Episode Transcript
[00:00:01] Speaker A: Hello, I'm Jeremy Rivera, your unscripted podcast host. I'm here with David, who's going to introduce himself, his company, and tell us why we should trust him.
[00:00:14] Speaker B: Wow. Okay. Quite an intro.
Thanks, Jeremy, for having me on. So, yeah, I'm David Bokweri, founder of CFGrass.
Why you should trust me, that's.
That's a tricky one. Right? That's the ultimate human question.
Look, I think maybe don't trust me. Trust what other people have said. We'll leave it at that.
[00:00:35] Speaker A: All right, well, what other have other people said and what are they saying it about?
[00:00:44] Speaker B: So I think they're saying good things. Right. About Grass.
Ultimately, I think we're here to help people. Right. Do better with Google Ads. We can go into what Grass does in a bit, but, yeah, generally the people are pretty positive that we are significantly improving what they were doing before versus now what they're doing with Grass.
[00:01:06] Speaker A: All right, so do you have a background then, of working in the ad space online?
What is your background and history, some of the experience that you're bringing to the table to this adventure, just to let people know in general. And let people know in general, like the mission statement of your company as that context, and then kind of plug in a little bit of your history.
[00:01:35] Speaker B: Sure, yes. I've come into the industry from probably a very different point than a lot of people might have. Right. Given what we're doing.
So for context, I used to work in private equity and venture capital.
I've been programming since I was quite young, 13, 14, I guess, which helps.
But what led me to start with Google Ads was whilst I was working in private equity, in the evenings, I was running my own dropshipping and e commerce stores. Right. And that would take up my entire evenings, entire weekends.
And it was through doing that that I got pretty good, I think, with Google Ads, but also started to figure out that we could maybe do things a bit better.
And that's how Grass was born from that, basically.
[00:02:21] Speaker A: So what's been the primary approach?
Because I've. I've run Google Ads for, in different levels. I've audited Google Ads setups from small businesses struggling to actually, you know, one guy, poor guy, he was doing 4D ultrasound and he was spending as much on his ads to get an appointment through that system as the actual 4D system was actually costing.
Of course, that's because he actually called in to Google Ads and had them help him set up his account, which was horribly configured. It didn't work at all. So coming from the ground up, coming from the outside.
Tell us about the approach that you've. You have with your company to make that a better experience and a better outcome at the end.
[00:03:22] Speaker B: Yeah. So Grass is quite different. Right. So as far as I know, right now we are the first people to have genuine, fully autonomous campaigns end to end.
[00:03:36] Speaker A: Right.
[00:03:36] Speaker B: So there's a lot of tools out there, a lot of things that help with this sort of thing.
Recommending keywords, negative keywords, giving you good ad copy.
We are now essentially replacing entire Google Ads teams. Right. And they just connect the campaign to Grass and it runs itself 24. 7 and measurably better than what a human can do.
[00:03:56] Speaker A: Right.
[00:03:57] Speaker B: Why? Because it's ingesting more, more data points.
[00:03:59] Speaker A: Right.
[00:04:00] Speaker B: We did not start off like that. Right. There's. It took a long, long time, Right. I want to say probably close to a year to get to the point of full autonomy. Grass started off with much more humble beginnings. It was like a little tool. It helps you do certain things. It gave you dynamic landing pages.
But Grass effectively evolved from being a tool that people would use and come in every day to improve their performance. Hopefully to now no one uses Grass.
So we actually get alerts if people are logging into their account on Grass and that will piss us off. It is not something that needs to be used anymore.
It's been a long, long, long road without doing all the stuff we did back in v1. It would be impossible to be where we're at because doing this fully autonomously and beating what a human can do is technically an extremely complicated challenge.
But yeah, we effectively got to that point now.
[00:04:50] Speaker A: So just to kind of lay it out, you've set up internal agentix systems to handle all of the individual pieces for a Google Ad campaign. Is that both on the search side and on the display side.
And what does the setup and decision making process of what it needs to look like to have them go do this? Like, what is the input you're giving to this system to get it to do what it needs to do? If that question makes sense. Yeah.
[00:05:26] Speaker B: So it's our campaign types that we have now are certain shopping PMX is coming later this month. We've left that last because PMAX is Google's own black box. They like to obfuscate as much as possible from you.
Yeah, you kind of summed it up.
[00:05:41] Speaker A: Right.
[00:05:41] Speaker B: So Caress is a collection of hundreds of different agents all talking to each other simultaneously and reading in live data from the account and the campaign to make its decision.
[00:05:50] Speaker A: Right.
[00:05:53] Speaker B: A human can Maybe look at 10 or 20 data points before doing something before taking an action.
