We want more quality content for our websites, but it’s hard to produce enough. So how can we scale the content creation process, especially for e-commerce sites with many products?
If you were to pay a writer to publish thousands of product snippets from scratch, you’d likely be out of pocket pretty quickly.
What if you pay for 1,000 new product descriptions, but only half of those products are live a month later? Clearly, you need a faster and more cost-effective approach. This is where ChatGPT can help.
ChatGPT’s native web interface is very useful and a huge time saver.
But if we have hundreds or thousands of product descriptions to create, there is a more efficient way to use ChatGPT without copying and pasting prompts. Here’s how.
Mass production of content snippets: scale output
If you have an e-commerce website, you may want to produce product snippets using data from a product information management (PIM) system.
Let’s say you have the data in a spreadsheet.
We can use Excel formulas to concatenate (or join, using the “&” operator) into rich, ChatGPT-ready prompts. For example:
Note that your formula may require one or more “IF” statements. This is because your data may have holes in some areas.
For example, some products may not have certain parameters (data within certain columns) specified. You need your formula to be flexible and you can always ask ChatGPT to help you write the formula.
Once your formula returns a request for each row (in this case, for each product), you can copy and paste some of the generated prompts into a word processor, even into Notepad.
It’s good to check a few to make sure the text makes sense, even when some data is missing.
Once you’ve verified that your Excel (or Google Sheets) formula generates the types of requests you want, you can send some to ChatGPT (manually, via the web interface) to see if you like the results.
Generated snippets will likely require human editorial oversight, though you want the AI to do as much of the work as possible. That’s why we engaged in such a deep “rapid crafting” process.
Are you satisfied with your initial questions and answers? Well, then it’s time to move on.
Get the daily search newsletter marketers trust.
Getting your new product content snippets from OpenAI
So now you have a list of products (or other types of web pages) that you would like to generate content for.
In this example, let’s go with a dummy sample of 100 products. Now you have a list of all your products (separated by URL, SKU, or some other unique identifier).
These products also have assigned rich prompts that you’ve generated. But ChatGPT’s web UI is limited. So how can you send them all at once?
For this, you’ll need to familiarize yourself with basic scripting and API request handling. You can create an OpenAI API account to access the ChatGPT web interface.
I’ve created a basic Python script for my agency. While I can’t share the script, I can review some of the processes and necessary documentation.
If you want to syndicate this script later, you’d better build it with marketing-accessible endpoints and technologies. As such, I first produced an Excel sheet:
The sheet just provides an area to dump items for processing (identified by some unique identifier in the “Item Name” column, in this case, the product name). In addition, requests to be processed can also be placed here.
Another tab contains the parameter settings for the request. (You can learn about all of this via the OpenAI documentation.)
Some of these settings refine the granting of content creativity, the deployment of unusual texts, maximum token spending per request, and even content redundancy. It’s also where the OpenAI API key is saved.
Once a certain button in the spreadsheet is clicked, the Python script is automatically launched and takes care of the rest:
First, the script defines the request/endpoint URL. After that, the script sends the request headers and request data.
Most request header/data parameters can be modified within the spreadsheet shown above.
Finally, the response text is received from OpenAI and logged into the “data dump”, another separate spreadsheet.
I have three scripts for this deployment, although only one needs to be run. I also have two separate spreadsheets, both required.
After the script resolves all queries, all text fragments will be saved here:
If you look at the output above, you might have some concerns about content uniqueness.
Although all excerpts begin with the exact phrase (“Introducing the [product name]”), the content produced becomes more diverse in the generated paragraphs. So it’s not as bad as it seems.
Additionally, there are things you can do to try to make each generated snippet even more unique, such as categorically asking the AI to generate unique content (although you have to be pretty firm and repetitive about this to get anywhere ).
You can also adjust the temperature and frequency settings to adjust the creativity of the content and avoid redundant language.
By combining these technologies (OpenAI API, Excel, Python), we can quickly check the generated text fragments for all input requests.
From there, it’s up to you what you want to do with this newly processed data.
I highly recommend moving it to a format your editorial team can understand.
We’ve mitigated quite a bit of this by crafting very rich prompts. However, you can never be sure until you check the output.
ChatGPT Exit Notes
Assuming you’re happy to work with ChatGPT, there is one A few things to keep in mind:
Let’s talk about the cost. It is difficult to give a cost breakdown for using OpenAI’s ChatGPT GPT-4 model via their API. It’s not just the indicator’s input word count or output word count. The price revolves around the “thinking time” of the AI. More complex requests will use more tokens and cost more (even if the number of input/output words is reduced). Our test batch of 100 sample data prompts only cost us $1.74 to run and return. We generated 22,482 words of content in total. 22,482 words of content for $1.74 sounds good, but there’s a lot more to consider. Due to AI’s propensity to infer, a human editorial process is still fundamentally required (in our opinion). However, using this technology transforms an expensive task of creating content from scratch into a much more cost-effective task of editing content. The data/AI specialist’s time to build and execute scripts should also be considered. In addition to inferring where data is missing, AI can also “creatively infer” things. In our sample dataset, the AI decided to infer the existence of a size guide (clothing) within the produced product content. If no size guide existed, it would look pretty silly. Always submit AI content through a human editorial review process for fact-checking, accuracy, and (most importantly) additional creative flair. You can further automate ChatGPT by connecting projects like Auto-GPT. These AI “agents” add more active processing and task power to ChatGPT. However, projects like this still need your OpenAI API key. And because of their infancy, they can chew through a lot of credits before they learn to perform tasks in a standard way.
Extend your content creation process with AI
AI can scaleably produce multiple fit-for-purpose pieces of content with minimal intervention.
For long-form content, it’s probably still better to use the interface and repeat the AI responses.
The views expressed in this article are those of the guest author and not necessarily Search Engine Land. Staff authors are listed here.
[ad_2]
Source link
