The guardian's Election Poll Projection collates all available polls of note and attempts to draw conclusions with regard to how many seats each party will win come election day. UK polling isn't anywhere near as spot on as in the States, with constituency polling limited at best. That said, it's still possible to build a broad idea of the impact of voting intentions, debates and so on.
The biggest surprise in this elections polling so far, is how it's been stuck in deadlock territory for so long. In fact so much so that at time of posting Labour and the Conservatives are both projected to win 271 seats. Unless there is a major shift before the election, or the projections are way out, a little movement this way or that in the polls seems largely academic as neither party will be able to secure a majority.
With that in mind, it all comes down to who is willing to deal with whom to secure either a coalition government or alternatively vote by vote agreements under a minority government. Should current projects be accurate it's very difficult to see how Tories would cobble together a coalition to begin with. Even an ambitious/fanciful pact with the Liberal Democrats + UKIP + DUP wouldn't get them over the line. As for Labour, an SNP coalition would likely comfortably see them gain a majority, though politically that move might not be the most popular in England. A Labour minority government operating vote by vote with the SNP could feasibly work though.
Who out of the Tories and Labour gain the most seats could at least symbolically impact the likelihood of this outcome or that. Currently the bookmakers have a Labour minority government at 2-1, a Conservative minority at 3-1 and a repeat of the Tory-Lib Dem coalition at 7-1. Whoever out of Labour or Conservatives claims the authority to form a government will need to get their Queen' Speech and Budget voted through first. It could get messy.