ChatGPT is changing the face of marketing (and the world). David Lawrence is on the pod to confirm that all is not lost - and if you use ChatGPT effectively, it can actually make your job easier! If you don’t use it well, you’ll be stuck with mediocrity. This episode discusses the ChatGPT’s limitations, how to perfect your writing prompts, and how it can be used to its best potential.
David Lawrence is the MD and Co-Founder of Rocket, an award-winning Australian digital marketing agency. He is also the co-author of the Amazon #1 best-selling marketing book 'Smarter Marketer'. David has presented at several events including Inbound Boston, Search Marketing Summit, Mumbrella360, CEO Institute and a variety of seminars and in-house sessions. Follow him on LinkedIn.
James Lawrence: Welcome back to the Smarter Marketer Podcast. Today I am joined by a smiling David Lawrence. David?
David Lawrence: James?
James Lawrence: Welcome back.
David Lawrence: Thanks for having me back.
James Lawrence: Third time on the pod. You're now in first place for repeat… I was going to say repeat customers. I'm not sure that's the right way of phrasing it.
David Lawrence: I'm not sure what I'm getting out of it.
James Lawrence: I know that's right. I think you've bumped Alicia from Red Wolf into second place, where it was previously tied for first.
David Lawrence: Quality over quantity.
James Lawrence: Alicia would be pleased to hear that. So we're talking today all things Chat GPT. We've kind of labeled it an emergency pod. And what Australian marketers need to know? I think we don't want to get too much into what is it? I think we presume that listeners have heard about it in the news. Have heard about it in industry news. Have probably jumped into it and played around with it in either a personal or professional capacity. Fastest platform ever to get to 100 million users. Faster than TikTok, faster than Instagram. On a scale of one to ten, how fearful, as marketers do we need to be that we'll have a job in five years time?
David Lawrence: If you decide not to use ChatGPT, then I think you should have a ten rating on that fear. Maybe a nine, if I'm being generous. If you're good at what you do and you can think strategically and you've got some experience about what works and what doesn't work in marketing, then I don't think you should be scared at all. I do think, like a lot of other disruptions we've seen in the industry, our jobs will be completely different in a few years time. For anyone working in the digital part of the industry, that's not going to come as any surprise. That's been the environment we've been working in. Probably the pace of change is going to be a bit quicker though.
James Lawrence: Yeah, totally. I think it's interesting, like, Dave did a presentation to the entire Rocket team recently and in it, I think that the thing that I found really interesting, and it's so obvious, is that what we're using now is the worst it will ever be. I think maybe just talk to that because I think that made a lot of people sit up and kind of listen.
David Lawrence: Yeah, it's funny, isn't it? When something new comes up, it's easy for the critics out there to look at it and find all the faults in it without realising that for every invention ever, the first iteration is by far the worst. And I think that the line I used in the staff meeting was that we will look back and laugh version that we were using late 2022, early 2023, and I think that's firmly true. I've even been finding on the tool that things that didn't work a couple of weeks ago are suddenly working quite nicely. And things that did work a couple of weeks ago that might have been perceived to be potentially a negative of the tool in terms of negative content, they don't work anymore, that the tool is already learning.
James Lawrence: And I think for listeners that maybe aren't as into it as we are, but I think version 3.5 is how it's labeled. Version four is coming out soon. There's no actual launch date or anything on that, but in kind of listening to people in the space connected to the space that have been in the space for years, the feeling is that it's going to be quite revolutionary compared to what we're playing with now.
David Lawrence: Yeah, I think that's absolutely true. And it's obviously really hard to get into the business of guessing where a large group of very bright people are going to take this thing, but it's going to be phenomenally better in future versions than it is now. And that's just something we've got to accept. And so when we're thinking about how we should use it, should we use it and how we should use it, you've got to factor that in that if there's things that aren't quite right at the moment, you have to look past those what are. The kind of the main limitations that we're experiencing.
James Lawrence: But we're using it in the agency, and later on I want to discuss because the feedback I've heard from other agency owners, some questions we've had from clients is like, is it right to be using it? Is it cheating? And whatever else. We're fully in on ChatGPT and we will use it. In a way, it's a tool to help us drive better results for our clients. But what are the obvious limitations that you've observed and that we've seen?
