Obviously, to keep the discussion useful/generic, you can't really promote the site, nor can an editor defend the action. And since you've not mentioned what business you're in (thoughtfully -- and thank you for following the forum policies conscientiously!) whatever we think we might know about whatever we might guess it is, can't be relevant.
But I can say this: some business websites deliberately implement an information flow model that makes them unlistable. The websites are unlistable, not because the editors DON'T understand the business, but because the editors DO understand the business. (And that's OK: organizations build websites for their own defined purposes, and the ODP sponsor allows volunteers to help with its own defined purpose. No rule says there has to be any correlation between their purposes.)
All we can really have here is a discussion of the form, "here are some generic circumstances" ... "then in such a situation, here are some considerations that should apply, or may apply." We can't even say whether those considerations were applied--or ignored--appropriately in any specific case. All we can say is that there are internal forums, where specific cases are discussed by any interested editors.
You've mentioned a number of statistics. A business webmaster is naturally going to be interested in statistics. (I'm not a business webmaster, but I'm still interested in statistics about other people's interest in content that matters to me.) But guess what? I can't think of a single statistic that has anything at all to do with listability. I've listed sites that consisted of several pages at most; or a single article; sites whose counter mentioned less than 200 visitors; sites that may not even appear in Google search results for all I know; sites a few days or weeks old; sites with exactly one content creator; sites with hideous backgrounds and atrocious layout; sites ten years old; free sites with random ad banners on every page And I've passed over sites with hundreds of thousands of pages, tens of thousands of visitors, top Google positions, beautiful backgrounds and crystal-clear CSS, professionally designed for commercial purposes with domain name subnet and dedicated server farm. (And, of course, vice versa.) Because none of that actually MATTERS.
So if statistics don't motivate a site review, what does?
I shouldn't have thought it was that hard for a webmaster to say something that inspired me to want to ask what the URL was, so I could visit it for my own curiosity. It's not impossible, but it certainly happens less than 1 in 100 webmaster posts. And, you know what? The last two times it happened, before I actually contacted the poster, I read on: and the post eventually mentioned "it's already in the ODP."
Well, two data points don't prove anything. (And, if they did, barring confidentiality, I could certainly name two editors--even two senior editors--who had inappropriately deleted sites.) Each data point has to be evaluated on its own.
So where's the responsibility? It's to the community: to anyone who's able and willing to set aside other interests and spend time working on the website's defined purpose. Any edit can be reviewed and discussed by anyone in that community. And anyone can ask for an editing action to be reviewed by a senior editor.
Most of the time, an editing action will be confirmed. Sometimes, an editing action will even be carried through more consistently. And sometimes it will be undone. Sometimes the editing guidelines will be tweaked to prevent anything like it from happening again.
But the webmaster's business model (or other kind of proprietary interest) and the community's interest are often disjoint or even contradictory. And that's OK. There are lots of websites, each one with a different purpose.