AI Ads Will Be The Next Land Grab (Here's How I'd Play It)

0:00I got into Google AdWords and Facebook ads early enough that growing my business felt almost unfair. Like I had access to a lever that my competitors didn't know existed. Both times that advantage disappeared within a couple years as everyone else caught on and the cost normalized. Now that the AI chat platforms are introducing ads, I think we're at the beginning of another one of those moments and the opportunity here is real. The hard part is timing. If you move too early, you're likely to waste money. If you move too late, you miss out on what could be a great marketing channel, at least while ad costs are

0:32low. So, today I'm sharing a framework to help you get the timing right. This isn't a step-by-step tutorial, it's how to think about this whether you're just getting started or already scaling. And stick around to the end because after investing in and advising hundreds of SaaS companies, I've seen a few traps that aren't obvious until it's too late. I don't want you to make the same mistakes. First, I want to talk about why this is an opportunity. So, whenever a new ad platform launches that has scale on the supply side, meaning they have a lot of ad space to sell, there isn't enough demand to drive up the

1:02prices of the ads yet. And as these companies like Facebook and Google and Instagram, as they've tried to develop their ad tech and and honed it to make it a really efficient auction, as they're doing that, they're testing things out and a lot of clicks are cheap. Back in the early days, Google AdWords had five and 10-cent clicks if I recall. I was getting Facebook clicks in 20, I believe it was 2012, for like 20 cents, 25 cents consistently targeting my ideal customer profile. And that's how I was able to quickly grow one of the SaaS companies I was running at the time. So, when these new ad networks

1:37launch, there aren't a bunch of books and content about it. There the playbook isn't saturated yet and that creates opportunities for founders who are willing to take some risk. So, with these AI companies burning cash, at some point they're going to need to become profitable and make money. And obviously, they're charging subscriptions, but ads are the obvious

1:56path. It's what Facebook and Google did. It's how television has been monetized for decades. It's how radio was monetized before that. Any type of mass media or widely used free tool of any kind built for consumers is usually monetized with ads. And of course, I'm the person who will happily pay for a subscription to skip ads, but most people won't. They'll take the free version with the ads. That's the opportunity. So in this video, I'm going to walk you through two playbooks. One is if you have a new product, and the other is for an already successful product that you're trying to scale. So if you have a new product, you probably

2:28have low budget and a lot of learning to do. In this, your mindset should be that you don't need the perfect channel. You really just need fast signal. You cannot afford brand-only experiments. You need to think like a conversion marketer, a direct response marketer where you see results from the clicks to a trial sign-up or to a purchase. So before you spend a dollar on ads, there are a few

2:51basic things that are really important. Having at least some understanding of your ideal customer profile can help. And this can be psychographic data like where are they, maybe what their role is at their job. It's going to depend a lot on the targeting that chat GPT allows you to have. But realistically, knowing your ideal customer profile and the more you know about them is going to be helpful. In addition, having one core use case or value proposition that you can articulate very simply in a few words is going to be important. We've only seen video with a mock-up screenshot of what an ad might look like, but it's very sparse on the text,

3:24right? So you're not going to have a paragraph to explain the value that you create or what you're going to do for your customers. So that's where you're going to need to figure out, how do I come up with something compelling, this use case or value proposition, and then maybe you have three or four that you want to split test if you have the budget. Third, you need a landing page that actually converts. Ideally, this is not your homepage. Usually, you want to copy your homepage if it is already a high-converting landing page or you've done the best you can to make it that, and then you change the headline on that

3:52page to match the headline of whichever ad you're running. Fourth, you need onboarding that activates people. Maybe that's manual. You can do things that don't scale for now if you're still early, but if folks are coming into your funnel and they are signing up for a trial or they're paying you and they're not getting onboarded, that's a problem. So if any of those are broken, ads are

4:11basically going to be wasted money. They're just going to amplify that problem. Something to think about is trying to define the smallest success metric that matters. So you're not going to get a positive return on your ad spend by day seven, right? You're not going to see that in some dashboard. Something real is more like qualified booked demos or activated trials from your ideal customer profile. And realistically, just getting traffic early on as if it's cheap enough is fine. That's where you start and then you see if people are converting. You can install tools like Hotjar on the landing page to see if people are getting confused by anything. But

4:45realistically, you're trying to take it kind of slow and ramp it up as you want to learn. With this in mind, you want to run learning tests. You're not scaling these things. So as I said earlier, maybe you have two to four offers or value propositions, 10 to 20 total variations of your creative, right? So you can use AI to generate images or whatever. Again, I don't know for sure if they're going to allow images in the ads, but you use AI to help you either write different ad copy or to create these images. And then you have one landing page per offer. So two to four

5:17offers, one landing page per offer. And it does not have to be completely custom to the offer. As I said, I used to back in the day just copy my homepage, change the headline, the H1, the H2, maybe something on the page, but oftentimes I was just trying to continue the conversation that had started in the prospect's head when they clicked on my ad. Then in terms of budget, obviously you want to cap your spend to protect

5:40your runway and commit to a time box. You know, this might be a month or two months. Now I like to think if I have the budget, I want to test ads for 3 months in general. And especially if there's really no playbook on how to run ads in AI yet, it is going to take me a few weeks just to ramp up and get started and see what's working. Now, if

