How to avoid procrastination during the research phase of my PhD?

I have found (and continue to find) two kinds of strategies that have some success in preventing me from procrastinating.

Silly, but scarily effective:

• As silly as this might sound, implement a blocker like Leechblock or Chrome Nanny and block sites that you waste time on. I used to scoff at such things, but once I installed them, I found myself being forced to spend more time working.
• Along the same lines, disable email alerts. Once I did this, I found that I could spend many hours working without getting distracted by random emails.

Less silly, and also less effective

Much of the procrastination when starting research in grad school comes from the lack of clear structure (as you indicate). So the trick is to create structure. For example, if you're working on one problem, and you've spent a little time generating a few ideas, write them down, and methodically start working through each of them one by one, going as deep as you need to go in order to test out an idea. When you're doing so, try to forget about the larger problem, the context, your Ph.D, your future career, and all other "big picture" matters.

More likely than not, the idea won't pan out (most don't!). In that case, move on to the next one.

The problem with early-stage research is that it doesn't look like you're making any progress, so it's easy to slow down and waste time. But if you can measure progress not in terms of "papers published", but in terms of "ideas tried", you'll get some sense of the amount of effort you've put in, and that can help motivate you to try more ideas, and so on. All these failed ideas are teaching you valuable things about your project, and they will eventually be useful because of what you've learnt.

I write both weekly and daily 'to do' lists if I'm stuck in a procrastination routine. The weekly lists are more general, the daily ones as detailed as possible, with essentially everything I should do that day, however small it seems. It feels good to tick things off - it shows me that I made progress and gives me incentive to go on. Usually there's one or two things on the list that I really should not postpone for tomorrow, and many less important ones. I sometimes do the big stuff first, sometimes last, it really depends on my mood that day. I also allow myself to push the smaller things to the next day if I don't feel like doing them. That way I feel that I have some wiggle room, but it's constrained to the stuff that doesn't matter much anyway.

I also noticed that I tend to be more efficient when I have something planned at the end of the day, i.e. I know have to leave the office at five and stop working until tomorrow. This gives me a feeling of a small deadline, and deadlines are good at eliminating procrastination.

Finally, as someone mentioned already, nothing beats working with people. Having weekly meetings with my supervisor, where a decision is made on what I will achieve until the next meeting, tops it all.

IMHO to keep the pace, the most important thing is not to work alone.

First, when I talk about my project on daily basis (which, sadly, it's not the case now), it naturally makes me working (as I don't want to make others wait, and it would be kind of awkward to say "I did nothing, because I was reading blogs").

Second, often getting stuck happens for a reason - and just 'working harder' does not help. Then it's important to have some other "seed", or a broader perspective, or just to learn that "it is not me who is stupid - this problem is hard for everyone".

But in case you cannot yourself from procrastinating (sometimes I know that it is not a day when I am good at thinking), try do some "white procrastination" - i.e. procrastinate from your duties, but in a way, which is constructive (e.g. learning a new programming language, reading papers in a different discipline, mastering your skills in something else, etc).

EDIT: In a longer run, it turned out that some of "white procrastinations" were more valuable than my "standard stuff". Remember, that not always you know in advance what will be fruitful in the future.