This calculation is not "the market is forecasted to be $x million/billion and 1-10% of that is $x million/billion" nor does it have to be
an overly precise calculation initially.

It is one that requires more emphasis on how you calculate it vs. the end result as you evaluate a market opportunity.  If you apply some rigor,
you will have a better overall indicator of the size of the market and how to reach it.  In a way, it reminds me of the case questions from
consulting interviews where you were asked questions like "calculate the number of plumbers in the US."

The answer is not so much the point
as much as how you calculate your answer isolating out known inputs and assumptions.
Your addressable market analysis can be the first step in segmenting and understanding your market with your level of analysis influencing your marketing messaging, your sales approach (direct/indirect), and what the whole solution looks like from your target
market's perspective.

Let's say I want to try to size the total addressable market for CRM-based SaaS integration:

• Input:    Subscription CRM revenue is $1.4B in 2008, $1.9B in 2009, and $2.3B in 2010 (according to the folks at Forrester Research)
• Assumption:    "Subscription CRM revenue" = SaaS delivery models
• Assumption:    Average CRM customer deal size = $50k per year
• Calculated:    Today there are 28,000 subscription-based CRM customers
• Calculated:    That grows to 46,000 by 2010
• Input:    75% of SaaS will require at least 5 integrations to existing systems by 2010 (according to Gartner Group)
• Calculated:    SaaS CRM customers in need of significant integration by 2010 = 34,500
• Assumption:    Percent of SaaS CRM customers seeking to purchase a solution to this problem = 50%
• Assumption:    Average customer pays $10,000 to solve SaaS integration need

Addressable global market in 2010:  17,250 companies X $10,000 = $172.5MM

This is the revenue available to the companies competing for it in 2010 – both products and consulting services.  Your company will
get some portion of this amount.

Could the assumptions be wrong?  Yes.  Can you use this basic model to run various scenarios?  Yes.  Should you refine and improve the
accuracy of calculations like this as you enter and gain experience in a market?  Absolutely.

Designing, building, and launching a new product or service is both complex and exhilarating.  I believe one of the keys to doing this successfully is to write a narrative about how your target user uses your product/service in the context of a day in their life.  I can't claim credit for this approach as I learned it from the guy who gave me my first shot as a product manager (thanks again Jim). 

This sounds easier than it is but thinking through what your target does before, during, and after using your product/service forces essential features to the forefront, surfaces substitutes, flushes out bad assumptions, and adds a persona to your internal engineering and marketing meetings.  Instead of "the user does this or that" you discuss how "Jim does this or that."  This also further crystallizes thinking on target markets and go-to-market approaches.

If you can't tell the product/service story through the eyes of a target persona, you're not as focused as you need to be.  Start with what they do first thing in the morning and describe everything they do in a given day until they go to bed working your product/service into the story.  Give it a try.  I promise it will be a revealing exercise.