The downside of spreads involving so few cards is that they sometimes don't offer enough meat to make a meal, if you will. I also think the One Card draw can actually be counter-productive at times because it encourages very specific questions (such as, 'Will my partner and I have a baby this year') which I believe can only be answered definitively by time, not by Tarot. Sometimes it's necessary to elaborate by adding more cards to the spread, either when the querent still feels they have questions/confusion or when you feel, as the reader, that you can't glean a strong interpretation from the original cards.
I haven't always believed in the method of adding cards. Recently I've been working with Runes and there are some interesting rules about going no further with a reading if certain Runes come up. They are to be taken as signs that it's not the right time for an answer. Although I think that kind of occurrence is valid, it's not great if you're doing a reading for an excited family member or a paying customer. You need material!
So, I tend to add cards if the flow becomes stunted before the interpretation is complete, if the querent asks me an additional question or if I strongly feel that a card could mean one of two things. (This last dilemma only tends to come up when I'm doing a reading with three cards or less. When there are so few cards to work with, I feel it's important that I interpret them as clearly as possible.) I take additional cards from the top of the already shuffled deck or ask the querent to cut it but I recently read a technique that involves using two decks - one for the reader and one for the querent. That way the querent can add additional cards when they feel like it.
Using this technique too frequently can muddy the waters. It's better to punctuate key messages with no more than three extra cards per reading, but obviously there's no hard and fast rule.