Aidan

Welcome to episode number 73 of The Growth Booth. Thanks for tuning in today. Today we are talking about systematization and how you can streamline operations. I want to give you as many tips as I can that you can hopefully go away and apply to your own businesses, your own life. No one better in my mind to talk to about this than Bonnie Collins. She's one of the managers in my business, and she has single-handedly built operations for a whole raft of different projects that I work on, including this podcast right here. She orchestrated building a team for that, putting all the systems in place, and then building it up so much that she could then step aside and move on to another project. She's done the same thing with different blog websites that we run. She manages a team of people in our business who are located all over the world. She's sort of been somewhat of a linchpin in this idea of systematization.

So, Bonnie, thank you for taking some time out there. I've been looking forward to this call. I know you've got a lot to share with people. Where do we begin when we're thinking about this topic? 

Bonnie

Well, first of all, Aidan, thank you so much for having me. I've been excited about this podcast for ages and really was happy to give it over to some really amazing people in the end after we got the flywheel going. Yeah, I guess it's all about getting a flywheel going. For me, it's really to do with you've got a blank canvas in the beginning, sort of like a V-shaped funnel I have in my head. From this V-shaped funnel, you have all the creativity at the top, you've got all the research at the top, and then basically you refine it into a well-oiled machine that really becomes like a flywheel at the end.

Aidan

And for people that don't understand the term ‘flywheel’, what you're talking about is something that runs smoothly on autopilot, essentially.

Bonnie

Yes. Well, I guess it's not quite on autopilot, but other people take it over. With online marketing, there are lots of moving parts, but it's really quite easy to, once you know what the moving parts are, there's so much repetition that it's easy to create some sort of a pattern. That pattern creates a process, the process creates a system, and then from there, we really can see the patterns and the processes and systems and get people to just replicate and replicate what we're doing.

Aidan

So you mentioned this idea of a V funnel. Can you elaborate on that a little bit and share a bit more about what you mean there? 

Bonnie

Yeah, so I guess at the beginning, you're looking at this blank canvas and you're saying to yourself, “Well, what are the goals of my project and what's the best use of my time?” It's really a lot about mindset and planning. What if I succeed? What if we fail? What does success look like? What does the range look like, and how are we going to do this well, and possibly better than the best? The questions are, “Do you like the subject of the blog?” I personally liked all the blog subjects that we've been working with recently. It's important to like the subject because really to do this well, you'll be writing a lot about it. So if you're writing on and on and on about it, you're going to have to like the subject you're writing for. 

Aidan

So you're giving an example specifically there about starting out a new blog from scratch, which I think is a good one to move forward with. And that can be a bit of a theme, if you like, for this episode. So if we're sitting down and we say, “Okay, well, we want to create a new blog,” what next? How do we go about systematizing that? Do you try to break down what the key components are? Do you try to figure out what the daily operations are? What's your way of approaching this? 

Bonnie

Well, I love technology, but technology is really the vehicle for all of this. When I'm working with the site, I try to decide how niche the site is. If you have a very niche site, then what I mean by that, if you have one core subject that really doesn't have a lot to work around it, say you have a scientific subject, for example, and this scientific subject has one core topic that you keep coming back to and back to, then you need to make sure that this particular subject is well catered for in internal links within the blog, whereas if you have a broad subject like a gardening website, then you have lots of different topics within the gardening niche and within the outdoor niche. You could bring in all sorts of cooking as well into that sort of niche. That has a lot of subtopics, whereas the other more scientific website, it's a very sort of fine-tuned niche. 

Aidan

You're talking about how specific a niche is. In that case, we've got gardening at one end of the spectrum. Another one that comes to mind is photography. It's a very wide niche, but then maybe a much more narrower niche might be drone photography, and an even more narrow niche might be like, I don't know, military surveillance or photography or something like this. Are you saying that the narrower the scope of the project is? I think I missed the idea, yeah. 

Bonnie

The more internal linking you have.

Aidan

Is that because you're developing more content yourself, you're relying on your own content more? 

Bonnie

Exactly. You're relying on repetitive content because there are key areas of, for example, safety. Depending on what sort of scientific topic you're speaking about, there's only so much you can say about certain subtopics within that. That's all. 

Aidan

Yeah, got it. Okay, so one of the considerations then, when launching a blog is how wide is the scope of the topic going to be. But from a systematization standpoint, such a mouthful, that word. I was actually trying to think of a simpler word to use, but I couldn't think of a simpler one than systematization. So anyway, here it is. 

