What's an agile way to conduct market gap analysis to determine an underserved niche?

You could look into markets that are served by old monopolistic companies. If you for example could get a list of medium sized companies sorted descending by age, then some of them should be in monopolistic situations (more than 50% market share or so). Since they are so old, I doubt that they will be able to stay innovative. Then make a list of innovations during the last 5 years or so, which have been real game changers: bio technology, wireless communication, internet speed, ebooks etc. and consider whether any of these technologies could be a game changer also in their market.

If not, then make a list of recent fundamental social changes: Green movement, risk of terrorism, changed immigration, social media, piracy (somalia), news broadcasting (Twitter, internet newspapers), usage of communication media (SMS, Twitter), etc. and consider whether any of these social changes could change the game also in their market.

If not, then make a list recent large price changes for widely used products: increased raw material prices (copper, gold, rare earth minerals), food prices (grain), lighting prices (LED lamps), broadband prices, roaming phoning (in Europe), flight prices (low cost carriers like Easy Jet, Ryanair), cost of capital, home prices, robotics, cost of some types of chinese/indian low-cost labour, price changes due to changes in subsidies/taxation or opening/closing of markets. See if any of these changes also severely affect their market.

Chances are that an old monopolistic company will not be able to keep up with such fundemental changes, meaning that new markets/niches arise.

I think the most agile thing is to use the Google Keyword Tool and check:

1. How many people are searching the terms you want to target for your product
2. How competitively companies are fighting for those keywords. IF there are a lot of searches and no one serving ads, it seems like a good niche to me.

You can even take it a step further and run a very low budget campaign on the terms, setup some landing pages through some 3rd party service (many have free accounts), and see if you get anyone to sign up for your fake product. If you get enough sign ups, build it and start selling it to your premade customer list.

Random tip - My friend did some testing on landing pages trying to get visitors to sign up for a service by just inputting an email, and another page getting people to sign up by putting in their email and reserving their username. Reserving the username had much higher results (almost 10 times the conversions). It was a small test, but I thought it was interesting.

If there were a formula to identify lucrative market gaps, I guess this site wouldn't exist at all!

However, one technique I've used and taught is 'one step innovation.' This takes as a starting-point a strong existing solution, one where it's possible to identify

• The domain (the broad sphere this solution operates in)
• The problem being addressed (defined if possible without reference to the domain)
• The underlying technology utilised
• The method of distribution
• The monetisation method (where relevant, end-to-end through the system)

Hold all but one of those fixed, and explore all the ways you might change the other.

Wherever the new proposition feels like it might have legs, go and look for someone already doing this.

You might find a (potential) gap - there really isn't notable competition. Explore this!

Or you might find an established market with one or more comparable and established offerings. Move on!

Or you might find a strong incumbent with a better solution. Try the trick in reverse - taking their proposition and varying that same element, could you create new market space? (For instance: you found a great B2B solution. You looked how it would fare in a B2C domain. You found crushing opposition. So now, how would an equivalent to the B2C service fare in the original B2B context?)

This is admittedly a tedious process. But I've found it fruitful on more than one occasion.

The agile-ist way I've found is to figure out what I'm missing and see if there is something that fills it. If there isn't I make something quick and dirty that solves my problem. If I couldn't live without this quick and dirty little thing, then there are likely a large number of people in your boat who would pay for it.

I would iterate a bit, trying to see if there are related issues that the product could solve (and develop these quick and dirty), and repeat until the product is a reasonable whole and has enough features to be marketable and valuable.

Then you release it, and sell it, and iterate over problem areas.