The first thing that you need to do to perform a scenario analysis is to create your cases, or future states of your portfolio, based on assumptions of what the future could be like. These are:
Base-case scenario. In a base-case scenario, you essentially assume nothing much is changing, and things are going to kind of proceed as normal. This is the average, middle-of-the-road case.
Worst-case scenario. In the worst-case scenario, you assume the worst. When you do your calculations, use the least-favorable side of the range for that metric and the lowest realistic expectation of growth.
Best-case scenario. The opposite of the worst-case scenario, the best-case scenario assumes that everything is coming up roses. Use the most favorable side of your metric ranges, assume things will be peachy, really dress the thing up (within reason), knowing that it's incredibly unlikely this will come to pass.
Basic steps in scenario analysis
Your scenario analysis will look different depending on the actual scenarios you choose, but a simple scenario analysis will follow these steps to help you create your cases:
- List the assumptions you want to consider in your scenarios.
 - Determine the upper and lower ranges of those assumptions, such as that your stock's dividend will grow by certain percentages in a specific period of time.
 - Apply statistical modeling to your scenarios, based on the assumptions you've made, to determine the most likely best, worst, and middling outcomes.
 - Use more math to assess the risk involved in these scenarios (there are online tools to help you with formulas for Monte Carlo simulations, sensitivity analyses, decision trees, and many more).
 - Determine how these scenarios fit in with your risk tolerance and investment thesis. If only the best-case scenario is acceptable, it's definitely a good time to reevaluate your future plans. People with a very low risk tolerance should also be able to easily withstand their worst-case scenarios.
 
It's much easier to use a tool like an Excel module programmed to help, rather than trying to do this analysis by hand. Unless you're making very simple assumptions, it will get very dicey very quickly. Make sure all your scenarios are assuming things about the same variable -- for example, the growth of that dividend or the timeline for a business to be profitable.