Background: I was reading about Google's Hybrid approach to research. This prompted me to think about how in academia I try to get multiple output from the one set of inputs. For example, I might try to
- Present a talk and and then write up a journal article
- Write blog posts and question-and-answer combos on StackExchange to force me to learn something that I need to know in order to write a journal article.
- If I learn a new technique in the process of writing a journal article, then present a tutorial on how to apply the technique.
I've also heard senior academics talk about employing this approach. E.g.,
- publishing an article in a journal as well as reframing the content for a magazine or newspaper
- What are the most important examples of getting multiple outputs for a given input?
- What is a systematic way of incorporating this into your project planning?
Your answer to this question will depend somewhat on your long term goals. Everything you mention: talks, journal articles, blog posts, tutorials, magazine articles, etc. can be worthwhile in the right circumstances. Most of what you describe consists of taking variations on the same theme to different audiences.
In what circles do you most want to be known? More generally, how will you measure success? Say for example that your primary goal is to get tenure. If you're at a high research school, then publishing an expository article for undergraduates may well be viewed (by your tenure committee) as a waste of your time. In contrast, if you're at a liberal arts school, this may be considered just as valuable as getting a research article in a top journal.
You asked about incorporating this into project planning. I suggest that you decide which audiences you are most excited to reach. This will likely be largely independent of the particular project. Then as you start on each project, think about how you can tailor your message to suit each of these audiences. You may be surprised by how often you find you have something valuable to say to each of your target groups.