Grass is analyzing hundreds of thousands per hour, which is why it has that edge. Right? That's one of the things. Also there's dynamic landing pages, which grass does, that's another topic. Which increase conversion rates.
But yeah, you're correct, it's effectively a logistics system that is working on that campaign 24 7.
[00:06:19] Speaker A: So is the objective for this to conquer Google Ads and offer an end to end where you as a small business owner, let's say you're, you know, trying to grow your Calgary SEO business and you want to run ads to do it. You know, you've got your SEO side of things covered, but you want to grow and get leads out of that. So you hop onto there and what does that onboarding process look like? What's the load on your site on the front end to kind of train or give the information needed to have it do what it do?
[00:07:03] Speaker B: Yeah. So since we've had full autonomy, which has now been live for about three months, three or four months, there's two types of people use grass. One is people who've just fired their agency, right. Maybe they were paying them like 5, 10k a month and they plug everything into grass for 1k a month.
Or two, it's agencies themselves plugging their own customer accounts into grass and not telling their end clients they're using grass. So they're just plugging their client accounts, letting grass do everything, collecting their retainer and charging the difference. Right. So those are our two types of customers.
So we, you know, a lot of Google Ads agencies have a lot of big crazy onboarding processes, set up fees, contract lock ins, that sort of thing. That's not how we operate.
The process is effectively you create an account, you connect your Google Ads account and then you decide one of two things. Either one, you have some running campaigns, you want grass to just get supercharged on. One click of a button, it takes that campaign in. It's now in the grass management. Or two, you let grass create a fresh campaign from scratch, fully optimized, basically from day zero. Right. So are you typically 20% of the usage is people creating stuff from scratch. 80% is people just supercharging and letting grass do its thing.
[00:08:18] Speaker A: I'm curious how you handle some of the rough edges within the ad system. I know I've run into my own fair share of difficulties. For example, if you're setting a budget for a particular campaign, Google can by its tos, spend up to twice as much as that designated daily budget.
So are there processes that you've had to invent to try to keep the spending within the correct spectrum, despite the fact that Google, you can set a budget, but the budget that actually gets spent within that may not accurately reflect. So on a manual side, you know, one of the tasks of a human is monitoring that campaign, its effectiveness, but also adjusting throughout the course of the month. If you actually want to spend specific dollar amounts, then you can't actually just designate that specific dollar amount to be the campaign spend limit. There are other things that you have to adjust for that. So has that been a challenge that you've learned how to overcome?
[00:09:29] Speaker B: Yeah. You mentioned two times there, we've seen it do up to four times in a day and they especially do that with sometimes newer accounts.
Yeah, look, that's Google for you, right? It's sort of pay to play as part of the game.
[00:09:45] Speaker A: Yes.
[00:09:45] Speaker B: Grass does have processes in place for that. Right. So for example, it might see one particular ad group is spending a lot, right. For some reason, but with no conversions, automatically it's going to pause that ad group.
So it's basically there. It will just pause the ad groups, pause the ad. Or it might say, hey, you know what it did get? It overspent. It did get maybe one conversion, but it should have got maybe two or three at this level of spend. So what Grass will then do is write brand new ad copy, brand new landing page and just deploy that automatically with the idea of trying to get you more conversion. So it's case by case. Right. There's a lot of rough edges in Google Ads. Absolutely. It's. This is why it's taken so long to get to full autonomy. It's really, really hard.
But yeah, effectively Grass tries to take into account all possible things that can go wrong. And there's a lot.
[00:10:35] Speaker A: Did you say, and I'm just curious, you said that it generates a landing page to sync up with each unique ad that you're writing.
That's dope. I just could say that that's dope.
I have a couple of questions on this SaaS side of things.
How are you approaching this business to gain visibility for it? I'm just curious what your marketing methodology is. Are you using your own dog food and using your own platform to run ads to gain visibility? Are you looking to network, find white label partners?
What's your game plan for growing your SaaS business?
[00:11:30] Speaker B: Yeah. So the short answer, Jeremy, is we've done zero. Okay. All our customers, 80% of them come from. I guess LLM SEO. Just recommending grass is what they're looking for. The other 20% are customer referrals. I haven't spent a dime in marketing so far. Right. Because everything, all my resource, everything has gone into building the absolute best thing that exists currently.
[00:11:52] Speaker A: Right.
[00:11:55] Speaker B: We've had the opposite problem, which is the last two months particularly we have been effectively unable to handle the organic demand that's just been being generated. So I'm behind the curve on hiring. So I now need to hire more people which we're trying to do ASAP to handle the demand. Once the team is in a better place. Right. And we're not working 18 hour days, seven days a week, we will then start to do, I guess do marketing for ourselves.