David Lawrence: Yeah, there's two big categories of limitations at the moment. One of them is the things within the tool which I'll talk about, and the other things are the limitations of any AI in something as nuanced as marketing, and indeed in lots of fields. So if we talk about limitations in the tool, there's a big obvious one at the moment, which is the database stopped crawling data at the end of 2021. So if you're looking for anything that's truly recent, anything that's topical, it's going to really struggle. And if you're wanting to write about something that's fast changing, then it's usually going to be quite antiquated, if you consider a year and a couple of months antiquated. So that's probably one of the first big things you've got to be aware of.
David Lawrence: It has some obvious limitations around, sort of garbage in, garbage out, which we talk a lot about in marketing. So if you don't use it properly and if you don't feed it the information it needs to give you a great answer, it will always be incredibly limited. The other interesting thing is it's not a sentient being. It doesn't think and create the way that a human brain thinks and creates. So it's limited to the knowledge that's already out there that it's been exposed to. So if you're wanting to use it to come up with something genuinely unique and powerful, it's just not going to do that. But it's not designed to do that either. However, the reality is that there is so much knowledge out there in almost every subject matter that no one really knows at all. So it knows more than we all know on most topics. Of course, it's going to have trouble putting emphasis on the really important things versus the less important things it has to use. Some of the clever stuff that I don't understand in terms of how it's powered and also how often things are talked about and how important the sources are that talk about them.
James Lawrence: I think that's the thing to get over. When you think of it as basically pulling together and cataloging and understanding vast waves of information, it tends to rip through university texts and Wikipedia and as much of the internet as it can get its hands on. Billions and billions and billions and billions of pieces of data. You know, we've been doing digital marketing for 15 to 20 years, and it's probably fair to say we know, you know, 3%-5% maybe, of what there is to know in the space. ChatGPT has cataloged every single piece of content, reasonably, that there is on the topic and can then spit it back out at us almost in real time. So I think you've got just the idea that you know more or you have more information is kind of something you just have to give up and seed straight away.
David Lawrence: It does some funny things, though, when it gets to the edge of its knowledge, you can make some pretty funny moves. It's a tool that likes to give you an answer. And so a lot of people are sort of naming the errors. It comes up sometimes as AI Hallucinations, so it can come up with some genuinely strange things that sound sensible. They're written about in quite a nice way because it is really strong with its language, but they're just not true and they never were true and they probably never will be true. So you do have to have a pretty healthy level of skepticism. And probably that brings me on to some of the challenges that people who use the tool will have in terms of what it's not meant to be good at, what it sort of can't be good at. And when it comes to marketing, one of the real challenges for a marketer is working out what's going to perform and what's not going to perform. And the ChatGPT at the moment, and I suspect for a very long time, perhaps forever, will struggle with that.
David Lawrence: Because the thing that allows us to know what's going to work is past experience of running campaigns. And often, as much as we talk about being data driven in digital, so often success is very difficult to define. Long sales cycles, complex buyer journeys, dozens or more touch points can make figuring out what worked versus what didn't work really tough for a human brain. So much harder. Impossible for an AI that doesn't actually have access to the information of all the different campaigns in the world, let alone all the campaigns that, say, an agency might be running at one point in time to know what's working and what's not working. So if you go in there and say, ‘what's a great way to do an email marketing campaign?’... ‘Top tips for high converting email marketing campaigns’... it's going to come up with a really fantastic list of things that are going to be close to best practice and they're going to be really compelling. Most experts are already going to know those things. What it won't be able to do is tell you if the subjects that you're playing around with, which of those email subjects are likely to perform the best. Someone who's an expert in email marketing is going to have a pretty good gut feel on that and they should realign their expertise and their strategic mouse rather than hoping that the tool will do it.
James Lawrence: Yeah, it's very quick to get you to above average again. It's passing university courses, university degrees almost in real time. I think if I punched in, write a blog article on, even though it only catalogs to 2021, you can kind of say, write an SEO article on the best trends for 2023 and it'll kind of spit out something that it thinks is right. It's reasonable. You'd read it, you go, yeah, it's probably about right. Is it cutting edge? Is it challenging you as a marketer? Is it making you kind of ask questions that you haven't thought about in terms of SEO? Probably not, but it's very quick to get to outsource copywriter type level of someone who's just punching out 500 to 700 word web blog articles to fill out a website. Right?
