Most deals, for investment or acquisition, happen in two phases. There's an initial phase of negotiation about the big questions. If this succeeds you get a termsheet, so called because it outlines the key terms of a deal. A termsheet is not legally binding, but it is a definite step. It's supposed to mean that a deal is going to happen, once the lawyers work out all the details. In theory these details are minor ones; by definition all the important points are supposed to be covered in the termsheet.

Inexperience and wishful thinking combine to make founders feel that when they have a termsheet, they have a deal. They want there to be a deal; everyone acts like they have a deal; so there must be a deal. But there isn't and may not be for several months. A lot can change for a startup in several months. It's not uncommon for investors and acquirers to get buyer's remorse. So you have to keep pushing, keep selling, all the way to the close. Otherwise all the "minor" details left unspecified in the termsheet will be interpreted to your disadvantage. The other side may even break the deal; if they do that, they'll usually seize on some technicality or claim you misled them, rather than admitting they changed their minds.

It can be hard to keep the pressure on an investor or acquirer all the way to the closing, because the most effective pressure is competition from other investors or acquirers, and these tend to drop away when you get a termsheet. You should try to stay as close friends as you can with these rivals, but the most important thing is just to keep up the momentum in your startup. The investors or acquirers chose you because you seemed hot. Keep doing whatever made you seem hot. Keep releasing new features; keep getting new users; keep getting mentioned in the press and in blogs.