So in terms of systematizing a project, like setting up a new blog about drone photography, where do you actually start with that though? Because at one end of the spectrum, you've got it so that you can be doing 100% of everything yourself, from writing articles, from buying a domain name, to publishing content, to potentially building links and advertising and so on. At the other end of the spectrum, if you want to build out that blog and dozens more blogs, then you build a system. If people want to build a system, what would be that next or that first step that they would take? What would you suggest there? 

Bonnie

Okay, so I look at what would the finished product be. In terms of a blog, what sort of subject matter authority figure are we creating for the site? We're then looking to the actual technology. I've recently been working with AI and there is a huge element of “How does it work? What can it do for my audience? Is it the best tool? How much can it improve productivity of the team? Does it matter if AI itself predominates? What about plagiarism?” So all of these questions, what do I need to know about AI? And then you go on to the WordPress website, if we're doing WordPress, and what's the technology that we need to bring in with the WordPress site. The domains, so it's all the particular elements of the technology. 

Aidan

In terms of the technology, we've got WordPress as a platform, a place we're publishing, and it sounds like you want to immerse yourself in that and explore the limitations, what that can and can't do in an automated fashion, but the same with AI technology, what can we actually use this AI technology for. Why do you do a deep immersion into this? What are you trying to achieve by learning more about the AI technology? As an example here? 

Bonnie

Yeah, I guess we're trying to learn about the range of capabilities of the technology as a vehicle to support the project. Again, we are the experts, so it's really about exploring as much as we can, sort of on an academic level, really. And again, immersion. I also look at YouTube videos on 2X speed because my attention, as it were, is valuable for the project and for the team, so I need to do it as quickly as possible. The idea is that you don't want to get paralysis by analysis. Instead you want to gather as much as you can on a 360-degree level and then move forward armed with that information to the subject matter. 

Aidan

So deep dive into AI thinking about it from a content development, content creation standpoint mainly, understanding of WordPress because it's like the platform that we're going to be using to run a website. What next? What about team elements and how do you go about starting that side of it? 

Bonnie

It's about choosing the right people for the team. Interestingly enough, with a content team with AI, you really need people who are more process-driven than English majors these days because the technology can support really high level of writing. The people that I picked for the team are actually very strong in process. We did have a really high level writer in the beginning, but we just found that the AI was so great in terms of the volume that it produced with a certain level of quality, so went to GPT4 level and this produced a really high level of quality. 

Aidan

Perhaps in the past, if we roll back the time machine, a year ago, one of the key members in a team would have been the person who was writing the content. Nowadays, we're able to solve the content using AI tools, and not just AI tools, but a combination of AI tools and making sure the things that we're asking AI to do are the right kind of things. But in any case, we can solve the content using AI technology. The team focus is more around people who can create or follow a process rather than maybe have creative abilities. Not to say they're not creative, but creativity is probably less of an important skill. 

Bonnie

Yeah, and I do want to come back to the actual website creation itself. As you know, we have stellar people who can help create websites, that's more of a one and done sort of task within the whole project as a whole, but equally important to have high level website creation, website development, people who know what they're doing, doing it quickly. We've got a team that helped with that, but ongoing, we need the content creation wheel running. 

Aidan

So what kind of a framework have you come up with then for the kind of articles and content that we're publishing nowadays? What's the process there? 

Bonnie

The big learning element within AI really is prompt engineering and creation of the prompt to you really had to be refined over time. We've now created a prompt that we use that can be used across all our blogs. 

Aidan

Could you just explain to people who might not know what you mean by either prompt or prompt engineering? Like, what is a prompt? 

Bonnie

Well, a prompt is simply a question that you ask the AI. So whether it be Chat GPT, GPT4, or another AI tool that you're using to really help you to create something. One of the important points in creation of content is to have unique content. We wanted the content to be SEO-optimized. I mean, AI can't do everything regarding SEO, for example, backlinks and things like that, but it goes a long way in terms of supporting keyword elements, supporting competitor research, etc.

Now in the last week, I've learned that the limitation of AI being up to September 2021 has been overcome with GPT4, and now we can actually take real-time information.

Aidan

Yeah, we want just to expand on that a little bit. People that are on the Chat GPT Plus program, which is around about $20 a month, you get access to different beta features. One of the ones that they've rolled out in the past couple of weeks has been where you can put Chat GPT4 in web browsing mode. 

For example, I asked a question a couple of days ago, “Find me all the different flight routes, find me the different flights from Buenos Aires to Miami on Tuesday next week,” and I put this specific date. And because Chat GPT can now go and ask questions on the web itself, it can go into like web browsing mode, that's what it's called, and pull back real time data. 

