There’s a better way to do this! Read Using MarketMuse AI to Create Content and watch the accompanying video.
Let’s turn a previously built article outline prompt, along with its output, into another ChatGPT prompt we can use to create the actual draft. Now that sounds way more complicated than it is!
So let’s get started.
But before we do, let me point something out. If you’re using the MarketMuse ChatGPT integration, you’re already miles ahead of the game — the article draft prompt includes MarketMuse data.
We’re just going to level it up a notch.
Here’s the 3-step process I use for turning an outline into a draft with the help of ChatGPT.
- Generate an outline using a detailed prompt for ChatGPT.
- Modify the outline prompt for use in generating the content.
- Add the outline (created by the first prompt) to the end of this new prompt
Why go Through All the Trouble of Creating a New Prompt?
You could ask ChatGPT (in the same conversation) to use that outline in creating an article. But there are a couple of issues with that:
- You’ll probably want to tweak the outline first.
- You want to be sure that all the parameters used to generate the outline get carried through to content generation.
I’m not sure that would happen. But I’m sure that if it didn’t, ChatGPT wouldn’t say anything.
So to be 100% sure, I prefer creating a new prompt by making some minor modifications to the previous one.
Generating an Outline Using a Detailed Prompt
We already did this in the previous post, How to Create a Blog Post Outline Using ChatGPT (and MarketMuse). Briefly, we tweaked the MarketMuse Outline prompt to include additional information around:
- Target audience
- Position in the marketing funnel
- Tone and style
Here’s what the prompt looked like:
Although it took some effort creating the initial prompt, the good news is that it doesn’t take much to turn it into a content-generating prompt.
Modify the Outline Prompt to Generate Content
There are a few minor changes to make that can have major ramifications. So let’s be safe and make sure we’re perfectly clear. Basically, we’re changing any references from “outline” to “article” — but it’s a little more than just finding and replacing.
Here’s how the new prompt looks.
Add the Outline to the Article Prompt
We built the previous ChatGPT prompt to generate an outline. Now we’re going to use the output from that as input for this new prompt.
Literally, all we need to do is paste it to the end of our newly modified article generation prompt to ensure it follows the prescribed outline.
The Generated Output
Here’s part of the output generated by this new prompt. Now, is it perfect? Absolutely not!
There definitely are some edits that I would make. But as someone who’s grown tomatoes for many years, I can tell you that this is a solid initial draft.
And in case you were wondering. The Content Score for this draft is pretty decent too.
While it may not meet the Target Content Score, it’s well on its way. And there’s something else I’d like to draw your attention to.
The word count is far below the target — to a greater degree than Content Score. Look at it this way. Dividing Target Word Count by Target Content Score tells us how many words it takes to earn a Content Score point. The calculation is a shortcut to determining information density — scoring well with fewer words as opposed to writing fluff.
In this example, the calculation using Target Word Count and Content Scores rounds up to 50 — so in an ideal situation it takes 50 words to score one point. For the article itself, the information density calculation comes in at 36, meaning it conveys more information per word.
As content marketers, I think we should focus less on word count and more on information density. It’s not the number of words that’s important, but rather our ability to convey information. This calculation is one way to look at that.
Like anything, this can be taken to the extreme. So you want to find the balance between conveying information in a concise manner and being verbose.
Creating a ChatGPT prompt that generates an article from an outline only takes a few tweaks to the original prompt. All you need to do is change any references from “outline” to “article” and add the outline to the end of the prompt.
The resulting output is an initial draft that requires some editing. However, you’ll probably find that it does a good job at conveying information as determined by its information density score.