5:58there's early success, it's really nice. If there's traffic and some book demos, it's really easy to tell. But if stuff isn't working, you just have to tweak and tweak and tweak. And I would get myself a minimum of a month of tweaking and learning to reevaluate. And then your stop rules where you're just going to cut it off is if you're not seeing any signal, right? You're getting no click-through rate. Your lead quality is abysmal. There's no activation. And this is when after a month, if if you haven't gotten any signal and you've tried and

6:24tried and tried, you pause. Maybe it's a month. Maybe it's 2 months, depending on the budget you have. But there really are these four things. It's your your offer or your value proposition that you're putting in the ad. It's the ad itself. It's the landing page and how you're collecting info there, whether it's to get a demo or to start a trial. And then it is your onboarding. One thing that applies no matter where you are in your journey is eventually the AI companies should figure out how to make their ad tech work. Now, Twitter and LinkedIn haven't really done that in my opinion, but we

6:54have seen companies like Google and Meta do it. So, I think the AI companies are going to figure it out. So, you probably don't want to just try this once and walk away. The models are moving fast and I expect the ad tech will also move as quickly. So, what doesn't work today might work in 6 months. And that's where I personally would keep testing. Now, if you're a more seasoned founder with an existing funnel, I want to talk about what you should do if you have an already successful product. This is where your goal is to exploit the new opportunity without breaking what already works. But you have a baseline

7:26cost to acquire a customer, lifetime value, payback. You have creative history. You probably already have a bunch of offers that you know work. Your advantage is that you have data, you have discipline, you have some distribution muscle, and you probably have more budget. So, in this case, I'd be taking more of a portfolio mindset. So, I'd allocate a fixed percentage of my entire marketing spend to experiments. You know, maybe that's 10%

7:49of everything I spend on marketing, 20%. And I would set clear gates for increasing or decreasing that allocation. This is an exploration or experimentation budget, and I'm not looking for return on ad spend right away. I'm just trying things. And this is where you're probably going to want to add a how did you hear about us? Ask

8:07them if they heard about you through AI. And if so, dig in. What were they trying to accomplish? What platform were they using? What did they ask ChatGPT or Claude or whoever it is? What made your website appear? Ask them if it was an ad. You know, you don't know if it's an ad or it's organic, and that's going to be tricky unless ChatGPT passes something in a you know, a URL parameter, which which hopefully they

8:28do. So, this is all free intelligence. It's more valuable than anything the platforms will tell you probably, having a conversation, assuming you're doing more high-touch sales. It's going to be important to have some type of attribution because as much as we want the click-through attribution to be perfect on 100%, there's just no chance it will be. Two other things to keep in mind is you do have an existing brand if you have an existing product. So, if you're using AI to generate creative at scale, someone needs to be monitoring that, the claims, the tone, the offers, to make sure it still sounds like your voice. So, someone needs to be

9:01protecting the brand. Your edge here as a successful entrepreneur isn't discovering the ads before anyone else cuz a bunch of people are going to be rushing into this, but it's being able to test systematically and to scale fast when your math works. You're a bit like a pioneer out in the Wild West, and you want to get out of head before the ad costs increase. In a minute, I'm going to talk about some of the traps that I've seen founders fall into when there's a new channel like this. But first, I want to tell you about Tiny Seeds SaaS Institute. For SaaS founders who are doing a million a year or more,

9:32who want community, coaching, and mentorship, they're not a great fit for our startup accelerator where we invest in early stage companies, but if you still want that community of like-minded founders and a roster of world-class mentors to help you break through your next plateau, you should head to SAS Institute.com and apply if you're a good fit. We have weekly calls, we have mastermind groups, we have paid coaches, and this is a premium offering and it's one of the best SAS coaching programs in the world. Now I want to tell you about a few risks and caveats and some mistakes that I've made. One thing to keep in mind, I've said it already,

10:06attribution will be messy. Tracking may not be very mature, so I would have multiple sources. Like I wouldn't trust any single source. You want UTM per ams if you can, you want whatever ChatGPT allows you to pass through, and you want to be asking customers. So whether you ask them in their dashboard manually, how did you hear about us, or you're asking them on sales demos. Another thing is platforms will change rules quickly. So what works today might not work in 3 months, and that's not a reason to avoid testing, but it's a reason to stay light and not over commit. One other thing to be aware of

10:39is not to trust platform case studies. So platforms are going to show you wins that make it look like an easy no-brainer, but they're selling ads. So trust your own numbers, not their cherry-picked wins, and not necessarily the attribution they have in their own ad dashboard. All of these are mistakes that I had made when I was trying out Google AdWords and Facebook ads in the early days. Two more things. One is to remember that they're not going to work for everyone. So if you have a really low annual contract value, you have a weak differentiation, or maybe if you have a tiny total addressable market,

11:10the math might not work. Be honest about whether your business is a fit. And finally, this is a pay attention and test moment. It's not a moment to bet the company. Sign up for advertising waitlists. I'd be doing that right now. Well, whatever AI provider you can think of, I believe Anthropic has said they're not going to have ads, so Claude is not a candidate. ChatGPT has announced they are, so that's certainly one I'd want to stay in the loop on. But be smart about it, right? You're not going to bet your runway on something that you're still being figured out. Try it, optimize it, and scale it as it makes sense. If

11:39you're thinking about ads on AI, you're likely trying to figure out marketing in general. In this next video, I talk about how I would approach content marketing. I still think there's real opportunity in SEO while the models are hungry for fresh content to train on. Check it out. Thanks for watching, and I'll see you next time.