As with the election polls, there's no point, it skews the results as they won't be counted in the real referendum anyway
Undecideds, like me, will be counted because, one way or the other, I'll vote. I'm just not finalising that decision quite yet, when I don't have a final picture on either what I'd be voting to stay in, or as clear a view as possible on what 'out' looks like.
Not having that option leaves the result even of a poll like this less clear than it might be.
Suppose you get 56 vote leave, 54 vote stay and 3 undecideds. That suggests that, in as far as this poll can ever be representative, it's narrow out result, with 'in' barely within grasp.
If, on the other hand, you get 56 voting leave, 54 remaining, and 1000 undecideds, it suggests the vote is absolutely wide open.
Nearly all general elections in modern UK history have actually been decided by a modest-sized floating group of voters. Many have voted Labour all their lives, many Tory all their lives and they cancel each other out. And another chunk vote, fairly consistently, for smaller parties and aren't likely to change, be it LibDem, UKIP, Green, SNP, various NI parties, etc.
Elections are won and lost by floating voters floating. A 6-7% swing has often been enough to see one party lose power and another gain it. And the majority of those swing votes consistently come from floaters, from undecideds, not from the party faihful. That's why some constituencies are ass-deep in visiting MPs come election time, and others are more likely to see political tumbleweed rolling down the street than an actual politician.
Similarly, all the signs are the referendum has hardcore in, and hardcore out, for whom nothing is going to change their minds. This referendum will be determined by which way the undecideds eventually decide.