David Lawrence: Yeah. Fast mediocrity.
James Lawrence: Which isn't a bad thing. That's not a bad thing at all. It's a thing that it is what it is. Fast mediocrity is excellent because a lot of the content that gets created in the world is below mediocre.
David Lawrence: There's no excuse for being below mediocre anymore. But it does raise it - maybe we're jumping ahead - but it does raise an interesting question, which is; if the ability to create mediocre content quickly has been democratised, what is the value of that content anymore? And I obviously have to say it's not very valuable.
James Lawrence: I reckon that's a really good segue, because I do want to jump into a lot of the kind of the positive, practical applications of the tool for marketers. But I think it's a really nice segue. When we first started playing around with the tool back end of last year here, it was almost like it had this kind of dirty connotation to it and it's kind of like, is it okay to use it? Are we allowed to use it? Can we copy text and put it onto a client's website? Do we have to change it? If we do change it, what percentage? You've got schools and universities around the world banning it, saying we're not letting our students use it. Then there's talk of, you can copy a suite of text from Chat GPT, put it back in and go, was this written by ChatGPT? And it'll say, yes, make some variations. It'll say, I'm not sure, or no, it probably wasn't. There is technology to watermark. It's probably not going to work very effectively. There's different schools of thought on that, but I guess what's your position as an agency owner? Doing work for lots of clients where the stakes are high. Are we allowed to use it as marketers? Are there some caveats around that? What's best practice for Australian marketers? To stay out of trouble?
David Lawrence: So speaking for Rocket Agency, I'd be upset if someone in our team wasn't using it. Our job is to do the best job by our client and deliver the most value. And this is a great tool for that. It's a great tool for delivering value. I would be even more upset, though, if I discovered that someone at Rocket was only using Chat GPT to gain knowledge, make decisions and produce content. If you're aiming for more than mediocre, and I believe that agencies in general are aiming for more than mediocre, then you have to be doing a lot more than just using Chat GPT.
James Lawrence: That's it. I think I'll bring up Google. Basically, there was a lot of talk back end of last year and in January about Google considering content that's been taken from ChatGPT and put onto websites as blogs and website copy and white papers as being spammy content. They've updated in Google Search Central in the blog section - they posted a piece on February 8. So a couple of weeks back from when we were recording this and essentially, it’s saying it's totally fine to use it with some kind of caveats rewarding high quality content however it's produced. Google aims to reward original, high quality content that demonstrates the quality of what we call E-E-A-T; expertise, experience, authoritativeness and trustworthiness. They're basically saying that they themselves use AI, that it is the future. The appropriate use of AI automation is not against our guidelines. This means that it is not used to generate content primarily to manipulate search rankings, which is against our spam policy.
James Lawrence: So if you get ChatGPT to write a paragraph of text on your website that is bang on, you look at it and go, couldn't have written a better myself. Absolutely happy days. If you get it to write a blog article in an area of kind of deep expertise and you want to modify 20% of it to make sure that it is on point, totally fine. But if you're kind of spinning up blog articles for the sake of just beefing out your site that the user itself has no real interest in it, then Google will have a problem with that. And then that kind of parlays into what you said before about mediocre content. We started in the search business and it was all about just pumping people to your website, any kind of content marketing, subject matter, expertise area. Then it kind of went to, well, if you've got blog articles that talk deeply about what you're doing, you will attract as to your website, the quality didn't really matter too much.
James Lawrence: We've now moved to a point where we don't really advocate just blog content for the sake of it, for our clients. We're trying to create valuable content that users interact with because that's what actually helps you generate better SEO results. I don't think it's controversial to think that in two years time, whether it's through ChatGPT, or whether it's through Google's alternative offer, which is going to be coming soon, why would anyone turn to Rocket to read an article about SEO trends in 2023? You just pump that into Chat GPT or into Google's version. Unless Rocket has something to say that is so unique and different that it sits above the content on those types of websites, I think just content itself has forever changed.
