This is a question that was brought to mind by the current episode of the NextNYers video series, where every week they interview entrepreneurs who are building companies in the New York area. This episode is about a company with an online video speed dating web service called Camlink.
Indeed it wasn't a bad interview, but the host, Courtney Nichols, asked one question that troubled me. She asked what sort of market research the company had done to determine if there was a market for this kind of product.
This, it seemed to me, was an unfair question, because it implied that market research is a needed and/or valuable way of determining whether a web service will be successful.
I don't generally believe this is the case.
In fact, I would go as far as to say that without actually building a consumer service, it is rarely possible to determine whether it will be successful. This is because with these kinds of products, the devil is in the experience. Who would have thought that Facebook would have been able to come into the market so strongly and to catch up to or beat MySpace? On paper, it would probably not have been compelling.
The only thing you could have perhaps determined from market research is whether people want to connect with each other. But the truth is it's like asking someone if they would prefer Coke with lemon or lime when they have never tried either. You can ask people if they like lemon. You can ask if they like lime. You can even ask if they get thirsty. But the only way to know what they really are going to like is to observe their behavior after trying both.
In this case, I don't think there is much question that there is a demand or need for tools to help people find dates, partners, or spouses. Will their particular approach work? I don't know, but I don't think any market research would tell me.
But this does bring to mind the question of how you determine if there is a market for a web product. I think the best way is by solving a problem that you personally relate to. Of course its never that simple, particularly if you do not have sensibilities that match those of the mainstream. No matter what, some combination of insight, luck, and execution will always be required, and indeed I think finding that sweet spot is more art and iteration than science.
For developers doing consumer web products, doing pre-launch market research surveys, or anything other than testing product usage is a silly exercise. Post launch, real market reaction is the only research you can really trust.