For those who explored Google FeedGen some months ago and didn’t see the expected improvements, I highly recommend giving it another try. Significant enhancements have been implemented recently. The necessity for high-quality feed inputs is a thing of the past; FeedGen now has capabilities to extract data directly from web pages and images. Below is a summary of the latest updates, including an anticipated new feature:
January 2024
Added support for fetching product web page information and using it for higher quality title and description generation.
December 2023
♊ Added support for Gemini models (gemini-pro and gemini-pro-vision).
✅ Unified description generation and validation – now handled by a single prompt.
Added support for image understanding for higher quality title and description generation (only available with gemini-pro-vision).
Added LLM-generated titles which should avoid duplicate values at the possible loss of some attribute information.
What to expect in near future?
Soon to launch: Product categorization support. This will leverage product feed, web page, and image data to refine product categorization, boosting visibility in Google Shopping search results.
I’ve personally leveraged FeedGen extensively and am now offering tailored courses for agencies and in-house teams eager to master its application. Having successfully conducted our inaugural session yesterday, more training sessions are scheduled this month. Interested in elevating your team’s proficiency? Reach out for details!