Another way that you can use it is you can say this is something else we've been doing, particularly with social media, is we can say “Analyze this article,” and I can paste the URL of the article. Let's say I've got a post or an article on my website about drone photography. I can paste in the URL of the article and I can say “Analyze this article and give me a 70-word blurb along with hashtags and a link that I can then use in social media.” In the past, it couldn't do that because it didn't have the ability to go and analyze things on the fly. It's gone from being something which was rather amazing, but still 2-dimensional, to amazing and all of a sudden 3-dimensional and has sort of become alive. This is probably the topic for another day, but it's amazing what that has sort of opened the doors to. 

Anyway, going off on a bit of a tangent there, but you were mentioning that we start off by creating a prompt, prompt engineering. This is basically asking a question or giving instructions to the AI tool such as give me a 1000 word article on the topic of drone photography, whatever it might be and much more elaborated than that, so that we get content back in a way that is search engine-optimized, meaning keywords have been built in, meaning it's got a good structure and a good flow to it to give the reader of that information a good user experience and much, much more. We can do all of that part of it quite easily and quickly now with AI. 

Bonnie

Exactly. In the beginning, we did a lot of keyword research, competitor research in the old way, and that was actually helpful as an anchor, if you like, to make sure that we were able to check back to see was GPT4, Chat GPT, working the way we expected to. So, again, that speaks to your last point on sort of a three-dimensional or four-dimensional, five-dimensional type of approach, if you like. 

I then created a spreadsheet. The spreadsheet had a Google Doc outline. We had images that we put into the document. We had Amazon affiliate links, because this particular couple of blogs was doing that, but we would put the links separately in the document in order to be able to check, double check, and triple check them. And then we let the team loose on writing the articles. We've done videos for the system in action. It was really quite imperfect to start with, but we were testing, testing, measuring progress, and we were able to create quite a stunning number of articles per day and per week in the end, and then we tweaked that as we went along.

Aidan

How have you been able to ensure the quality and originality of the content? Because one of the concerns that a lot of people have is Google is not going to like AI content. My response to that has always been, if you're using AI tools in the right way, then it would be 100% impossible for Google to know if your content is human-generated or AI-generated. It should be impossible. If you can actually say, “Oh, this has been made by AI,” then there's a major, major problem there. What's been your learnings or your take on the whole concept of just ensuring the highest quality content possible? Because that's really the objective that we have for every single blog post that we put out there. 

Bonnie

Well, in real life, there are different levels of conversation, if you like. You've got an academic conversation, you've got a conversation with a teenager, you've got a conversation with a five-year-old. So it's recognizing that and asking the AI to really speak to that, if you like, to speak to that level. For us, it's more conversational English. However, we do like to have a high standard of knowledge and understanding by an expert, so we put that all into the prompt. That's how we're able to get a high level of output. 

We use a tool called originality.ai in order to analyze the level of AI and the level of plagiarism. One of the elements of our prompt is that we ask for unique content. Unique content means that the AI will troll the search engines in order to find the unique content or recreate that, but super-fast. 

Aidan

Yeah. Now, how do you take yourself from engineering a system and putting the pieces of the jigsaw in place to then be able to get out of the project and move on to the next one? Because that's always been one of the things that I have appreciated about you, is that you can sort of systematize yourself out of a job or out of a project so that you can move on to the next one. I think this is really a key thing that people should consider for themselves. You want to get to the point where your project is running in a way that is so streamlined that you're not needed in the day to day operations anymore. That's when you can step away and that's when, regardless of how many humans or tools are needed, it becomes passive for you. What do you do? What do you put in place to allow yourself to then eventually step away from a project?

Bonnie

I guess I'm looking for the identity of the project avatar, if you like, as it were, and I'm tweaking and tweaking and tweaking as I go along. Different elements of the system, like, for example, the Amazon links weren't coming out well at one stage, and I had to dive into why that was. The content with Chat GPT wasn't as good as I'd like to, so we rewrote a lot of conclusions, we wrote a lot of introductions. It was through that that we got better and better at the output that we were writing. Certain images were fuzzy that we got as featured images, so it was actually talking to the team about that and seeing, just examining without judgment, if you like, observing what's going on, and just seeing how that might look if it was even better. 

Then there's obviously repetition. So distilling the plan, removing the fluff, monitoring, introducing accountability by observing, not judging, and empowering the team that way, just to let them know they're doing a great job. A lot of the time I'm as positive as I possibly can be, but they're also an amazing team.