David Lawrence: Yeah, I couldn't agree more. And already that's largely the case, isn't it? Google has indexed more pages than it would ever need on every topic you could ever imagine. So as it is at the moment, writing the words is not enough. And I think that's one of the concerns that I've heard from lots of marketers is that because there's a tool that can bring up answers to questions quickly, and because it can write well, that suddenly they could become irrelevant. And if we look at how we spend our time, unless you're a copywriter, writing words is a small part of what you do. In otherwise quite complex environments where we all spend a lot of time managing people, truly understanding audiences, clients, products, services, offerings… trying to find out as much as we can about the competitive landscape, finding differences, places that our clients or our company, if we're working in house, can compete in.
David Lawrence: ChatGPT is not built for that kind of work at the moment, and it may never be built for that kind of work. It's effectively lifting the standard of the written word to a certain point and it's allowing it to be created quickly. And you're quite right, if you were going to write a generic SEO article about trends in 2023, there's probably not a lot of value in that in the future. But to be honest, if you did it today, you're probably going to struggle to really rank on that anyhow, unless you've got a website where an awful lot of work has already gone into, where actually writing an article on that topic at the moment.
James Lawrence: I'll have it done before this pods finish recording.
David Lawrence: Perfect. But it's a great example, isn't it? Because the ability for an agency like Rocket to write an article like that and rank on it is not just about the words on the page or the screen, it's about all the work that's been done on that website over a number of years. The connections we have with other websites, technical structure and all the rest of us. The words are really important, but mediocre is not going to cut it. And it's interesting if you do think about those SEO trends in 2023 type articles, they do get more valuable when you bring them down into a particular location or a particular industry or a direct analysis for client. And that's where markets have to be playing. They've got to be using their experience and their strategy. Now, the ability to communicate, that's what they should be doing. And ChatGPT will just get them geared up quicker than they might have done before.
James Lawrence: I think it's kind of just fast tracked where things were going with Google anyway, if that kind of makes sense. And I guess we're probably in the weeds on talking about SEO at this point. It has just meant that if you haven't got something valuable and interesting and unique to say, it's kind of not really worth kind of pumping it out there. And whether it is through ChatGPT or whether it is through Google Bard, at the moment, people are thinking about it, how do I use it to help me kind of market when we're talking about digital marketing? Better in Google, better SEO, better Facebook ads, better Instagram ads, whatever else. But as we've seen, we've got the fastest growing platform ever. 100 million users. Mum and dad may or may not be on it,
David Lawrence: We’ll have to check.
James Lawrence: But it is fast becoming a destination of its own. And sure enough, as every other platform that has eyeballs, it will be a place where you have to buy advertising. You want to be there buying ads is there a way of manipulating the returns itself to start bringing your brand into that space and it will become just another kind of territory that us as marketers are trying to get our clients in front of.
David Lawrence: Yeah, that's right. And it's probably outside the scope of the discussion today, but it'll be very interesting to see what Google in particular and being in other platforms do to ensure that they're very, very lucrative. Search businesses remain lucrative. It's definitely not obvious to see how you can monetise an AI response, which often will just require one view, one read, compared to the way searches worked after.
James Lawrence: They’ll find a way, Dave, they'll find a way. They always do. I think that's really good. What I want to talk about now, the obvious use cases. The obvious use cases like, obviously try a blog article, chuck it in there, whatever else, but love to hear just how you think ChatGPT can be used for marketers on the pod.
David Lawrence: Yeah, excellent. So let's start right at the beginning of how we often have to work as marketers, which is doing some research. So you've got a new product or service or new client or whatever. It's a great fast way to get some good top level knowledge. All the caveats about it being out of date apply, obviously, but it's great at basically doing a query to say, tell me about a certain industry in a certain location, what are the current issues? What are the trends? What are the strengths, weaknesses, opportunities type stuff? So it'll be good at that. At the moment, you wouldn't rely on that research wholly because it definitely is not always correct, but it's a great starting point.
