I think the article's author is missing a key point - there are equally valid reasons for taking 6 weeks. Say, it is company policy and Feedjit is not high-profile enough to warrant going against a policy that errs on the side of caution.

Or it could be the case that, yes, it does take 6 weeks to update the database without doing some kind of emergency, invasive fix. But you know what? If you wanted instantaneous fixes for "obvious" mistakes, you are either going to need a damn good definition of "obvious" and be prepared to still handle complaints when it doesn't work properly, or you are going to need to have some person dedicated to this issue.