I am immensely skeptical of playing the estimate-size-of-market game because the addressable size of the market (i.e. how many of them you can contact) is the limiting factor for most small businesses.

Anyhow, here's how I did it for my company back in 2006. I used two methods: top down and bottom up. Both involve a whole lot of hand waving.

http://www.kalzumeus.com/2006/06/29/guestimating-market-size...

Of the two these days I'd be much more inclined to use the bottom up approach, but either way these estimates are so laughably inaccurate I don't think they have that much utility. Example: the bottom-up approach predicted 2k searchers for my product in a month. I now personally achieve roughly that much in a day. Market size, shmarket size.

(Instead, I prefer estimation by axiom: if you can sell it to anyone, you can sell enough of it to matter, when multiplied over the entire Internet.)

I'm going to suggest something crazy. You probably won't be able to estimate the size of your market using the internet. Go to a research library where you can read and copy paid industry and market research reports for free (the library pays the fees.) Bring a USB key and you can download the reports (at least in SIBL in NYC.)

I would suggest you do some old fashioned market research by looking at the age/sex/demographics/industry of your target customers and reading market research reports on them and industry reports on your competitors and other replacement products. A replacement product is one that isn't a direct competitor but people can use instead of your product (for example: Gmail and Google calendar is not a direct competitor of a CRM system like Salesforce.com, but people can and do use it instead of a CRM system.)

I think you could do all of this with 1 day spent in the library. You can find the same information on the internet, but it will take you a lot longer and you would have to pay for many of the same reports you could get from the research library for free.

I've used Google's Keyword Tool (https://adwords.google.com/select/KeywordToolExternal) to estimate the number of people searching for keywords that might indicate they'd be a customer; however, you should be careful about any conclusions you draw unless you're keywords are very specific.

Also, Google trends (http://www.google.com/trends) can give a more impressionistic idea of whether a market, as judged by search requests, is growing.

if you don't know it, Google Search Insights is much more powerful and gives better estimates

I suspect most approaches (barring methods so time consuming your spending will be out of control before you even start) aren't going to give you something something remotely accurate (or useful for your question). Having said that, there might be unexpected side-benefits to research though, like finding out exactly which features people are key to people downloading some competitor programs and neglecting others.

It sounds as though you want to estimate how much un-addressed market potential there is (otherwise you would be taking the alternative approach of estimating the resources needed so you make a product that's flat out better or better value for money than your rivals, thus competing for market share in a tapped-out market).

But don't count on it or waste too much time - why not figure out what maximum resources you would put into this based on a tiny chance of success, release that first attempt, and improve on it as an when either it starts gaining traction (suggesting real demand and possible success in future if you go through with this) or when new resources open up to you that you don't mind risking on this to take it a little further?

You could run a test with Adwords:

http://news.ycombinator.com/item?id=357674

Use the various suggestions below to estimate your market size. Do both top-down and bottoms-up. Then, for both estimates, consider what would happen if you were off by 10x (in a bad way - you'd overestimated your market size by an order of magnitude).

Do you still have a viable business? Yes? Great, go for it and stop worrying about market sizing! No? Revisit your assumptions in the market sizing calculations - if they are as good as you can make them, and you find that being off by an order of magnitude will kill your business, then proceed carefully, because the actual market size may in fact be 1/10th what you originally estimated.

I did a similar estimate using bunch of download sites and Google Keyword Tool/Trends. The hardest part, from my experience, is not the estimation of market size (people who need a product) but how many of them will actually turn into paying customers (people who want _your_ product).

Especially if your market has free alternatives.