How well does ChatGPT understand consumer behaviour?

Facebook Twitter LinkedIn

While ChatGPT is on everyone's minds, Dave Chaffey explores why a 'human' approach is still essential for getting the best out of AI marketing tools.

Why you still need to apply Behavioural Science to gain an edge in your marketing in the ChatGPT era

I’ve found that people who work in marketing are from a variety of backgrounds. Some started with college marketing degrees, but they seem to be in the minority. Like many in digital marketing, I’m from a tech (Software development) and science (Geochemistry) background, but now I’ve worked in digital marketing for quite a while, I sometimes think what I would have chosen to study if I knew I was going to be working in digital marketing. Given my time again, I would choose a Psychology plus Marketing degree, since so much of hands-on marketing is creating effective communications to engage your audience, so understanding the reason behind how people think and act the way has to be really useful when crafting value propositions and copywriting.

That boat sailed long ago now, but there is always the option to learn and, short of completing a whole degree, one of the best ways to learn about the intersection between Marketing and Psychology is to learn more about Behavioural Science. At ClickThrough Marketing we're now encouraging our team members to develop their skills in this technique since we see it as so important to success and it helps set us apart. Even in 2023, where everyone is sharing examples of the power of ChatGPT to generate marketing copy, it pays to know the principles of behaviour science and how to apply them to marketing.

To learn, the principles direct from an expert, with examples of how you can apply behavioural marketing to make your paid media work harder, I recommend you re-watch our recent webinar with Sarah Clarke, our head of Paid media (and a psychology graduate herself), where she explains how you can integrate Behavioural Science to boost your marketing campaigns.

Behavioural Science - Sarah - pre-webinar

Can ChatGPT help you follow behavioural science principles?

Like many in marketing, I’ve recently been using ChatGPT to understand its implications for marketers - read my article on Using ChatGPT in marketing if you haven’t yet had the opportunity to understand the principles.

Many marketers have already shared examples of ChatGPT dialogues to create headlines for blog posts, ads and emails. For me, the outputs are undeniably impressive, but people will still need to apply their own intelligence to create the inputs and review the outputs.

One of the key takeaways for me from using ChatGPT, is that marketers will need to develop skills in prompting the AI. This is a new skill that we will have to learn since, although ChatGPT is an advanced natural language generation (NLG) model that doesn’t require prompts (known as ‘Zero shot’ amongst AI professionals), we will see it works much better if you can be more specific, i.e. you give it specific context through examples and more detailed instructions (‘One shot’ or ‘Few shot’). So, we can’t just rely on the AI, we have to apply our experience and intelligence as, what some are calling, Prompt Engineers. It’s a skill just like learning Google - in the way that some people are better at getting answers from searching Google because of their experience.

Let’s take a look at an example. I asked ChatGPT to create 3 marketing email subject lines by prompting it about the purpose of the three emails and the the type of business they related to. Here are the results:

ChatGPT and Behavioural Science - Email 1 - 20.01.23

You can see that on first glance, the three subject lines look compelling. They’re natural and engaging. Even without asking it to consider behavioural science, the recommendation about creating a sense of urgency or scarcity and using personalisation in the subject line are well-known principles. However, what I was trying to achieve was to get it to produce a series or sequence of three communications that were in a logical order. No matter how I phrased the question, it just gave me three alternatives, not a sequence!

To gain more ideas, I asked for the copy to focus on ‘switching’, since that’s the main way the charity I’m working with on this gains most of its customers. These are results; it’s cool the way the AI remembers the B2B context of the conversation and it understood the principle of switching, developing some emails that are more compelling than the first.

ChatGPT and Behavioural Science - Email 2 - 20.01.23

As well as learning the prompts, I’d encourage users of ChatGPT to use it as a collaborative learning tool rather than a black box. One of its biggest strengths compared to Google is that you don’t have to scan and select multiple responses, but one of its biggest weaknesses compared to Google is that you don’t have the source to follow-up to see whether you can trust it. I also got the AI tool to create an email and then asked it why it had structured it the way it had. It’s logic in structuring in that way is, again, impressive.

ChatGPT and Behavioural Science - Email 3 - 20.01.23

As another learning example, I asked ChatGPT to explain the principles of persuasion applied to marketing. It gave a neat explanation of each of Social Proof, Scarcity, Reciprocity, Authority, Consistency, Liking. However, it didn’t explain why it had recommended these. Now I and some of you will recognise these as based on Cialdini’s book, Influence: The Psychology of Persuasion, but I had to ask the AI to give this and recommend further reading from him and others like Dan Ariely and Richard Thaler. You can read the full transcript of this exchange at the end of this article.

So, where does this leave marketers looking to engage their audience and increase response to their communications? Well, certainly it has become much easier to generate copy alternatives. But it’s very difficult to second-guess which creative will give the biggest response. This is all good, since it encourages testing of creative. We have always recommended to build in testing regardless of media at ClickThrough Marketing and Sarah will explain processes in her webinar.

Watch the Webinar

I think that within different forms of media, the best results for combining AI with behavioural science won’t come from generic tools like ChatGPT, but from AI Martech tools like Persado and Phrasee which have been used by savvy brands for years to optimize their communications based on behavioural science principles. Returning to my original question actually, I think it would be useful to study a combined degree in Marketing, Psychology and Artificial Intelligence!

 

what can ChatGPT itself tell us about behavioural science?

Exported with AIPRM https://www.aiprm.com by **@gmail.com on 13/01/2023, 09:10:32

User: how can behavioural science be used for marketing?


