When writing an essay on globalisation, I had to compare arguments that addressed different questions. Some explained overall gains, others examined the experiences of different groups, and others asked how proposed improvements would be put into practice.
This made me more careful about the order of an argument: first identify what is being evaluated, then ask whether the evidence supports the conclusion. I find that distinction useful when thinking about projects too.
What does the overall result tell us?
Evidence about an overall change may tell us little about any one person's experience or whether a particular measure worked.
In writing, I want to check whether a conclusion asks too much of its evidence. A case can help a reader understand a problem without having to explain every situation on its own.
How are different people affected?
Alongside the overall result, there are questions about who gains more options and who has to adjust. The same change can show up differently in people's work and lives.
Comparing cases can make those differences visible. A contrast, however, does not establish causation. If two things happen at the same time, their relationship still needs explaining.
How will the proposal be carried out?
Once a recommendation is on paper, who is responsible for putting it into practice? What steps are involved, and what conditions need to be in place?
I pay particular attention to the difference between support being available and someone actually receiving it. The first describes an arrangement; the second involves a process. Leaving out that process makes it harder to judge whether the proposal has helped.
Similar questions apply to business improvements. If something is described as “more efficient”, which step has become faster? How has other people's work changed? Who is responsible for the new process?
One thing I take from research writing is the habit of keeping questions, evidence and conclusions connected. Being precise about what we are evaluating makes it easier to see what we still need to find out.