You will be doubling your workload, essentially starting a brand new website from scratch in the same niche. I feel I need to make that very clear.

You asked about efficiency and that's pretty much the core of what I'm talking about, subdomains are inefficient.

Niche subdirectories are by far the better option.

I can see what you're getting at, but I agree with Lewis-SEO. Given that meta description is not visible on the page any changes (other than perhaps old school spammy) are just noted and displayed. If you were to change something on-page at the same time it may well get re-assessed, but that's not what we were discussing.

Even if it were to trigger a re-assessment all it would be doing is reviewing that page and putting it in it's "correct" place in the SERPs. Had you not done anything to it, it would still have dropped as soon as it was re-crawled regardless, all initiating a crawl would do is bring forward the inevitable.

Of course if we go on the theory that Google has started using user data; if the meta description changed how the page converted or how people reacted to the page (and you could check your analytics) it could effect your rankings.

Basically if users are now leaving your page in droves because of a misleading meta description that may have a knock on effect.

Whilst I do agree with Daniel Deceuster, subfolders are historically much nicer than subdomains. You're not going to get in trouble for having one on a different IP address. We use a subdomain for our members section (decision was made before my time), customers like that mental partition, and it's not done us any harm.