Hey all,
my client needed an qualitative audit on their landing pages advertising. More specifically needed a way to make ads context more targeted an associated with the related landing pages.
The challenge is to audit 4000+ pages promoted by up to 4 different ads, which means 16.000 combinations to be completed every month.
Having a review on the issues and examples where ads copy was not to the point we thought about the following process:
- Scrape all landing pages content & take a screenshot for each page
- Get text from ads and landing pages comparing in quality match using the following fields:
- Target group
- Location
- Language
- Value proposition
To automate the process I used Javascript node to open and get content details from each page:
Depending on the page type we extracted different elements and all those details are saved on
central place where landing pages and ads content is already linked.
Then a separate scenario feeds the context from ads and pages to AI asking for comparison on paragraph’s level and on title’s level.
AI could respond us if the fields had a match or not win full explanation. We also asked AI to return separately the result in boolean to perform automated calculation on the base.
The two scenarios run on demand when a webhook is triggered and processes minimum 4000 lading pages and their associated ads and every month we know our pages’ scoring automatically so we can look on low scoring records and make improvements immediately.
Feel free to post here any questions about the case 