Summary
This video presents a detailed case study and key lessons derived from working with a dental client who faced severe financial difficulties due to over-reliance on advertising for patient acquisition and reactive decision-making, particularly influenced by AI-generated strategies.
Case Overview
- The client depended heavily on digital advertising, with about 95% of patients coming from platforms like Google and Meta.
- Advertising results naturally fluctuate over time due to platform algorithms and budget constraints.
- The client reacted to short-term dips in performance by frequently changing advertising strategies, often based on recommendations generated by AI tools such as ChatGPT.
- These frequent and drastic changes wiped out accumulated platform learning data, causing a reset in ad performance and ultimately leading to significant revenue loss.
- The client was unable to pay the agency fees and had to cut costs, including layoffs.
Timeline of Key Events
| Timeframe | Event Description |
|---|---|
| Initial period | Client highly dependent on advertising; stable but fluctuating results. |
| Reactive phase | Client reacts to weekly dips by implementing AI-suggested strategies leading to worse outcomes. |
| Major strategy shift (1 month ago) | Client pushes for a large-scale AI-driven strategy overhaul, wiping out platform learning data. |
| Result | Advertising results collapse, client faces financial distress and cuts costs, including agency fees. |
| Agency response | Agency offers free support until client recovers financially. |
Key Insights and Lessons
1. Avoid Reactivity in Strategy
- Reacting impulsively to short-term fluctuations, especially by altering the core strategy rather than tactics, is detrimental.
- The hierarchy: Goal → Strategy → Tactics; changes should first be tested at the tactical level before revising strategy.
- Strategy includes fundamental positioning decisions (e.g., quality vs. cost leadership) that should not be shifted frequently.
2. Build Intangible Assets
- Success should not rely solely on advertising; instead, businesses should build assets such as:
- Trust
- Reputation
- Relationships
- Networks or membership lists
- These assets reduce vulnerability to advertising variance and provide more stable, long-term growth.
- Asset-building allows businesses to endure fluctuations in paid advertising performance without drastic disruptions.
3. Exercise Caution When Using AI for Strategy
- AI tools like ChatGPT often provide:
- Outdated or generic recommendations lacking full context.
- Responses influenced by confirmation bias due to training on user feedback favoring agreeable answers.
- Average or status quo solutions which may lag behind current best practices.
- AI can be useful for challenging assumptions or generating alternative perspectives but should not replace expert judgment or strategic thinking.
- Effective AI use requires knowing what questions to ask and when to critically evaluate AI output—a catch-22 for many users.
Additional Observations on Advertising
- Advertising effectiveness is declining due to:
- Increased competition (auction system dynamics).
- Privacy regulations (e.g., HIPAA), which reduce available user data.
- This leads to higher costs, lower data quality, and slower optimization cycles.
- Many dental offices depend heavily on advertising (80-95% of new patients), which is increasingly risky.
- Diversifying patient acquisition methods—especially through word of mouth—is critical.
- Word of mouth acts as a stable asset and diversifier, creating patient evangelists who organically promote the practice.
- Reducing reliance on paid channels can increase stability and reduce vulnerability to platform changes.
Summary Table: Comparison of Advertising vs. Asset-Based Growth
| Aspect | Advertising-Dependent Growth | Asset-Based Growth (e.g., Word of Mouth) |
|---|---|---|
| Stability | Low; subject to platform variance and costs | High; built on trust and relationships |
| Cost | Increasing due to competition and privacy laws | Lower; organic and referral-driven |
| Control | Limited; dependent on external platforms | High; owned by business |
| Scalability | Variable; constrained by budgets and auction systems | Sustainable; grows with network and reputation |
| Risk | High; sudden drops in performance can be damaging | Low; diversified patient sources reduce risk |
Core Conclusions
- Dependence on paid advertising without building intangible assets creates vulnerability to market fluctuations and platform changes.
- Reacting impulsively to short-term advertising dips by changing strategy can erase valuable performance data and damage results.
- AI tools are helpful but should be used cautiously and supplemented by human expertise, especially in strategic decisions.
- Building assets like trust, reputation, and networks provides long-term stability and reduces dependency on costly and volatile advertising channels.
- Diversifying patient acquisition via word of mouth is a critical future-proofing strategy for dental practices and similar businesses.