James Lawrence: Not bad. We've just started working with an Australian whiskey maker and that was an example of what one of our team members did, basically give me some research in the Australian whiskey industry. How big is it? Who are the biggest players, the value, the challenges, opportunities, whatever else? And given that we know the data is not up to date, in terms of being a couple of years old, definitely fast tracked, I wouldn't rely on that information by itself. But in terms of just having 700,000 words written within minutes, pretty effective.
David Lawrence: Yeah, that's it. If someone said, you've got two minutes to become a semi expert on a topic, I'm going to Chat GPT instead of Google at the moment.
James Lawrence: Yeah, really interesting.
David Lawrence: So that’s a good obvious one. There's lots of fields where it's already providing some really exact work, and coding is a great example. Marketing shows the sort of work you do with data tracking and analytics, so you can go in there and talk in plain language about something you want to achieve. So, for instance, I did Facebook Pixel, which is going to track when someone scrolls 35% down through a page and it will just write JavaScript code in a matter of seconds or less. And it's not always perfect, but it's pretty good. The same if you wanted to add an FAQ section to one of your web pages, you can just put the content in and ask for the code and it'll come back with all the front end code that you need to drop into your website. I haven't used it for this because I'm not a coder, but it's certainly able to punch out pretty good code in all the common programming languages as well. And I've heard people talk about the fact that it's better than a junior developer. So usual challenges we find as well as marketers. It's not going to be strategic in terms of how it writes code, but once you know what you need to do, you can do a pretty handy job of saving a lot of time in what is otherwise quite repetitive work.
James Lawrence: I think an interesting one that we're playing with was just putting together email sequences, maybe just kind of how you framed that was quite, I think, quite interesting giving the brief.
David Lawrence: Yeah, that's right. And maybe this is a good time to talk a little bit about how you can make sure some of these more specific use cases actually give you quality results, which is the way a lot of people using ChatGPT is not surprisingly, quite novice, because it's quite new for most of us. But you'll go in and you'll say, write me an article on Blah, it comes up with something that reads pretty well, but the deeper you look into it, the more you realise you couldn't actually run with that as a quality piece of content. So what you have to start doing is giving a lot of information in the form of a brief. So if you're briefing a copywriter at the moment, you wouldn't just say, write me an article on Blah; you've got a client, this is their industry, this is their location, this is their audience. Tone.
James Lawrence: What's that? Tone.
David Lawrence: Tone 100%. You'd be talking about tone. You might list the USPs, basically everything to make sure that the copywriter can come back with some content that the client reads and says, wow, this is perfect. Like, I've been working with you guys for years, so you have to do exactly the same thing with ChatGPT, and that's the number one thing you'll do to get a step change lift in the results that it's going to provide to you. So getting really good at understanding how to write a brief for ChatGPT. Expect that to change over time as the tool changes over time. But in general, plain language seems to work the best with most things.
James Lawrence: How do you do that for listeners that have kind of gone in there and said, hey, write a website, homepage copy in electrical engineering for a business of 20 people. How would you elevate that? How do you actually put those prompts in?
David Lawrence: Okay, so first of all, obviously any marketer should have all these bullet points or more detailed bits of information about how they would brief someone working on the campaign for their client or their employer at the moment. You basically want to take that and you just want to convert it a little bit to make sure it works nice to ChatGPT. So, short sentences, you don't want too much ambiguity. Some things you put in just don't seem to work at the moment. Other things seem to work really well, especially been driving me crazy at the moment, because you put in use Australian English and it keeps on throwing back words with Z's in them, which is frustrating, whereas I'm sure that was working a couple of weeks ago. So the brief is essentially just a series of short sentences with all the information a real person would need to write something that's personalised to your needs.
James Lawrence: So put together an email marketing series for this type of business. The target market is X, the USPS are Y, 2s tone, playful conversational, formal, whatever it might be.
David Lawrence: That's it. And the way I tend to do it is I put all my briefing information in there first and I have that in a library of assets that I would have. I do a lot of writing for Rocket, so I now have a nice little ChatGPT library, which is briefing information for Rocket, and I have a big collection of lines, but I wouldn't use them all for everything. Like if I just need some assistance with a paragraph about a particular thing, I'm not going to put the entire brief in. That’s sort of overkill. But you'll have that brief, and then once that's in there, the next line you'll put in is you might say - a lot of people are using the phrase ‘act as so act as a copywriter’. So if it's an email sequence, you could go create an email sequence, for someone who has, for instance, downloaded a guide from our website. I'd like the first email to go immediately, suggest the next three emails and give me titles and copy for them. And if you brief it really well and you use the right instructions for that prompt, then you're going to get something back which is actually quite good.
