Well, there's a couple of ways. First, and easiest: copy someone's proven idea. Something you can do, you'd love to to, is proven to make money, and the margins work for you. Bonus: if you can do it better, cheaper, faster. This is the safest route.
Second, pick a problem you, your close family members, or your close friends have. A problem that's painful enough that they'd pay for, interesting (for you) enough that you can spend a few years solving, and practical enough that you can do it with the means you have today (i.e no investment). If it's a "new" idea, it's a lot of risk.
For most people, a small, niche, proven idea that makes money is good enough. It'll take a few years to try a bunch of experiments and see money, though. Grind it out...
Yup. Business problems only rear their ugly heads after you've immersed yourself in the daily work AND if you reflect critically on what you're doing. This usually takes time because when you first start working you usually don't even know what you know or don't know.
I did that a few weeks ago, but I think the "one email a day" format is really bad for such a thing, it's just too slow to find something that suites you.
It's probably a bit like dating.
I'm using dating websites since 2002 now, leaving one, joining another, using about 3-5 simultaneously.
I find like... 4 interesting women a year and only 1 in 2 years finds me interesting.
And most women I had relations with, I met outside of these portals.
-which business will vanish and how will today's customer migrate
-What are the currently growing company's hiring, and what is amiss to replace this jobs, yet again?
-What is missing in modern life, and how could a app re-engineer social-life and society to provide it?
The last one is the most noble, but also the most tricky.
I've been messing with Mechanical Turk a bit, getting a feel for the platform, as I have some ideas I could use it for. When I was in worker mode yesterday I saw a task that asked the question that oppsdaily seems to be sending out to people: "What problem do you face at work that software might be able to fix?"
There is no one way to do this. Everyone does it differently.
Keep in mind that finding a thing you can monetize does not mean it is a thing you would be good at or that you would want to work on for years. Yet, those two pieces are critical to the equation.
Identify your own problems. The reason is you have a much better understanding of what you and therefore your target user would want. Fixing someone elses problem requires working very closely with that person.
I don't think you're going to find an exact, repeatable, cookie-cutter approach to this, because if there was one, somebody would already be using it (and they probably wouldn't share). But while there isn't necessarily an answer, there are plenty of answers to be had, some of them trivially obvious.
So break it down... who do you think has problems? Well, businesses for one. So think about businesses. What do businesses need? Well, they usually need more customers. Or maybe they need to reduce costs. Or both. So think about how the technology you're familiar with could be used to help a business find more customers, or operate more efficiently.
Of course doing this in detail is going to be easier if it's a domain where you have personal experience, but if you don't, just come up with an idea, and then go talk to people about it. If you feel really strongly about it, maybe build a prototype to show off. But be careful of spending tons of time building something before you know if anybody wants it (note: I haven't always followed my own advice here. Also, never take advice from me.)
Another element is: read books. Lots of books. Preferably books about business (sales, marketing, promotion, operations, organization design, strategy, etc.) This will help give you the understanding needed to link technology with business problems like "find more customers" or "reduce costs". What you read in books will always be somewhat non-specific though, so you have to - again - loop back to "talk to people. Lots of people."
If a particular industry interests you, read up on it specifically. Subscribe to the trade journals in that industry, and go to the conferences and trade shows that people in that industry go to. Talk to people there.
Read The Four Steps To The Epiphany by Steve Blank.
Use LinkedIn to find people to connect with and talk to. Favor having actual conversations with people over doing surveys using SurveyMonkey or the like.
If you do enough of all this, at some point, you'll probably come up with a pretty good list of possibilities.
Note that while this is pretty simple, it's not easy. People won't return your phone calls or emails, or will agree to meet you and then not show up. You'll come up with what you think is a great idea, then start looking around and find that 375235028372512.7 other companies are already doing something in that space. Or you'll fall in love with one of your ideas too early, spend a ton of time building it, and then find out that that A. nobody wants it AND B. 375235028372512.9 other companies already built something similar. Etc., etc., etc. Don't get discouraged, just keep plugging. Read this essay by pg: www.paulgraham.com/die.html
Second, pick a problem you, your close family members, or your close friends have. A problem that's painful enough that they'd pay for, interesting (for you) enough that you can spend a few years solving, and practical enough that you can do it with the means you have today (i.e no investment). If it's a "new" idea, it's a lot of risk.
For most people, a small, niche, proven idea that makes money is good enough. It'll take a few years to try a bunch of experiments and see money, though. Grind it out...