ChatGPT: Behavioral science can be used in marketing in a variety of ways. One common application is using principles of psychology and sociology to understand consumer behavior and create more effective marketing campaigns. This can include using persuasive techniques such as social proof and scarcity to influence purchasing decisions, as well as understanding factors that drive brand loyalty and customer engagement. Behavioral science can also be used to inform product design and pricing strategies, and to improve the customer experience through design and communication. Additionally, some companies use behavioral science to segment their target market, and personalize their marketing approach to different groups.


User: What are the use cases in search marketing


ChatGPT: In search marketing, behavioral science can be used in several ways to improve the performance of search campaigns. Some examples include:

  • Understanding user intent: Behavioral science can be used to understand why users search for certain keywords and phrases, and how they interpret the results they receive. This can help marketers create more relevant and effective ad copy, titles, and meta descriptions.

  • Optimizing landing pages: Behavioral science can be used to optimize the design and layout of landing pages to increase the likelihood that users will convert. This can include testing different headlines, images, and calls to action to see which are most effective.

  • Personalization: Behavioral science can be used to create personalized search experiences for users based on their past search behavior, demographics, and other factors. This can help increase the relevance of search results and improve conversion rates.

  • Bid optimization: Behavioral science can be used to understand how users interact with search ads and to optimize bids on keywords in real-time based on their behavior and engagement.

  • Search retargeting: Behavioral science can be used to understand how to target users who have previously searched for certain keywords or products, and to serve them ads based on their past behavior.

  • Local search optimization: Behavioral science can be used to understand how users interact with local search results and how to optimize local search listings and campaigns to be more effective.

 


User: and specifically for organic search?


ChatGPT: Behavioral science can also be applied to organic search to improve the performance of a website. Some examples include:

  • Understanding user intent: Behavioral science can be used to understand why users search for certain keywords and phrases, and how they interpret the results they receive. This can help website owners create more relevant and valuable content that meets the user’s needs.

  • Optimizing website design and navigation: Behavioral science can be used to optimize the design and layout of a website to increase the likelihood that users will stay longer and engage with more content. This can include testing different layouts, colors, and calls to action to see which are most effective.

  • Personalization: Behavioral science can be used to create personalized search experiences for users based on their past search behavior, demographics, and other factors. This can help increase the relevance of search results and improve user engagement.

  • Content optimization: Behavioral science can be used to understand how users interact with different types of content and to optimize the content on a website to be more effective. This can include testing different headlines, images, and format of content.

  • Search retargeting: Behavioral science can be used to understand how to target users who have previously searched for certain keywords or products, and to serve them relevant content based on their past behavior.

  • Local search optimization: Behavioral science can be used to understand how users interact with local search results and how to optimize local search listings and campaigns to be more effective.

  • Search engine result pages (SERP) optimization: Behavioral science can be used to understand how users interact with SERP and to optimize the website’s organic listing to increase visibility and click-through rate.

 


User: Thank you. What are the principles of behavioural science that affect consumer behaviour through persuasion tactics on a website, e.g. social proof and scarcity.


ChatGPT: There are several principles of behavioral science that can be used to influence consumer behavior through persuasion tactics on a website. Some examples include:

  • Social proof: Social proof refers to the phenomenon where people tend to look to others for guidance on how to behave in a given situation. On a website, this can be used by displaying customer reviews, testimonials, or the number of people who have bought a product. This can help increase trust and credibility in the product or service being offered.

  • Scarcity: Scarcity refers to the idea that people place a higher value on items that are in limited supply. On a website, this can be used by displaying limited time offers, limited stock, or by highlighting that a product is selling out fast. This can create a sense of urgency and increase the likelihood of a purchase.

  • Reciprocity: Reciprocity is the social norm that people feel an obligation to return a favor or a gift. On a website, this can be used by offering free samples, free trials, or even free shipping. This can encourage users to return the favor by making a purchase.

  • Authority: People tend to follow the lead of experts or authority figures. On a website, this can be used by displaying certifications, awards, or by featuring the company’s leadership team. This can increase trust in the product or service being offered.

  • Consistency: People tend to want to be consistent with their past actions or decisions. On a website, this can be used by asking for a small commitment first, such as signing up for a newsletter, before asking for a larger commitment, such as making a purchase.

  • Liking: People tend to be more influenced by people they like or that they perceive as similar to themselves. On a website, this can be used by highlighting customer stories, showing pictures of happy customers, or by making the branding relatable.

These principles can be used to help create a more effective website, by considering how users interact with a website, and how to make the website more persuasive.

 


User: Who first identified these principles - which researcher or author


ChatGPT: Many of the principles of behavioral science that affect consumer behavior through persuasion tactics on a website, such as social proof, scarcity, and reciprocity, have been identified and studied by various researchers and authors over the years. Robert Cialdini, is an American psychologist and professor emeritus of psychology and marketing at Arizona State University, He is considered one of the leading experts in the field of persuasion, and he is the author of the book “Influence: The Psychology of Persuasion.” He describes the six principles of persuasion, including reciprocity, scarcity, social proof, authority, consistency, and liking. Dan Ariely, is an Israeli-American behavioral economist, and a professor of psychology and behavioral economics at Duke University. He is known for his work on behavioral economics and irrational decision making, and he has written several books on the topic, including “Predictably Irrational: The Hidden Forces That Shape Our Decisions.” Richard Thaler, an American economist and professor of behavioral science and economics at the University of Chicago Booth School of Business, He is a co-author of “Nudge: Improving Decisions About Health, Wealth, and Happiness” and a pioneer in the field of behavioral economics, which combines insights from psychology and economics to understand why people make certain decisions. These researchers have made significant contributions to the field of behavioral science and have helped to identify and understand the principles that affect consumer behavior through persuasion tactics on websites.

 

Talk to our experts about how you can use behavioural science and ChatGPT to create quick, effective marketing campaigns.

Get In Touch

Talk To Us About Accelerating your Digital Performance