David Lawrence: It has a good knowledge, a lot of stuff written online about email marketing and sequences, so it has that knowledge. And I think the next thing that's really important with Chat GPT again is on that prompt side of things. So if you think about it, we've done our brief, you keep prompting and think about prompting as the way you give feedback to an email strategist or a writer who'd come up with a sequence. You're unlikely just to look at it and go; ‘it's perfect, I'm going to run with it’. You would typically give feedback and chat GPT is no different, so you want to give it feedback. So you might want to say if it's given you four options, you might want to say for option one, give me three more variations of the headline with different tones. Or you might say, specifically, want the tone to be playful or educational or informative, and you can play around with different words that you'll get interesting results for there.
David Lawrence: And so it's all about the, what we've kind of named it internally, is ‘pick and edit’. So what you want to do is you want to pick the option that you like, continue using chat GPT to improve it as much as it can. But at some point chat GPT is done, you take it offline and then you edit it. If you become really good at briefing the prompting, the amount of editing you'll do might be a little bit less, but most times there will be editing to do because it won't capture the full requirement of the brief. There'll be things that, you know as an expert marketer that it won't know.
James Lawrence: My experience with it was I first started playing with it and I was like, this is unbelievable, it's incredible. Then I started going, hang on, it's actually not as good as it's so light on, it's so mediocre. And then I realised, hang on, I'm just not actually using it correctly. And once you do go through that process of properly briefing it, putting in smart prompts that you know right after using it more, reading up on different ways that people in your industry are using it, but then also not expecting it to be. A complete panacea. There's still going to be changes you have to make at the end to kind of put the finishing touches on the item list here of the things that we're using it for writing articles, emails, ads, social posts, web content. Create a Facebook ad. Create a Google ad. Limit my characters to 50 characters or 60. Those things are just so difficult to actually get right normally. Generating top level ideas for a campaign, designing email sequences, creating scripts for videos. It's awesome at scripts, even just from a personal context. Like write a poem, turn it into Old English, turn it into gangster speak. It's so clever at moving it around. That could also be very professional if you needed it for it. I don't know. Depending on what industry you're in.
David Lawrence: Writing policy, that's a common task for you.
James Lawrence: Yeah, absolutely. That's why Mum likes me the most. Perform market research. I think it's excellent for search. Build me a table of the ten biggest businesses in this particular vertical. What it said? Turnover. Put your sources in. You can put that in. So it'll actually give you sources sometimes and endlessly rephrasing existing content, creating lists with sources, keyword research, writing code, creating complex Google Sheet formulas I think this is a really good one, just generating fresh ideas to avoid writer's block. I often really struggle to get started writing something, but I'll just have it write something pretty average and I'm so happy to jump in there and play around with it. Creating tracking events querying KPIs to add it to a spreadsheet. It's a bit of a cliche, but it really is just limited to your creativity with how to use the tool, right?
David Lawrence: Yeah, it's so true and I think one of the big takeaways for me with where the tool is at the moment is it's a tool for experts. And that kind of means two things. You've got to become an expert on how to use Chat GPT; by no means the most complex thing to use in the world, but you'll get better the more you use it. But the most important thing is that you're an expert in the subject matter you're working in. So it's easy for someone to think that with this tool, you can go from knowing nothing to becoming an expert marketer in an hour. And certainly if you're crawling TikTok, you find all kinds of people claiming to have built amazing agencies with incredible results without doing any work other than firing up Chat GPT. It's kind of fantasy land. It's completely dishonest. The strength of Chat GPT is in assisting an expert to get to a quality result faster and it allows them to bring in more fresh perspectives and interesting ideas than they might have got otherwise. So it's sort of almost like you're brainstorming with a bunch of people, but you're still going to be someone who knows what great looks like.
