Strategic planning is an umbrella term under which companies do a LOT of things. There are no right or wrong ways to do it (excluding obviously dumb ways like using astrology or tea-leaf reading), there are only appropriate and inappropriate ways of applying various conceptual approaches and analysis methods, and appropriate/inappropriate people to be charged with such tasks.

You seem to be thinking in terms of some CEO on a retreat with a few top reports, smoking something potent and charting the future course of the company with brilliant visionary insights powering the thinking.

If this is what you mean, Eisenhower had the best tweet-sized answer:

Plans are nothing, planning is everything.

Marc Andreessen has the best short answer I've seen to date:

The process of planning is very valuable, for forcing you to think hard about what you are doing, but the actual plan that results from it is probably useless.

I quote both Eisenhower and him very frequently, and also have several pages in my book devoted to analyzing this notion (warning, shilling ahead), http://tempobook.com

If I attempted a definition based on my own ideas in the book, which are based on narrative models of decision-making, I'd offer this:

Strategic planning is the process of imagining different scenarios in search of robust unfair advantages to invest in.

These 3 definitions all map to a particular formal analysis technique in the strategic planner's toolkit called scenario analysis. It's been around since the 60s and Dan Rasmus is one of the modern practitioners (he did a lot of it for Microsoft and has written a couple of books about it). I am personally guarded about the tool. It was much more valuable in the 50s and 60s when predictably growing markets made its use much more tractable and there was a lot more good data to fuel it. There are tools you can buy to help do this stuff, such as Crystal Ball, which works with Excel. Today, for most strategic questions, I rely more on qualitative/narrative analysis than data, simply because the reliable data is usually available to all, and privileged data is both expensive to obtain and much less reliable.

But more broadly, if you look beyond CEO-and-buddies-at-a-retreat exercises, "strategic planning" as a business function (i.e. in large companies, usually the main job of a staff function with its own reporting hierarchy with the CSO or VP Strategy at the top), the term covers a bunch of different activities. None of this is "wrong" or "right." It's all a question of which techniques and organizational practices are appropriate for a given company in a given time/place/market. Common activities include:

1. Everything Michael Porter teaches MBA students to do (SWOT, 5-forces, structural modeling, competitive analysis)
2. Strategy Operations, a term generally covering collection and maintenance of data that provides relevant input to the Strategy Planning function.
3. Evaluation and selective adoption/use of all business flavors of the month as they come and go. Some, like experience curves or disruption theory or Moore's "crossing the chasm" models, have been with us for decades and continue to be valuable. Others, such as Magic Quadrant or Hype Cycle, wax and wane in popularity, partly thanks to the firms that promote them. Still others, thankfully, are short-lived fads, like Blue Ocean.
4. Short-term, heavy-lift activities like absorbing a major acquisition, spending a large M&A budget.
5. For companies with mature analytics capabilities in the sense of Davenport and Harris, analytics is usually a strategic function. It replaces Strategy Operations in a sense, since it produces live firehoses of analyzed, visualized data to inform the strategy planning.
6. Strategic succession planning: in big companies, who does things and when and where is usually more important than what gets done or why. So a lot of strategic planning actually involves how you move people around in roles/responsibilities. Think of it as chess playing with people. In large companies, CEOs correctly spend more time on backend casting decisions than on front-end goal-setting decisions or middleware process-decisions. This aligns with Larry Collins' wise dictum: good people with a bad process will always beat bad people with a good process.

Not all these activities deserve the adjective "strategic" but these are the things that have traditionally been put in the strategy bucket.

I'll specialize the general model above for big companies.
In big companies, the essence of strategic planning from those more abstract definitions earlier, pretty much boils down to "get the right people on the bus, the wrong people off the bus, and then decide where to go." This is the Larry Collins model in Good to Great.

Eisenhower had a great version of this too: Leadership is the art of getting someone else to do something you want done because he wants to do it.

I'll add a note for startups, since there are so many entrepreneurial types around here. For startups, all strategic planning pretty much boils down to "look for PMF by any means necessary."

I was introduced to Marc Andressen's thinking through his classic The only thing that matters, where he popularized the PMF idea. It fits in perfectly with his view of planning in the quote above.

http://pmarca-archive.posterous....

If you are in startup-land, and you read only one thing about strategy, the above link would be it.

If you are in a big company and are the practical type, I'd recommend Walter Kiechel's Lords of Strategy to get the basic background, and Dan Rasmus' book Management by Design.

If you like philosophical approaches/meta-cognition, I'd recommend my own book.