The results were listed as "very good" because 20% of new signups tweeted. However, you're giving away revenue in hopes of generating more long-term revenue. The success or failure of the experiment should be given in terms of customers who signed up from seeing the tweets and what the lifetime value of those customers are, taking into consideration the initial revenue loss.
The product is valued at "a tweet for a free month". Does this support the pricing tier? There's something about the perception of price and brand value here.
On the other hand, you do comment about perception–if the users perceive the product to worth less because of this deal, "a tweet for a free month", the rationale behind why the deal is good for the product might not do anything to offset that loss in perceived value.
That's why you give a user-specific shortened link with UTM values for people to tweet out and don't simply reward them for tweeting your site name or address.