I had jotted the following headings down on a notepad some days ago when I was in the midst of solving a problem. I immediately found it to have broader applicability, because I observe many peoples' endeavours stall — including my own — for lack of inputs. Furthermore, those that barrel ahead notwithstanding, I often see crash and burn for treating conjecture as fact — for neglecting knowledge that if acquired would have been a cheap and effective hedge against failure.
To encourage responsible and effective behaviour in problem-solving and execution, I saw fit to capture this pattern and elaborate on it.
Before I can attempt to define or often even understand a problem, I must first get the data.
Data do not care about how I get them or where they come from. It is a mistake to be picky about the origin of the data. If they are too messy or incomplete, I can often go back for more.
But, the data do need a home. So before I go get the data, I must make sure there's somewhere to put them — preferably somewhere I'll remember and from which I can retrieve them.
Getting the data is the difference between guessing and knowing, between inference and evidence. If I find myself hesitating, I will ask myself what I expect to learn, and assess the risk of relying on an assumption instead.
Getting the data can be a painstakingly dull and boring task, but at least it's well-defined. The sooner I start it, the sooner I finish, so quit stalling and go get the data.
Once I have the data, I will often need to clean them. This entails repairing noisy records and separating what is useful from what is erroneous or irrelevant. This task has the likelihood of being nearly as dull and boring as getting the data, but thankfully it is also reasonably well-defined.
Closely related to cleaning the data is shaping the data. I can chop them up into the appropriately-sized pieces, or normalize them to make them easier to operate on. It is also at this point that I can begin to see productivity gains from automation.
Before I can operate on the data, they must be logically, and ultimately physically concentrated into the correct place. This may entail acquiring a structure which will either push or pull the data where they need to go.
Here is where I extract whatever I am looking for from the data, possibly in correlation with other data. If I do this correctly, I should now have information.
Once I have my information, I will likely need to make it medium-sized so that my medium-sized brain can absorb it.
I should now package my newly-acquired, medium-sized information in such a way that is appropriate to its shape, and such that it is hard to confuse the package for part of the content. I should also make sure anyone who ought to see it sees it, because at this point I should be in possession of a new and hopefully useful body of knowledge.