The most effective way to learn is to go out and find the people who should be using the product, and talk to them.
Problem is? if they are taking the time to talk to you, they are giving you free advice. For whatever reason, they like you.
If you are doing a product that does face to face sales? that's fine. People like you, they buy your product.
However, this tells you very little about how good the product is on it's own, as a low-touch commodity product; so it's not all that useful if you are building something to scale (rather than scaling a marketing organization, which is a very different sort of problem.)
Also, most people? they are trapped in their own worldview, so they can't tell you what they need. "If I had asked people what they wanted, they would have said faster horses" - Really, to actually innovate? You need to both understand customer needs /and/ the technologies you have on hand that could serve those needs in ways they are not currently being served.
My philosophy has always been to build into a market I knew. Sell to yourself. If you want to sell to a new market, learn about /that/ market first. I mean, this approach has it's own problems[1], but eh, it does mean you know when you are providing actual value, rather than having to sort out "this person likes me" vs. "this person likes my product"
AdWords is great for validating B2C products because you NEED the opinion of MANY people to judge the viability of your product/idea.
In B2B your customer typically has money on the line, and it's quite straightforward for them to evaluate your value proposition: "Does this make me money, or lead to making me money, how much?".
In B2C the question becomes: "Do I like/enjoy/have fun with X product?" typically far more subjective and whimsical. Subjectivity shines in small samples (<10 people) and shrinks in large samples (>1000 impressions).
I agree regarding features/press however.
This doesn't directly translate to people opening up their wallets and paying for said product.