If you know a technology well, should you use it for new projects?
Or you should try to objectively choose the "right" technology for the problem at hand?
If you go with what you know, you can get started faster, but you might paint yourself into a corner or waste time later.
Doing the "right thing" is good in 3 years, but what if your startup never makes it that far?
If you are a developer creating your own product, then as a general rule, I think picking the technology with which you are most familiar is the most important thing, because it enables the multiplier effect called experience. Experience accounts for a lot, and in ways that may not be obvious.
In addition to minimizing ramp-up time, avoiding rat-holes, and being able to reuse components you've developed before, experience with a given technology probably means you know people that you can get help from if you get stuck, and you know the best books, blogs, add-ons, and developer tools. But maybe most importantly, you can evaluate the talent level of other developers should you ever need to hire employees or contractors to help you develop your product.
If you already know a specific technology, then, it is already shaping the way you think. As such, at the moment, you are bound to the limits of that thought process and the paradigms that come with it.
I'd probably start with the tech you're most familiar with and perhaps experiment with something you think might be better. But you have to experiment in a meaningful (measurable) way and ask for opinions from others and see if it matches your actual goals.
If you want something that can scale better than your current technology, you might want to see if you need to scale at all.
When you're choosing a language or a platform, the community and available libraries are often more compelling arguments in favor of a technology than your own personal experience. Given solid technology fundamentals, it shouldn't be too hard to learn a new language or a new technology platform. Additionally, you can always hire expertise.
How active is the developer and support community for a language or a platform? This will determine improvements released over time and if there will even be support for your choice in the future. What kind of libraries, plugins, and extensions are available? A large selection of modularized solutions to common problems both saves time and mitigates risk in the long run.
We've chosen Ruby/Rails for these reasons given the active development on the Rails platform, the great developer community, and the huge array of gems/plugins available. This has paid long term dividends over the last few years of development. The fact that Rails serves as a great rapid protoyping and development platform is a nice extra.
The above points assume your product development has achieved reasonable product/market fit. If you haven't validated your fundamental concepts then you should be using whatever tools meet that short term goal the fastest. You're more likely to throw away code until you've reached that level of stability.
There's no right answer for this one. Do what makes you sleep better at night.
If you're the kind of person who prefers eliminating market risk first, by getting an early prototype out and iterating on it, then you probably want to use whatever technology you're most productive in.
If, on the other hand, you're worried that your choice of technology is bad, maybe it's better to play with other options first. Maybe just experiment with them for a bounded, finite period of time.
A good question to ask: maybe you know technology X, but would you be able to recruit other stars who know (and want to use) this technology?
I am going through the same dilemma. Do I create a quick prototype with the technology I have years of experience with or select a technology which I have no experience with but which seems to be in favour with investors.
Does anyone know of a case where a successfull startup changed the course after 3 years (or X years) from one tech platform to other?
Is it correct that investors favor open source technology and do not want to invest in something which uses licensed tech?
So, even if I am expert in Java or .Net , do I need to learn Ruby on Rails and PHP?
I don't think the two options are mutually exclusive because part of the method for determining the 'right' tech is how long it would take for you or your team to get up to speed with it. Effectively, for each plausible option you need to work out the trade off and choose the most appropriate based on that.
If you can get 5% extra performance but it takes twice as long to learn and build, then (depending on your industry/product) you might be better sacrificing the 5% performance to get the job done faster.
It only gets a bit dicey when you don't know the benefits/disadvantages of pursuing different techs...
Doing the right thing is one of the best decisions you can make to ensure that the Startup lives long enough to celebrate it's 3rd birthday!
So my advice is to think about how you would build your product with the technology you know (so that you at least have an idea of the effort and cost). Then, if you know of some new technology that might be a promising new option for what you need to deliver, take a look at it, and see how it compares with the estimates you have.
How this helps
You should objectively choose the right technology for the problem at hand, hands down.
The goal for a startup is to start up, and be successful. Or, to be appealing enough to someone (individual or company), so that your work impresses them enough so they choose to purchase the company for a good amount.
For either case, whether for your own future, or for the potential sale to a good buyer, use the right technology that will be scalable in the future.