To solve your problem you need to ask how you can solve it—that is, identify various solutions. But should you do so from the get go?
Ask how to solve your problem only if you already know its root cause(s)
Formulating your problem with a how is attractive because it forces you, right from the start, to think about potential solutions: you are dedicating yourself to finding the various ways in which you can solve it. Soon, you’ll decide which one is the best, implement it, and give yourself a pat on the back.
The problem is that much of your effort can be misplaced thinking how if you don’t know the root cause of your problem.
Consider having to identify how to go from New York to London.
If you treat it as a how, you can classify the means of transportation as air-based or sea-based. Breaking it down further, you’ll find further means for each branch (for air: plane, helicopter and so on). So, you’ll now be considering anything from swimming to taking an airliner. You’ll be exhaustive in your identification of solutions. But why do you want to go to London in the first place?
If your primary motivation is, say, a business meeting, then you’ll probably be most interested in being in London on a specific date and not being tired when you get there. Also, price will probably not be too important. So while an airliner or a private jet will do, going at it swimming or using a submarine won’t. Alternatively, if your goal is to go to London while breaking a world record, using an airliner won’t do but swimming just might.
So, if you don’t know why you’re solving your problem in the first place, jumping straight to a how can be counterproductive (for an example, see here). You’re better off starting with a why analysis to identify the exact scope of your problem, capture this additional information in your problem identification card, and then do a how analysis.