I would definitely recommend a research notebook.

There are several different ways you can go about it.

• The old school method of just using a paper notebook, where you enter data and computations sequentially by date. One is often encouraged to start a new page if one is moving to a new topic (and label the topic at the top of the page); but another possibility is to make use of the margins for noting the topic at hand.
• An electronic notebook would be more searchable. On the other end of the spectrum from a paper notebook is a lab wiki. Some universities even have a university-wide Wiki platform available (for example, here's EPFL's wiki portal). In those cases you won't have to worry about administration issues, and many of those services are regularly backed up, and come with access control system so that you can limit your lab notebook to be only viewable by those in your research group.
• An intermediate method is to just keep a private electronic notebook. I use a customised document class that I wrote for this specific purpose. If you use a good indexing package in LaTeX and make the effort to keep good indices, such a notebook can be very easily searchable. It also has the advantage that when preparing lecture notes or papers for submission, you can just copy and paste directly from your notebook.

If you prefer not to have a single notebook, what you'd need then is a sophisticated document managing system. For paper documents this will generally involve a filing cabinet, folders, and sticky labels. For electronic documents (say you digitize all your notes either by typing them up or scanning them), a lot of the citation managers, especially those that support multiple databases, can easily be co-opted for organising notes. On the even fancier side, you may want to use some sort of mind-mapping software.

Another option if you do not mind "showing how the sausage is made" is to follow the initiative of the Open Science Project and blog about your research as you go along.

I would definitely recommend a research notebook.

There are several different ways you can go about it.

• The old school method of just using a paper notebook, where you enter data and computations sequentially by date. One is often encouraged to start a new page if one is moving to a new topic (and label the topic at the top of the page); but another possibility is to make use of the margins for noting the topic at hand.
• An electronic notebook would be more searchable. On the other end of the spectrum from a paper notebook is a lab wiki. Some universities even have a university-wide Wiki platform available (for example, here's EPFL's wiki portal). In those cases you won't have to worry about administration issues, and many of those services are regularly backed up, and come with access control system so that you can limit your lab notebook to be only viewable by those in your research group.
• An intermediate method is to just keep a private electronic notebook. I use a customised document class that I wrote for this specific purpose. If you use a good indexing package in LaTeX and make the effort to keep good indices, such a notebook can be very easily searchable. It also has the advantage that when preparing lecture notes or papers for submission, you can just copy and paste directly from your notebook.

If you prefer not to have a single notebook, what you'd need then is a sophisticated document managing system. For paper documents this will generally involve a filing cabinet, folders, and sticky labels. For electronic documents (say you digitize all your notes either by typing them up or scanning them), a lot of the citation managers, especially those that support multiple databases, can easily be co-opted for organising notes. On the even fancier side, you may want to use some sort of mind-mapping software.

Another option if you do not mind "showing how the sausage is made" is to follow the initiative of the Open Science Project and blog about your research as you go along.