Aidan

Yeah, I think I think a lot of it comes to really sort of diving into the process, creating the standard operating procedure and diving into it and then fine tuning it over time. It's an Iterative approach, it's not something that we did right on day one. I remember when I was studying at university, I had a quite charismatic professor at one stage and he was always saying, “Look, the best can even get better. The best can always get better,” and it turned into a bit of a joke because he would say it to us every single day, but boy, did he sort of engrave that into my mind. I think a lot of that is true in what we're doing here. 

I'm conscious of your time, Bonnie, so I've got a few final thoughts that I want to sort of leave people with, and then if you've got anything else that you want to add.

Bonnie

We do have a really great culture in our team. With yourself and with everybody else, there's a whole element of forgiving yourself, for mistakes, there's a whole element of accountability. There's a whole element of because we're all remote working, that we really are the people that are responsible. It's that culture that does bring things forward as well. 

Aidan

I think it's hard to ever perform at your best if you're not in in a culture that that fosters that kind of performance. For some people, I think it comes naturally, and I think for some people, they sort of need to work on that. In any case, it's a big opportunity. 

Going back to what I was saying, a few sort of final thoughts from me here, and then if you've got anything else that you want to add, I'd love to hear it. 

But I think in this whole idea of building processes and systematizing, I've always taken the approach of first I do it myself, then we do it together, and then you do it. I like to be able to at least have a high level understanding of every part of the system. I don't have to be an expert in it, but I want to know, for example, about how to use AI tools to create content, and I want to be able to do it myself, pass that on to someone else and do it together and then make it 100% theirs. 

I like to learn from mistakes. I think making mistakes is fine, but I think it's fine to make the mistake one time. If you make the mistake over and over again and you're not learning from it, then that's a problem. So try not to make the same mistake twice. If you keep pushing forward like that, then you end up with a team of all stars because you've embraced learning from mistakes and getting better and better and better. 

Another thing that comes to mind is diving in and just trying to break things. And we do this with the different AI tools and really sort of push them to their limits and ask the question, “Can this do this for me? Where can we get to with this?” It's great to do that with technology, because that's when you really learn how it works. We've done that with a lot of the prompt engineering content creation. We've seen that if we do things one way, ii creates stellar content, but it's prone to such and such an error. If we tweak that a little bit, then it gets better. 

Another thing that I think about with this whole systematization approach is layering the gains on top of one another. It's incrementally winning, if you like. It's like you make 1% improvement here and a 1% improvement here, 1% improvement here, and you wake up one day, a couple of months after you've started a process and all of a sudden you are a through and through expert and you're brilliant at what you're doing. But that's because you've been making deliberate improvement and often micro gains day after day after day, hour after hour after hour. It is amazing how these build on top of one another and it's amazing how they compound because there's a story about making a 1% improvement day after day after day. You end up being 37X better over the course of a year and that's huge when you can do that. 

I think the last thing that really comes to mind here for me is investing in the team and the tools. If you're trying to really do your best work, but you don't have good tools or you have subpar tools, then you're not going to get the very best out of what you're doing. We recently realized that one of our team wasn't using the paid version of Chat GPT, they were using the free version, which is still fantastic, by the way. The very first thing we did was make sure that they had the $20 a month plan of Chat GPT because that would allow them to get better work done faster. It automatically pays for itself because of the speed gains, and that's just one example. You want to make sure that you're providing the investment in the team and the tools needed to really be at your best. 

Any final thoughts from you, Bonnie, that come to mind on this topic? 

Bonnie

Well, I guess the key is to create the process, to work the process and then fine tune it so it can be almost self-sustaining and then put new personnel in place. Then after that, it takes minimal monitoring and you can move on to the next thing. 

Aidan

Fantastic. Well, look, Bonnie, thank you so much for your time here today. I appreciate that. And as I've said, I appreciate everything you do in our team. You're a real linchpin of the operation and definitely I like to get you involved with the systematization of new projects because I think that really plays to your strength.

So this has been a bit of a brain dump here on this episode. If you want to check out the show notes and download the transcription, you can do that by heading over to thegrowthbooth.com navigate to episode number 73, and we will put links in there to all the goodies and important things that we've spoken about today. We would love your feedback. 

I look forward to getting you back here again, Bonnie, in the near future, to dive into this even more. I feel like there are a dozen topics that we could branch off onto and speak for hours about. So thanks very much, Bonnie. 

Bonnie

Thank you for the opportunity. Take care. 

Aidan

All right, guys, that's a wrap. This is episode number 73. Check it out on Spotify, check it out on Apple Podcasts, check it out on YouTube, or check it out on thegrowthbooth.com. I look forward to seeing you on the next episode of The Growth Booth. Bye for now.