I mean, yeah, the obvious thing we're going to do is run grass on grass.
That's going to come asap. We just need to hire more first.
[00:12:31] Speaker A: What's next on your product marketing list that you know that you're going to be tackling? What's that?
[00:12:41] Speaker B: Yeah, good question. Look, I think about that a lot. Right. We still have a lot to do in Google Ads directly. Okay.
Uh, it's still in my mind day zero. Uh, we need to consistently keep improving our own models that we've trained to basically be like the top standard, which I think we are already. But short answer is number one, bring more campaign types, continue to improve models and that already is a lot of work technically, like a lot.
Uh, then who knows, right? Uh, obviously for now we've focused on Google. Who knows where, where we'll go. Right. Um, maybe look at the organic side, maybe look at other ad platforms.
We'll have to see. Right. We'll talk to our existing user base, see what fits there. I have some ideas, some. But also we're in a space that's evolving so rapidly that what I think now in three months might just be there's no point anymore. So we'll see
[00:13:37] Speaker A: when it comes to different verticals or niches that are running their ads what have been some kind of the learning curves of, you know, hey, we need to address this specific challenge. We need to do this differently for this niche or this article or type of company what have been some of the learnings on that side to set up a, you know, because saying we're going to do Google Ads. Google Ads can address all kinds of different business models from service to leads to, you know, high end products to E commerce.
Talk about some of those unique challenges and niches and industries.
[00:14:23] Speaker B: Yeah, it's been so Hard. It's been so hard. I think, I think, I think we're past those, those humps now. But yeah, we obviously cater to all possible niches. So we have, and not just that, but geography. So we have niches in countries I can't even name running on grass currently, you're right. Legion, E, everything, right?
Each niche has a different way that works well with Google Ads, right? So actually some people doing like a skag structure might work really well in some cases. We all support that.
Sometimes broad match works really well for a particular sort of like keyword, let's say dimension. Sometimes it's trash. There's so many possible things right, that could go wrong and can go wrong. The short answer is we've kept adapting and evolving with the users that have come to us, right? So now we're like very well covered where we've seen all possible campaign types of what could go wrong and what could go right. Some, we have some niches that are like, you know, kind of in the, like the tricky area. Right.
We have to adapt to that.
So yeah, it's, it's a constant learning curve. But effectively, if we hadn't had all Those users from V1 back when grass was just like a little tool, we would have never have solved full autonomy for all sorts of niches and campaign types. It wouldn't have happened.
[00:15:41] Speaker A: So what about that process of learning that, you know, taking the leap from, hey, we have an LLM based tool, there are these AI tools and now there are these agents capabilities.
How did you connect those dots as far as a solution and you know, problem solve? Okay, here's how far we can go with this agentic approach.
[00:16:14] Speaker B: Good question. So like when I first started Grass, I didn't, I never imagined ever we would be where we're at now ever. I was just building like a simple little tool. It may be recommended keywords, negated certain keywords.
We did the landing page stuff. Like it was quite basic, right. And we had a lot of like really small spenders working with us. I never imagined ever we would get to full autonomy.
It's actually, it's funny, it's the smaller spenders who didn't understand a lot about Google Ads, who were kind of like, huh, isn't this supposed to do everything for me? And it kind of got us thinking like, all right, is this even possible? Like, is this like, how could we do this but actually ensure it's doing correct things that are making sure that your rest is doing this and your CPA is doing this like Taking action on a campaign is one thing, anyone can just set that up. It's making sure you're doing the correct actions around the clock better than a human might. Really tough.
So the short answer is it came from a lot of early users who didn't know much about Google Ads, who kind of questioned what was possible and with them we effectively were able to get to where we're at now.
[00:17:19] Speaker A: So I'm curious because you know, on, on the front end side, you know, I use Claude heavily and any different AI tool, you know, has its level of hallucination, you know, has its level of false positive answers like that, it gives back.
What do you think has been the source of that and how have you approached agentic using data, executing something very real world, very real money, where a hallucination is, let's say very deadly or very potentially dangerous. So how have you addressed that? Elephant in the room?
[00:18:02] Speaker B: Great question. So two things. One, custom trained models, which again would not have been possible like fine tuning our models, which would not be possible without our big user base that we collected in D1. Right. So like without this I don't know where we'd be.
Two, context matters. So business context, Right. If you can feed as much possible information about the business or the particular offer, what they're trying to achieve, the brand, all these sorts of things, the results you're going to get actually are drastically better. Right.
