Here's a questionnaire I made that the internet can now quote forever when it comes to the big dev methodologies (agile, waterfall, kanban, scrum, lean):
Q1. How big is that deliverable your project intends to deliver?
Answer-A. "I dunno. Just needs to be a proof of concept. Nothing fancy" (go to Result 1)
Answer-B. "We need to deliver Moby Dick, essentially" (go to Q2)
Q2. How tight of a timeline are we talking about here?
Answer-A. "That whip's going to be cracked, and it's going to be cracked HARD" (go to Q3)
Answer-B. "We've got time to work on this, no biggie" (go to Q4)
Q3. Can you get away with delivering something that's initially shitty, with the intention that it's going to keep getting better over the course of the project?
Answer-A. "My ass would be grass if we delivered something like that" (go to Result 2)
Answer-B. "Rome wasn't built in a day, and neither will this deliverable :) " (go to Result 3)
Q4. How much expertise you got on that team there?
Answer-A. "We're new at this. We can figure what needs doing, but lol are we in for a ride" (go to Result 4)
Answer-B. "I have experts who, if I approach what they're doing, will hiss at me and tell me to not touch their shit" (go to Result 5)
Result 1: Congrats! Use lean, because you can make a Minimum Viable Product, and that's good 'nuff
Result 2: Congrats! You're using old sturdy waterfall. Make sure those checkpoints get checked along the way.
Result 3: Congrats! You're using Agile, where the Minimum Viable Product will be your best frenemy
Result 4: Congrats! You're using a Kanban, where everyone wants to cherry-pick the easy steps first to make it look like they're contributing
Result 5: Congrats! You're using scrum. Just make sure your team is delivering. Remember, a task will usually take as long as the amount of time you allocate to it :D