than shiny buzzwords.

The reality is that agencies should be talking about entity mapping, schema markup, LLMs.txt files, AI crawlers, and they should have all of the technology and tools to map and track all of that.

We’re in this new world where everybody wants ChatGPT to recommend their business, and businesses are quickly picking up on this trend.

So when we say that, if you’re looking at a marketing agency and they’re using all these newfangled, shiny buzzwords like GEO, which we use and which are on our website, ask the agency, “How do you track this?”

At JAR Consulting Group, we have a multitude of tools that we use. We want to track how many AI crawlers are running through your website. We want to track what the actual sentiment is. What does Grok, ChatGPT, or Gemini think? What sentiment rating do they have toward your business?

When we bring a client on, we do a lot of analysis and audits upfront to show the client all of this. At the end, we want to know: can Perplexity and ChatGPT see your website? Are they scanning and crawling your website? What is the sentiment? What do they think about your website and your brand?

Real GEO agencies are using tools that can simulate AI queries, build matches, and look at the architecture of the website. It’s a lot more than just shiny buzzwords and saying, “We’ll put content on your website for $299 a month.”

I would probably stay clear of companies that are just trying to say, “Here’s an all-in-one solution. Sign up today, put your credit card in, and we’ll do this for you.” In my experience, almost every single time, those solutions are not going to deliver for you.

Mike Downer: Perfect. Okay, now I actually understand that. That’s good.

Number three: you said siloed web development. Tell me what that is.

Kevin Wosmansky: I say siloed web development and SEO. Here’s the mistake: letting a web development team build a site that’s entirely separate from your search engine optimization or generative engine optimization team.

What I’m talking about here is that they’re building a site, but it’s not really focused on your SEO or GEO strategies. This leads to JavaScript-heavy, slow, siloed, or gated sites that AI crawlers and AI robots can’t read.

The reality is that large language model crawlers need lightweight, accessible code.

Your web development team might build this beautiful website with all these great, shiny bells and whistles, but if it’s uncrawlable, what good is that? Humans see the website, and it’s aesthetic and beautiful, which you need for a great user experience. I’m not negating the user experience for humans.

But you also have to build the UI so that machines can read it and crawl it. I think that’s really important. As I’m trying to educate our clients, we need a website that is very readable for machines while still creating a great experience for humans. Does that make sense, Mike?

Mike Downer: Yeah, it does. That cleared it up because I did not know what you meant by that when you and I first started talking about it.

Kevin Wosmansky: Let me drive that point home for you.

If your website is a black box that requires heavy rendering, has big walls, and is very heavy, cumbersome, or large, large language models are going to treat you like you don’t exist.

Web development and AI optimization are married now more than they ever have been. Your code structure will dictate your visibility.

That’s probably one of the most important takeaways for a business owner: you might have invested heavily in a website one year, two years, or five years ago. Guess what? It’s most likely that the code is set up in such a way that large language models are not able to easily read it.

Mike Downer: That makes complete sense. You have to keep up with the times.

The fourth thing we discussed was lack of third-party brand presence. What do you mean by that?

Kevin Wosmansky: We’ve talked a lot about your website, right?

A business that only focuses on its website probably isn’t going to get recommended by large language models. Why? Because large language models are trained on the entire web.

When an AI assistant or agent recommends the best digital marketing company or the best digital ad software, it’s looking at a website, but it’s also looking at third-party recommendations.

That includes Reddit, product reviews, whether you qualify to be on Wikipedia, whether you’re on tech forums, and whether you’re mentioned in other articles.

The point is that you have to focus on your website, but you also have to ask: what do third parties think about your brand?

An agency can optimize its blogs perfectly, but if no one else is talking about you online, AI is not going to trust you.

Modern GEO requires digital public relations and community presence because LLMs establish trust by cross-referencing other authoritative sources.

So when we talk about this, yes, your website is critical, but now we need to focus on the broader spectrum of your business’s entire digital footprint.

Mike Downer: Okay, now that makes complete sense. You’re doing a great job. You do a really good job of talking to me like I’m a five-year-old, Kevin, so it makes things very easy for me.

Kevin Wosmansky: I have teenage daughters, Mike, so I have to make it really digestible. By the way, that’s what large language models want, right?

The days of 3,000 words of fluff are gone. Mike, you know me. I can drone on and on and on, so I have to get very specific and explain it in five-year-old language. If you can understand it, even better.

Mike Downer: So that I can keep my job, I’m just going to go ahead and move to number five.

Number five was measuring clicks and conversions only, and not tracking citations and conversions. What do you mean?

Kevin Wosmansky: Obviously, if your business is doing any type of digital marketing or advertising, the top-of-funnel metrics are always clicks, click-through rate, ads served, and impressions. Those are important. You need to track them, and we track them.

But as another top-line metric today, we also need to track citations, specifically citation conversions.

What I mean by this is: where is your business being cited?

The mistake is businesses accepting monthly reports that only focus on impressions and organic clicks.

Here’s the reality, Mike: as AI search answers users’ queries directly on a search page, traditional organic clicks are going to start decreasing. We know this. Highly qualified AI referral traffic goes up.

The takeaway is that you should still track website activity, clicks, conversions, and things coming into your website. If you’re doing paid advertising, we absolutely want to look at click-through rates and all of that.

But we also need to start tracking citations. We need to look at things as simple as your NAP: your name, address, and phone number. Is your business listed and cited in all of the traditional business directories? Is it starting to appear in other top-of-funnel directories?

When you think about the conversion funnel, it has shifted.

When a user asks a large language model for a solution, the LLM is going to provide a brand name and a citation link. When the user clicks through, they’re ready to buy.

So if the agency you’re talking to isn’t tracking these AI citations and share of voice, then they’re leaving out a massively important metric in terms of things they should be tracking for your business.

Mike Downer: Okay, that actually did make sense. Again, you did a great job of making it digestible, as you said.

Kevin Wosmansky: These come to us from things that we hear from prospective business partners and existing clients. That’s why it’s fun to talk about this. Everything I’m talking about is real-world stuff. This is what I see on a day-to-day basis.

Mike Downer: Yeah, 100%.

So we’ve gone over the five mistakes businesses make when hiring an agency. I think you’ve got this pretty well covered, and we know that anytime you deal with JAR Consulting, it won’t be a mistake.

Kevin Wosmansky: I’m saying this because if anybody out there is watching this, I want them to know the top five mistakes we see businesses make when they hire a marketing agency in this new AI world.

There are five specific things to look at: keyword stuffing; GEO, when they just list it as a service but can’t define the tech stack; siloed web development; lack of third-party brand presence, or not focusing on your entire digital footprint; and measuring clicks and conversions only, without tracking citations.

When you’re looking at hiring a marketing agency, I think those are five really important, digestible things you can take a look at throughout that process. I think that will help someone make a really informed decision.

Mike Downer: I agree. I’m going to lead us into our next episode.

If you’re looking for a marketing agency, you probably do what everybody else does and look at Google reviews. So maybe we should talk about that next time.

Kevin Wosmansky: That’s a great one, Mike. Google reviews and online reviews would be a great topic to talk about here.

Mike Downer: Let’s try that next time because I want to learn a little bit more about that, if that’s okay with you.

Kevin Wosmansky: I love it. I appreciate you being on today, Mike.

Mike Downer: Thank you, man. Thanks, Kevin. Take care.