Looking for ways to refine and improve your social media marketing efforts in 2023?

Over the next few weeks, we’ll be sharing a range of key tips and pointers to help you maximize your efforts, as well as essential tools and processes that you can use to connect with a broader, more receptive audience online.

In the first post of the series, we looked at how to define your brand purpose statement, and how that can then guide your process moving forward.

With that established, we move onto the next element - researching key queries in your industry, in order to map out a more effective content approach, both for SEO and social posting purposes.

Understanding what your audience wants to know, and how you can meet them in their search process

The key to getting your brand and products in front of your target customers is to understand where people are looking, and what they’re looking for in relation to your business.

Optimizing for search is the core consideration here, but increasingly, people are also searching on social platforms for more information about products and businesses. And then there are ads, and meeting demand that people didn’t even realize they had. All of this requires research, and understanding both the common queries and questions, along with competitor insight, and how they’re delivering (or not) on the same.

So first off, at a high level, you need to get some measure of the most common queries relative to your brand, products and region.

The best starting point here is Google’s Keyword Planner tool, which provides key insight on search trends relative to whatever term you enter.

So you need to have at least some idea of the key search terms for your business to begin with, but from there, you can build a large corpus of potential search terms, queries, questions to be answered in blog posts, inspiration for social polls, and more.

As you can see in these examples, if I’m looking to sell basketball shoes, I can enter ‘basketball shoes’ as a search term and get a full listing of related searches, which I can then sort by popularity, competition, trending interest and more.

I’ve removed all of the directly branded searches in this second listing, in order to provide a clearer idea of generalized trends for content ideas.

As you can see, there are some good notes of inspiration in there, including ‘best basketball shoes’, ‘league basketball shoes’ (i.e. what sneakers NBA players are wearing), and ‘best basketball shoes for jumping’.

Building on this, I can start to get an idea of certain content directions - like, for example, creating my own video reviews of new sneakers, rundowns of the most popular shoes worn by NBA stars, feature overviews based on purpose (running, jumping) and more.

If you’re looking for the most common questions specifically, Answer the Public is another handy resource.

Enter a search term, and Answer the Public will provide an overview of all the most commonly asked questions about it, based on Google autocomplete data.  

Another tool to consider here is Google Trends, and checking in with the latest trending topics, to see if there’s anything of relevance for your brand.

You can search by key terms to get a gauge of interest, and related queries, while you can also browse through popular trends to see if any might relate to your business. If there’s an angle there for your promotions, it could be a good tie-in opportunity, which could be another way to gain traction on social networks.

You can also use ChatGPT to assist in your research efforts.

ChatGPT’s data isn’t updated in real time (OpenAI notes that it can only access data up to 2021), so it can’t, for example, give you up-to-the-minute insights on search trends.

But it can give you some notes on common questions, for a specific region, which could give you some more food for thought.

A key limitation of AI-based systems is that they are just that, systematic responses to the outputs you enter. They can’t qualify the results, so you can’t know, for certain, that all of the info they provide is accurate. But as a starting point, when trying to uncover key angles and notes, it can be a handy, and more intuitive way to dig into topics of interest by region, and you can also prompt it to provide more answers, with different qualifiers, based on the initial query.

Or you can ask it to show you the top questions on a given subject, with whatever qualifiers you want to add, including region, audience, etc.

You can then ask ChatGPT to put these questions about a product into a table, which you can then add to your Google/Answer the Public lists.

In combination, you can build up a pretty big collection of prompts and notes to guide your content strategy, which at the least will give you a means to build connection with your audience, in various ways.

Can you use ChatGPT to create posts on these topics?

Yes, you can, though I would exercise a level of caution here also.

For the most part, it depends on the purpose of the posts that you’re looking to create, and the experience you want to provide visitors on your site.

So yes, a simple, cheap and time-saving strategy, in expansion of this, could be to enter the questions you list as prompts for posts in ChatGPT, with further instructions to optimize these posts for certain key terms.

ChatGPT can do this, and legally, you can use it on your site. Kind of.

Google has explicit rules against the use of AI-generated content, and there are tools being created that can detect ChatGPT-originated material. There are also various legal debates underway around the use of such tools, and who essentially owns the rights, and all of these are developing areas that you need to keep an eye on if you are going to use AI-originated material.

Also, the copy that you’ll get from ChatGPT is based on the inputs that it has access to - which, in most cases, is a lot of average website copy from all across the internet.

So, if you don’t mind your website reading like everybody else’s, and you’re okay with the potential risks (which are difficult to measure at this point), then using this content for SEO purposes could be a way to go.

But as you can read in the example above, it’s pretty generic, pretty bland - I personally haven’t found ChatGPT’s outputs engaging or rich enough to actually use, beyond basic summaries and overviews.

You could use them as a guide, and freshen them up with your own editorial flair and input. That’s possible, but in most cases, you will find that you need to rewrite a lot, to the point where you might be better off creating something from scratch, in your unique voice.

That, in itself can be a challenge, so there may be ways to use a hybrid approach to save time in building an SEO foundation on your site.

Basically, ChatGPT outputs are the median of everything else you read online, and as more businesses pump out more ChatGPT summaries, it’s only going to get more bland and generic. I personally think you’d be better off taking a fresh approach - but there may be options that you can find which incorporate its current outputs with your own additions.

Another option - enter part of your own writing, then ask ChatGPT to follow the same style.

It’s not always great, and you always need to fact-check and edit each piece that the app gives you.

In my view, it’s an additive tool, which requires some experimentation, and can provide a range of benefits. But it’s not the singular solution to all your content needs.

AI for images

While we’re on blog posts, you’ll also need visuals to go along with your updates, both for your social posts and your blog and website updates.

AI-based tools like DALL-E and Midjourney can also be used for this purpose, with both producing some pretty good results based on specific prompts.

A lot of them do, however, come out looking a bit weird, and you may have to try a few times to get something that you actually want to use, but you can also edit and refine them yourself, and enhance the AI creations.

But here’s where they could be particularly valuable in a brand marketing sense:

Using a DALL-E, you can upload your own product images, then use AI tools to iterate on them, by editing, for example, the background, which can reduce the focus on the computer-created visuals, and limit the impact of the weirdness at the edges. 

The example prompt that I’ve used here is pretty average, and you’ll get better results with more specific guidance (some pointers on how to best write AI prompts here). Try enough times and you could end up with some pretty good product images – but do note that the free version of DALL-E uses a credits system, and you only have a limited amount of free usage (you can buy more credits if you need).

Legally, you can use AI created images for commercial purposes, so long as you don’t upload images of people without their consent, you don’t upload images to which you do not hold usage rights, and you don’t create images of public figures.

You should also provide a transparency note - something like ‘Main image created by DALL-E’, as per the respective usage terms.

Do these tools replace stock image libraries wholesale? No, as there’s generally some customization and editing required to get what you’re after, while specific use cases will still work better with actual photos - either your own or via a third-party source. But as with ChatGPT, they can be handy tools to have in your creative arsenal to help improve your process as a complimentary element.

At the least, it’s worth experimenting with these evolving systems as you go, and seeing what results you get.

Use them or not, it costs nothing to try them out. For now, at least.