So yeah, short answer is custom models and feeding in the correct context, that matters. And also knowing like what to ignore, not just dumping somebody's website and saying, hey, there's a lot of other things that you have to be careful with.
[00:18:53] Speaker A: That makes sense.
So with that in mind, I'm curious, is that something that you have to address like in your marketing materials that we, hey, we have this process to make sure, you know, AI is running it. Right. Or you know, how do you ensure that people are communicating that communicated that the process you have is, you know, safe for their money?
[00:19:28] Speaker B: Yeah, look, it's hard.
But you know what, you're surprising, I'm surprised by the number of people that are very excited to have something fully autonomous so they don't have to check.
And that's obviously smaller businesses will appreciate that. Right. Like the alternative is maybe an agency or a freelancer which is hit or miss and obviously very expensive.
But also even at the much, much bigger end, right. We have people spending seven figures a month through grass campaigns. Like that's just one customer, they're super excited about it. Right. Even before we even got on A demo and showed them about this stuff. They were just keen right from the off. So it's not something that we have to address too much.
Obviously AI is getting bigger. People are knowing more about it. That builds confidence. Most people use AI tools in their day to day life.
The models, when you go to like a ChatGPT or whatever, seem to be getting better. So that also I guess gives people some confidence. Uh, but yeah, I mean had this would, would Grass have existed two, three years ago? Yeah, maybe it would have been a different, a different question.
[00:20:30] Speaker A: A little bit of a different question.
Um, here's a soapbox. I'm going to put it down in front of you and you can step on that soapbox and what's your hot take? What do you think that people should know about anything? About anything that you're passionate about? What you wish the world was more passionate.
[00:20:55] Speaker B: Whoa, okay. I don't know about, I can.
If we're keeping it in the context of, of marketing and Grass, I'll say this right, I know, I don't mean to be disrespectful to anybody or inflammatory. I would, I would just say truthfully, like what we have seen in this, in this journey we've been on because it's been, it's been quite the ride when we take over accounts from other agencies, right? Or maybe some, some freelance person they found on upwork.
It's quite, it's quite drastic what we see. It's really drastic.
I want to say 90%, 95% of those scenarios. What we see in those accounts is really tough, right? Just things horribly configured. Business is losing money on Google Ads. They have absolutely no idea why they've been charged a fortune, right? So they correlate the price they've been charged with the quality of work that's been delivered, right?
They think, oh, Google Ads are not for me. It's just like a dead horse, like what's the point? And that's really tough to see. We've seen this countless times where we have to then take over that account completely build everything from scratch again and then a month later, like, okay, why wasn't Grass around like a year ago?
We've seen this so many times in this space. We have agencies using us, by the way, and we have like a great relationship with them. Like a lot of them are great.
But I think in five years time the agency landscape will look very, very different because you can't, I think, justify charging, you know, 15% of, 15% of ad spend plus this like with these Lock ins for a long time because things like grow as are just going to I think keep, keep making the market evolve.
See I feel quite strongly about that because I've seen a lot of like particularly smaller business owners get really badly burnt by multiple agencies consecutively and I don't like to see that. Personally
[00:22:48] Speaker A: I think that's a fantastic and fair hot take.
I see that often too. You know, taking over on SEO accounts elsewhere and just saying, good lord, what is going on here? Like why is this set up this way? This is a terrible decision.
I've done audits on a number of Google Ads accounts of like hey, you're doing the SEO? Do you know anything about ads? Can you check this out? I'm like, yeah. Oh my goodness. Hawaii is. You haven't even set a geographic boundary. Like you're advertising in Omaha when you're in Southern California, like really basic stuff. So I could definitely see misconfiguration on a wide basis being a challenge with, with, with Google Ads.
That being said, what's, what's some resources or information or if you had anybody check out something out after this conversation about what you're doing in the ad space, where can they find you? Find white paper or some information about what's happening in Google Ads space?
[00:24:08] Speaker B: Sure, I guess our website, right? So growass AI.
I think we have a blog on there. I don't keep too much of a track that sometimes publishes some stuff I think people sometimes tell when they read it. So maybe check the blog.
Otherwise look, feel free to get in touch. I think actually that LLMs are a good source if you want to learn about Google Ads and just like start to tweak YouTube videos. We might start doing YouTube videos. But look, there's a ton of free content out there. Like if you're not running Google Ads, there's a ton of free content where you can just sort of get going.
And obviously if you have any questions then feel free to get in touch with. Gross.
[00:24:43] Speaker A: I guess.
Fantastic. I appreciate your time Dave.
[00:24:47] Speaker B: Thank you Jeremy.
