This is the whole reason I named the post 'an implementation'.

This particular pattern has worked GREAT for us in the product we are building out as it allows us to roll back, in case any of subscribers failing to do what they need. Both for anything that happens as part of an outbox pattern (where we queue jobs that run asynchronously and are stored db side), but also changes to other aggregates. Which has proved very valueable to us.

But as with anything in software engineering, your milage may vary, depending on the problem space at hand.

Thanks for the comment though!