Keywords
- Advertising variance
- Strategic reactivity
- Intangible assets
- Word of mouth marketing
- AI limitations (ChatGPT)
- Privacy impact on data
- Patient acquisition diversification
- Confirmation bias in AI
- Platform learning data reset
This analysis provides a comprehensive understanding of the importance of strategic patience, asset-building, and critical AI use in managing advertising-dependent businesses, particularly dental practices.
00:00:00
Over the weekend, a client reached out telling me that he couldn’t pay our fee anymore, that he had essentially run out of money. He was in the red, and so he was cutting costs everywhere he could. Um, he had to let go of an employee, and it was really rough. And so, we worked out a uh well, not really an official plan or anything, but basically, we’re helping him out for free until he gets back on his feet and he’s able to afford us again. Um, but there is a couple of lessons to take away here and let me
00:00:34
give you a bit of context before we get into the lessons. The context was that his office without exaggeration probably gets 95% of their patients from our advertising. So he is very dependent on our work and also the uh effectiveness of platforms like Google and Meta for bringing him new patients. All right. And when you run advertising, unless you have a significant budget that local dental offices usually never do, you will run into inevitably periods of variance. Times where results are very good and amazing and times
00:01:21
where results aren’t very good. And this is just natural. So the approach that we as advertisers take is to look at the average across a month or so and use that to gauge our improvements to the advertising strategy. But because he was so dependent and because he had created a ecosystem that was so vulnerable to variance that when there was that dip in a week, he would become very reactive and reach out to us um with ideas. ideas he had got by leveraging chat GPT, having overnight sessions with it and
00:02:05
developing a new game plan that he would present to us and many of the times or and and he would just email us straight the um chat GPT responses. So I was basically reading what he was going through with CHAGPT and um yeah needless to say a lot of the times I push back saying that this isn’t how we should do things. Um, this is straight up a bad recommendation that sometimes ChachiPT would tell him. Um, but because of that reactivity, we would implement a couple of things and we would show him and it would make the
00:02:44
results even worse. So, we would revert it back to how it was before. time would pass, another dip, more ideas and so on and so forth until finally about a month ago we experienced another dip. And this time he requested this massive list, this massive strategy for how they could eliminate variance and get more patients than ever ideas by Chad GPT. And even after push back um we ended up implementing them. long story short and it completely wiped out results. So what had happened was advertising platforms they leverage past
00:03:27
performance data in order to understand where the advertisements work best who it’s best shown to what kind of behavior users need to have and when it comes into contact with them the advertisement it works best with those kinds of people. All that data was wiped away because of the severity of the changes that he wanted us to implement. And so the platforms were like, “Oh, this is a completely new objective. Therefore, this past data that we had collected through years of working with this guy
00:04:04
is kind of useless. And so therefore we as platform should start learning and regenerating better data for this specific objective. And so all of the past performance kind of got wiped out and now we were starting from ground zero. And so uh predictably his results were wiped out as well and hence why he wasn’t able to afford paying us. All right, couple of takeaways I think we can have from this and I think there’s three main ones and the first is to not be reactive just not in terms of advertising definitely don’t
00:04:43
be reactive there but just in terms of strategy as well that if we think back to the hierarchy of goals being at the top strategy being our approach to accomplishing that goal and tactics the implementation of that strategy When we see that our work and our efforts aren’t bringing the results and the goal that we seek to accomplish that we shouldn’t immediately go back to the strategy and start looking at what is going wrong and not necessarily even look at the tactics, but we should we
00:05:19
should at least start with the tactics first. But what he was doing was going back and being reactive to how his strategy would change. And that is not the right approach in terms of advertising. It’s also not the right approach in terms of marketing in general because strategy encompasses things like how you seek to show up, the positioning that you have in the market. So are you going to be someone who is leading with quality? Are you going to lead be leading with cost effectiveness? All these different ideas. And if we’re
00:05:48
just changing that for each patient, of course, it’s not going to build into anything greater. And so this leads into the second idea of building assets. And I’m not talking about assets like equipment or technology or any of those stuff. I’m talking about assets which are things you can build, you can things that you can affect in terms of development that will help you achieve your goal as time goes on. So things like relationships, things like trust, reputation, these are intangible things
00:06:21
that are assets. If you have more trust with the people, you will it will help you achieve your goal. If you have more better relationships, if you have a larger network, if you have even like a membership plan, a list of people is something t tangible that is an asset. And it will help your level of reactivity because now you’re less dependent on advertising because you’ve built these assets which are also turnurning out your like results towards the goal that you seek to accomplish. So
