by Matt Zimmerman
A discom… whatsit? Huh?
A discombobulator silly.
It discombobulates things of course!
Here’s how it works.
You’ll input a list of queries (or blog post titles) that you have already SERP scraped (or maybe you want SERP scraped). Next, you’ll load up some custom prompts, which are a weee~ bit different than the custom prompts you’re used to. Finally, you’ll set some additional options and press the button to create a discombobulated CSV file.
What happens when ZimmWriter starts discombobulating?
ZimmWriter takes the merged summary for each SERP scrape (which usually consists of a combined summary from five URLs) and performs your custom prompt on that merged summary. You can use it to extract facts, build your own meta description, your title, or just about anything you want.
List your search queries here. Usually, these will be blog post titles that you SERP scraped in the Bulk Writer or SEO Writer. Just put each query (or title) on its own line.
These are custom prompts that you created in the Bulk Writer, SEO Writer, or Penny Arcade. But… and this is a BIG butt… the prompt I feed to OpenAI is different than normal. Here is the system prompt:
[DATA TO PROCESS]: the merged summaries from up to five URLs from the SERP scrape [INSTRUCTIONS]: your custom prompt, which should probably refer to DATA in all caps, or DATA TO PROCESS in all caps. [RESULT]:
If you check the custom prompt page, you’ll notice that the formatting is different than in the discombobulator.
The reason is because the discombobulator doesn’t rewrite the text. Instead, it’s designed for using the summary as a data resource that you want to extract something from to create something new.
I highly recommend testing your prompts in the playground. You can easily get the summaries to feed into the prompt above by:
As a default, each prompt operates on the merged summary. But when you enable prompt stacking, then prompt 2 will work based off of prompt 1 (and not the merged summary).
As always, your titles and prompts need to be in English. However, you can enter a non-English language in this box and ZimmWriter will translate the output in the CSV file.
Note that when you enable prompt stacking, only prompt 2 is translated.
The name to give to your CSV file you want ZimmWriter to create.
The GPT model that you want applied on the prompts. The SERP scraping is always performed using GPT 3.5 regardless of your selections.
If you want to use the discombobulator on a non-google.com domain, then perform the scrape first in the SEO writer or the Bulk Writer (use 1 heading, no conclusion, nice cheap and simple). Then load up the same title in the discombobulator and it will pull the result from the cache.
The information provided on this website is provided for entertainment purposes only. I make no representations or warranties of any kind, expressed or implied, about the completeness, accuracy, adequacy, legality, usefulness, reliability, suitability, or availability of the information, or about anything else. Any reliance you place on the information is therefore strictly at your own risk. Read more in my terms of use and privacy policy. You can also contact me with questions.