Great breakdown as always, the depth is real. Engagement may be more of an ancillary signal than a primary one. I see this across clients as well, low to zero engagement and still getting cited. My AI Visibility subreddit https://www.reddit.com/r/SEO2AI_Bridge is another example, cited across all the platforms you mentioned with virtually no engagement driving it.
Much weight comes down to platform structure and machine learnable (structured) content. What is posted and its structure appears to carry significant weight. Simple things like word sequence, avoiding absolutes like "best" and "top," and steering clear of unverified claims matter too. Overuse of those tends to dilute the signal.
Fair point on the layered distinction - and you’re right that structure carries real weight. The way content is constructed, the language choices, the avoidance of superlatives and unverifiable claims, that’s a layer most practitioners haven’t gotten to yet. The engagement signal argument holds more strongly at the training data curation tier than at retrieval, so the two observations aren’t necessarily in conflict - just describing different parts of the pipeline. Appreciate you adding to the thread.
Personally and professionally, I’ve found that most subject matter experts (SMUs) don’t have the time or inclination to do any or most types of social media.
And if they do hire a social media company or group, they are rather substandard substitutes for actual SMUs. Fake reviews are fake reviews, popular or not.
---
Hey Duane,
Great breakdown as always, the depth is real. Engagement may be more of an ancillary signal than a primary one. I see this across clients as well, low to zero engagement and still getting cited. My AI Visibility subreddit https://www.reddit.com/r/SEO2AI_Bridge is another example, cited across all the platforms you mentioned with virtually no engagement driving it.
Much weight comes down to platform structure and machine learnable (structured) content. What is posted and its structure appears to carry significant weight. Simple things like word sequence, avoiding absolutes like "best" and "top," and steering clear of unverified claims matter too. Overuse of those tends to dilute the signal.
Fair point on the layered distinction - and you’re right that structure carries real weight. The way content is constructed, the language choices, the avoidance of superlatives and unverifiable claims, that’s a layer most practitioners haven’t gotten to yet. The engagement signal argument holds more strongly at the training data curation tier than at retrieval, so the two observations aren’t necessarily in conflict - just describing different parts of the pipeline. Appreciate you adding to the thread.
You nailed it with the retrieval layer. Its a different beast/layer. You're one of the very few that actually really get it. Your on it!
Joe
Thanks Joseph! 🙏🙏
Context moat.🤣🤣🤣🤣🤣
Personally and professionally, I’ve found that most subject matter experts (SMUs) don’t have the time or inclination to do any or most types of social media.
And if they do hire a social media company or group, they are rather substandard substitutes for actual SMUs. Fake reviews are fake reviews, popular or not.