We have software which can run as per user SaaS model and also as custom deployment. To advertise we are planning to host it and give it free for some time / users. But there is another argument as we should look for those who wants to install as custom deployment will pay more. Those who use it as free may don't really pay at any time.
It's not about right and wrong, it's about choosing a market/customer and being consistent.
These lists are not exclusive! Just gives you an idea of the division.
It's not impossible but rare and difficult to have both, even late in a company or product's life.
I agree with Jason's points but want to add the idea of "income diversification" for you to consider.
Like investing, you want to make sure that no single source of income disappearing (customer dropping your service) will put you out of business. For us the goal is relatively uniform income from as many customers as possible.
To measure this we have been using a variation of the Gini coefficient, a statistical metric typically used as a measure of income inequality in a country. We compute it quarterly across all our customers and look for trends.
If you land big fish they will bankroll you, but they will also OWN you. Not literally, but they will have the control in most situations. They may use this control to squeeze you on price and to steer the direction of your product.
In the book, 'The Knack' there are a couple examples of businesses losing money on their biggest big clients due to this squeezing and not even realizing it.
While these are more operational then pricing related here are some thoughts:
If you're a small company there will probably be less support with SaaS as you won't have to support all the users trying to install your software. With SaaS you also have less of a piracy problem which means you lose less control over your pricing in the long run.
It's also much easier to upgrade and add features to SaaS customers and you'll know who the actual customers are much more than with custom deployments where you might know the IT people but not the end users.
It is also easier to test out different pricing with SaaS as people pay in a recurring mode rather than all at once.
Yes, I agree this is a more of a business model question:
Who do you want your market to be? This will tend to drive your pricing and distribution strategy.
For example if your service is online auction software, you could:
Those are all completely different models, but as a bootstrap startup you'd probably lean towards #2 or #3. That tends to set your pricing and distribution strategy, since your customers will either have IT infrastructure and ability to install and host software on their own, or they really won't. Depending on the type of data being stored, they may also be wary of hosting externally, which also defines your approach.
From my own experience, I'd suggest starting with a market segment that you actually like. Even if a segment looks profitable, if you're going to hate dealing with that group from day to day, it's not worth it. MUCH better to pick a segment that you'll enjoy solving their problems, interacting with them, and supporting them.