00:06:52
you can get through weeks and even months of variance. Um so yeah build assets, right? And the third and final takeaway is to really be careful when you’re leveraging AI. And this is probably a sentiment that you’re hearing more and more now, but it’s also very common that when I go into consultations or meetings with clients or prospective clients that they will come having done a lot of back and forth with things like ChachiPT, with Claude, and I’ve noticed that a lot of the the
00:07:25
recommendations they’re receiving are outdated or ineffective or they miss a huge portion of the context that would have completely steered that recommendation in a different direction. But the problem is with AI is one, it’s trained on user feedback. You know, when you ask a question, it gives you a response and asks which one you like more. People will choose things that reassure them, that give them confirmation bias. And so, the method that LLMs are being trained right now is leading to a lot of
00:07:59
false confidence, a lot of confirmation bias. And so be very skeptical with the things that it’s recommending. And also that when we’re looking at how these LLMs work, they tend to get the average. So if you seek an answer to a question and there’s a whole range of answers, it will get you the one with the most um average, like the most cited, the most um this the status quo approach to doing stuff. And when you do that, typically when it comes to strategy, the status quo is always lagging behind what is
00:08:39
actually effective and working. And so therefore, this whole idea of leveraging chat GPT to determine strategy is backwards. It it’s going to have you show up as a follower. It’s going to have you doing things that are ineffective and outdated. Um, so be very careful with leveraging AI unless you know what to ask it, which is a catch22 paradox because you wouldn’t be asking it if you knew what the answer already was. But in order to arrive at the solution, you have to understand what to ask it and you also have to
00:09:12
understand where you should challenge it and push back on it, which if you don’t know the solution, you can’t do it, right? So it’s a very complex catch 22. That said, I would leverage AI. I would recommend you leverage AI by having it challenge your assumptions by having you assess in a different way to to give you a different perspective on things. But beyond that, be very careful with leveraging AI. All right. So to wrap it up, a client had a very um well detrimental results for his
00:09:44
practice for many reasons, but primarily because of his reactivity which was caused because he wasn’t able to incorporate assets into the entire ecosystem and his business model and because of his dependency with advertising and also when seeking an answer he was leveraging chat GPT which was giving him outdated strategy. And another quick note on advertising, I’ve also noticed more and more dentists who are starting to leverage advertising. And as more and more people start leveraging this regardless of uh the
00:10:23
number of people leveraging it, but advertising itself is just decreasing in its effectiveness. Um not just in terms of cost effectiveness, but also in terms of the results that you will see. lots of factors, but primarily because our ability to collect data is being throttled. Like privacy protections, HIPPO laws make it so that we get less data per activity with the user. And like I mentioned, data is how these platforms understand how to improve your ads, how these algorithms understand how to improve your ads. So with reduced
00:10:57
data, it’s going to perform worse. It’s going to get more costly for you to get that data and it’s going to take more time before the results improve. Now, with more people coming in and leveraging advertising again, the profit margins now diminish because it’s an auction system. Auction system. More people in that auction, the less likely that you’re going to come away with it a skew in value, right? All these factors make it so that advertising is in a really weird spot now.
00:11:29
contradictory to that effect. A lot of dental offices rely on advertising. Like I say 95% of his patients came from advertising. It’s not uncommon that some offices rely on 80% or 90% of their advertising or other new patients on advertising. And I think it’s more important than ever before in order to start understanding how to get away from that to build word of mouth as an asset to have that become the new new new patient acquisition method for your office. So eventually when you build that asset
00:12:07
large enough, when you have other people who are spreading the word about you and doing the marketing for you, you completely remove advertising and dependency on other third party systems for your um acquisition, right? Because right now what looks like it’s it’s an agency or it’s Google or it’s Meta who is responsible for your new patient acquisition. No wonder there’s so much variance. You have all your eggs in one or two or three baskets. But when we’re able to leverage word of mouth, another
00:12:39
factor is that it’s diversification. Now, every patient who comes in has the potential to become an evangelist, a salesperson, someone who’s spreading awareness about you, right? We’re now able to diversify our efforts and you get way more stability with that. All right. So, these are just some ideas and some takeaways that I think we should really consider. Um, if you’re interested in learning how to build word of mouth, I have a free course. Check it out in the description. Um, other than
00:13:06
that, hope you guys like the video and I’ll see you in the next