IRAC is the road to median, in my opinion. It's repetitive and unnatural.
What you do first kind of depends on the course and the type of question. For a classic issue spotting torts exam, for example, a good way to start is to write down each party in the question, and then all claims they might bring. Each paragraph should be a different claim (although you can of course break claims into multiple paragraphs if one starts getting too long). This should help you get over the problem of where to start. It's all kind of arbitrary, but just go in the order you wrote them down and cross them off as you go.
So for example: A may bring a claim for Battery against B. [Talk about the elements and whether they are satisfied]. For obvious ones just say that the element is satisfied because X reason (e.g. "harm" where B stabbed A just say that there was harm because B stabbed A; you don't need to get into some kind of esoteric discussion about what "harm" really means). For less obvious elements, you may need to devote several sentences to a back and forth: "A will argue this, B will argue this, the court should find A's argument more persuasive because..." It's important to stay objective and discuss both sides on a question like this, and discuss what each party would say. After you discuss the elements, you should bring up any affirmative defenses, such as self defense, as well as their elements, similar to what you did for the claim itself. After that, just write a quick conclusion that is consistent with your analysis (e.g. "The court will find for B because B acted in self defense"). Rinse and repeat for each claim.
For more policy-based questions, it's a bit less straight forward. Unless time is a serious concern, it can be a good idea to take a step back and find the part of your outline relating to the main point of the question. Look at your notes from that day if you can find them quickly, and see if you can't remember something that your professor said. I'm of the opinion that it's generally a good idea to parrot your professor's views if (s)he made them clear. Even if (s)he did, you still do a back and forth of why a given policy is or is not good and what you think the correct policy is and why (so for instance, whatever the professor's reasoning was).
One important thing to note: different professors have different expectations of how much you should use cases in exams. I had one professor who expected you to site the case and the PAGE NUMBER for the case on the exam for pretty much everything you did. I had one who straight up said "don't mention cases on the exam." I've had others who've said that they're not required, but can be helpful to support your analysis or as a shortcut for analogies (e.g. "this is like XX case where ABC happened"). Make sure you know where your professor stands on cases, and use them accordingly.
ETA: Some professors will tell you to write out what the rule is before doing analysis. I don't like doing that because it's redundant if you do your analysis properly, but if the professor explicitly asks you to do that, then plop it in there before the analysis: "A will bring a claim against B for battery. The elements of battery are XYZ." And then continue as usual.
2807 wrote:
No extra points for verbose flowery nonsense.
That depends on the professor; there are professor who will reward for totally off-topic BS, but you should have an idea by now who they are (if any). Generally I agree though: be concise.