Innovating on your strength
One of the great things about the Web is that the barrier to entry for a new enterprise is incredibly small - all the infrastructure you need to get started is a $300 computer, a $10/year domain name and a sub-$10/mo hosting plan. Your only limitation will be your own ability create something that will generate enough interest (or money) to sustain it.
Compare this phenomenon to almost any other industry - say restaurants or construction - and it's easy to see why there's a lot of innovation on the web. It's easy to go from idea to execution fast and inexpensively.
The Web has no Captcha
Unfortunately, because of the this ease of execution it means that there's not a lot of friction to prevent crap from getting shovelled onto the web. In other industries the need for physical space, goods and capital before launching into a business acts like a Captcha for innovation - forcing you to look long and hard at what you're doing before dropping a bunch of your or your investors' money and going forward with the enterprise. In developing for the web there's no Captcha - you don't really need to prove that your idea is any good to anyone if you can get it built by paying your neighbour's son $500 (You will get what you pay for however)
This is great in many respects, but there are downsides that people often forget. First, there's a lot of competition. If what you want to do isn't all that innovative, then there are going to already be a couple of thousand people doing it, and your chance of being successful is pretty slim. You'll need to be in it for the long-haul and be willing to iterate on your idea with the experience you get from the initial release in order to have any chance whatsoever of being successful. What this means is that, secondly, if your idea doesn't match up with your interests you are going to have a hard time generating the necessary follow-through to keep working on your idea and make it successful.
Just because it's quick & easy to start something on the web doesn't mean it's just as quick to become successful at it. While there are incredible get-rich-quick success stories (like Chatroulette) there are just as many companies that followed a more standard growth path. Amazon, for example took years, loads of funding and a bunch of changes of direction to before approaching profitability.
What this means is that if you have what you think is a great idea but it's not one that matches up with your particular skills and interest, you need to either find a partner with whom it does match up or pick a different idea.
As a web developer we've been presented with a lot of people's ideas (and been asked to sign one-too-many NDA's), and we've gotten to the point where we're not willing to work on a project if we don't think it has a chance of being successful, even if it's a lucrative project. I'm not just talking about the idea itself, the people behind the idea are much more important - their willingness to continue the project past the first round and put the legwork in to make it successful need to be there for a project to have a chance.
So, if you want to be successful on the Internet, you need to innovate on your strength - the type of project that you build needs to match up with your particular skills, or all the most talented-outsourcing in the world isn't going to help. I'd say there are roughly four different types of Web enterprises out there:
Product-driven, Content-driven, User-driven, Technology-driven
Now there's obviously some overlap between the different types, but these four archetypes are a good start.
These are sites whose purpose is to sell something physical. Innovating on a product-driven website means, for the most part, innovating on the product itself. You can obviously dress the site up with bells and whistles and that will help in the short term, but it's the product itself that needs to sell and unless you feel confident that you can provide something unique in the product or the way you sell it, the greatest web site in the world isn't going to help. Furthermore, if you are not excited about the idea spending your time manufacturing product, fulfilling orders and labelling and shipping packages you should stay away (there are ways to streamline this, like fullfillment houses, but if these aren't problems you want to solve, that should be a warning sign.
These are sites that provide generated content to the user - think blogs or newspapers. Functionality is a lot less important than keeping content fresh. The technology running a site like TechCrunch or HuffingtonPost really doesn't matter - it's the attention to fresh and awesome content that drives pageviews and thus advertising revenue on those sites.
These are sites that exist to allow users to generate content - whether it's a social networking site like Facebook or a communication tool like Twitter. These will often need to expand functionality to keep up with the demands of users, and spend a good deal of their time responding to users problems, requests and demands. If you aren't excited to interact with your users, you probably shouldn't be doing a user-driven product.
These are sites that people come to because of features of site itself and what you can do on it. Any of your online b2b apps fall into this category, site's that provide E-sourcing solutions, for example, need to have a value proposition that makes them better than using the competition.
So if you are a Web Developer yourself (like we are), it wouldn't make sense to build yourself a content-driven site if you're not interested in keeping that content up to date. On the other hand building a technology driven site will play into your strengths. You might be tempted to create a user-driven site, but unless the community, monitoring and support aspect appeal to you (hey, not all geeks are people-people), you might want to reconsider.
Now If you're a MBA-type, you probably don't want to create a technology-driven site unless you're going to be able to pay to continue to develop the site once you get the first rounds of feedback. As good as your idea is, you or your developer most likely isn't going to hit a home run on the first swing. Be ready to immerse yourself in the technology of the site for the duration or pick something else.
If you're a passionate quilter, you might not actually be happy building the worlds best quilting community, as that means you'll be dealing with user issues and trolls more than spending time quilting. A product driven site selling quilts or a content driven site that keeps up with the latest advances in quilting technology might be more interesting to you.
If you're a marketer that cares more about innovative ways to sell stuff than actually build stuff, a product driven company is probably a bad idea. SEOMoz and Hubspot both started out as content driven (good for marketers) and then as they gained traction and that ephemeral quality of "Thought Leadership" built themselves products that turned them into technology driven sites.
We would never have built something like GamesForLanguage, a combination of content and technology driven, but heavy on the content part, if we weren't working with two people extremely passionate about developing the content necessary to make the site successful.
Finding something you're passionate about to drive your startup is really just half the battle, you also need to ensure that the type of company you're building really matches your strengths, otherwise you'll spend all your time innovating on pieces of your business that don't really matter because that's what you'd rather be doing. Innovating on your strength means you'll be building the sort of company you would be happy working in, and since you will be working there a lot.