Google recently addressed some questions from the SEO community concerning neural matching and its role in search functionality.
Danny Sullivan, through the Google Search Liaison account, provided a series of tweets that clarified the distinctions between neural matching and RankBrain.
Here is a summary of Sullivan’s insights:
### How Google Uses Neural Matching
Neural matching is a technology that helps Google connect words to searches more effectively. This AI-based system has been in operation since 2018.
Sullivan describes neural matching as a “super-synonym” system, illustrating with an example:
“For instance, neural matching allows us to interpret a search for ‘why does my TV look strange’ as being related to the concept of ‘the soap opera effect.’ Consequently, we can display pages about the soap opera effect, even if those exact words aren’t present…”
Back in September 2018, Google mentioned that neural matching was being applied to 30% of searches. While the current extent of its use isn’t specified, it’s reasonable to assume its application has grown.
Sullivan’s explanation of neural matching closely aligns with what my colleague Roger Montti discussed last year in an article about Google’s Neural Matching Algorithm.
### What is RankBrain?
RankBrain assists Google in linking pages to concepts, even when those pages lack the specific words from a query.
This AI-based system has been operational since 2016, predating the adoption of neural matching by two years.
Some theories suggested that RankBrain considers user behavior signals, but those have been debunked.
In summary, RankBrain connects pages to concepts, while neural matching links words to searches.
### More Resources