David Lawrence: And you've got to know when the thing you're looking at that's just pretty good, could be turned into great by just taking it offline and doing some editing for it. So, yeah, if you're not an expert marketer, this is going to be a tool that just helps you become mediocre really, really fast. One of the challenges I found with using it is it's quite overwhelming. So it's really easy to say, give me 30 versions of that headline, and by about version four for me anyhow, I find they order stuff looking the same. It's genuinely hard to sort of find the real quality stuff in there. So that a skill, I think, that we'll develop. It's a skill that editors have had forever, but it's a skill the rest of us will have to develop, which is the ability to read work, not be impressed just by the words that are there, but actually see through to the quality very quickly.
James Lawrence: I reckon that's enough in terms of just our perspective on the tool, what it is, what it's not, different ways of using it. Anything else? Any other kind of areas you wanted to cover?
David Lawrence: I could talk about this forever.
James Lawrence: You're a massive nerd.
David Lawrence: I’m a massive nerd. This just taps right into great things for me. Look, I think the main thing for me is get into the platform, play around with it. There's a framework that we're playing around with internally, which you just pointed out was the four P's, which wasn't really when we put it together, but I think this is an important way to think about how you can use Chat GPT across an organisation to make the organisation effective. And I don't think it matters whether you're a marketing agency or any other company. So for us, it starts with process.
James Lawrence: It's the three P's, Dave.
David Lawrence: Okay.
James Lawrence: Wow. It's the P-B-P-P.
David Lawrence: Interesting. Well, we can throw that into ChatGPT and look at better options, but it's really simple for us. Process is where it all starts. So you've got to understand which parts of the things you do at the moment could benefit from ChatGPT. How are you going to achieve that? So you got to be really clear on that. Otherwise you could dumb down the work you do pretty quickly. The next you do is the brief. So getting really, really good at the brief is critically important. We've talked about that already. That's an art. I know some people have been referring to using ChatGPT as engineering and ChatGPT engineers. That's a disgraceful use of the word engineer. It's just not that hard. But you will get better at it as you go. The next thing is prompts are getting really powerful. Building a library of prompts that work really well and share them with your colleagues, that's where you'll get best at it. And then at the very end, it's pick and edit.
James Lawrence: I'm just busy in the tool here trying to change process brief. Pivot, progress plan, proceed. What do we like instead of brief? Plan, process, prompt, pick. We've done the four P's. ChatGPThas beaten us. It wins again.
David Lawrence: I had a funny interaction with a colleague the other day where they were doing a similar thing, coming up with three words to describe a company. Came up with some okay kind of stuff. And then I said, tell it to make those words rhyming. And I was able to put together five really great rhyming words to describe a brand. So those sort of things that humans struggle with, it is better than us of those things.
James Lawrence: Yeah, that's it. It's amazing. Cool. Okay. We always ask this question of people on the pod. You've been asked this twice before, so Kayley's asked me to tweak it a little bit. What's the one piece of marketing advice that you'd give to marketers coming into the new year?
David Lawrence: This is not a shocking one, but learn ChatGPT, so don't just. Use it a little bit and think it's okay. Put some time aside, do a one day deep dive, use some of the tips we've gone through in this pod. And that's just the starting point. Put a day aside. Social is great for finding people that are using ChatGPT really well. So have a look around, try what people are trying and then see how it applies to your role and think about where it fits into your process.
James Lawrence: Yeah, couldn't agree more. I think this is not something you can put your head in the sand on it. You've got to embrace it. There'll be lots of roles and lots of businesses where it'll be business as usual a little bit for a long time. And I think there's a lot of hyperbole around what it's going to do to kind of white collar jobs across all industries, not just marketing. But if you don't embrace the tool and I think you are in danger of being a bit of a dinosaur.
David Lawrence: Yeah. And I think one other tip, maybe this is a little bit dark. If you do a job at the moment where you've often thought, smart robot could do this, they probably can. And it's probably time to start looking at reskilling. Like we mentioned at the very beginning, the world we're moving into is going to reward people who are strategic, greater, strong communicators. That's the kind of stuff you want to be doing. You want to do the stuff that isn't simply a series of obvious rules or obvious language.
James Lawrence: Yeah, that's it. Otherwise you'll be replaced by a robot.
David Lawrence: Yes, you will.