Choosing the best software strategy for your software development business is significantly influenced by the stage you are at in the software business lifecycle.  Making the wrong choice can have a devastating effect on your business viability, profitability and longevity.  At the heart of it, there are three different approaches in owning a software development business, and making money from it.  I call them Race the horse, Milk the cow and Sell the farm.  Here's how they look:

Race the horse
It's pretty simple - get out of the gate as fast as you can, and try and cross the finish line before everyone else.  The finish line is up to you to define.  It's a fast paced approach, and you have to do it over and over and over again.  The more times you cross first, the more winnings you make.  Of course, it's not a race if there aren't competitors, and depending on your product, there could be tens or hundreds of them.  So picking the right races is very important.  If you (and your business) are kitted out for fast paced action, then this may be the best strategy for you.

Milk the cow
This is a longer term option - the key thing here is about getting 'lock in' to a market - some software businesses do this through proprietary data or complicated licencing.  My personal preference is to deliver quality product and quality service - build a reputation, and your clients won't want to leave.  Once you get enough clients, then you can start to wind back on 'feature creep', reduce development costs (on that product), and reap the long term rewards.  Milking the cow typically needs a supporting income model, such as a regular (monthly) service fee.  This is a great option for those businesses that can grow slowly, get well established, and then focus on fantastic service to keep clients happy.  However, be aware that eventually the cow will get old or sick, and eventually die - so you'll need more than one cow, or be prepared to sell the farm.

Sell the farm
"There's plenty more fish in the sea", and many of them are bigger than you.  So at some point (whether you race horses or milk cows), you might want to sell up and move on.  A strategy to sell the farm is complicated, as most think of this too late in the businesses life - ironically its at the start of your business that you should make this choice.  Selling the farm requires you to recognise that you have two types of customers - the ones that buy your product (or service) and the ones that buy your business.  They both require marketing, selling and support strategies.  For example, take a look at your competitors - if one day you plan to sell to one of them, then an option to consider is to figure out how to make your product complimentary to their existing product line.  This is just one option of many should you want to sell the farm.

I've seen a vast number of online products over the years.  For me, it's pretty easy to summarise the four key aspects of a successful online product.  I use the mnemonic "Play it SAFE".  Here's what that means:

S = Secure
A = Automated
F = Fast
E = Easy (to use)

Secure
Security is number one.  It involves every aspect of a successful product, and needs to be built in from the start.  It should be central to all staff in the company, and every decision should be measured against it.  Any security breach is news-worthy, and will damage the reputation of the product and your business.  Don't risk your customers data.

Automated
Being online means you have access to 24/7 computing power - so use it.  Desktop applications can be exited, online applications keep on living.  It's a significant point of difference.  Your application is also 100% connected to the rest of the world, meaning you can now access information all the time - find areas where it makes sense to do this.  Lastly, don't forget the power of referrals - automate the process for your existing customers to tell others.

Fast
Internet connectivity can never match the speed of a local area network, and applications running in the browser (in 2009) are clumsy compared to the richness of native Windows & Mac applications.  So don't make your customers suffer - find ways to make the online experience fast. Your customers will thank you for it.  (Take a look at google.com for inspiration on how to be fast).

Easy (to use)
Applications built for the desktop have become littered with hundreds of configuration options.  Developers are forced to do this, because they don't know the environment that the application will get deployed to.  In the online world, you have complete control over that environment, so do your customers a favour and make some decisions for them.