3 ms·
Depends on the ideas. Personally, I always feel it is better to go talk to people in person (ideally) about the idea. Gauging interest in person or at least o
by davismwfl 8y ago
Depends on the ideas. Personally, I always feel it is better to go talk to people in person (ideally) about the idea. Gauging interest in person or at least over the phone is a much better indicator of real interest than a signup form online.
Even if the product is going to be marketed and sold completely online, meeting people in person who would be target audience customers lets you get better data and validate the idea further than just a signup form. e.g. you can ask follow up questions, ask people questions around functionality etc. Also this method lets you find out how you find the right target audience, as in where online would you find them looking for your service, what keywords would matter to them, what media platforms do they pay attention too...
From there you can tweak the idea and if it seems promising then publish a landing page, start collecting more contact info and trying to convert some of those into commitments.
- wuliwong 8y agoBut face-to-face interactions are scaaaary !? :) I do a lot of talking to friends/family about ideas but I need to get out there and identify people that are actually in the target audience of the product.
- davismwfl 8y agoTotally, and talking to friends and family helps, but they are the least critical (typically) of your ideas because they want to be supportive. So finding people that are really the target audience and talking with them is critical. Personally, I find it is always hardest to start the first few conversations for any new project, but after that it seems to get easier for me. Good luck!
- wuliwong 8y agoThanks Davis. I received the suggestion today to actually try to get people to pay you up front for the product you are going to build. That was pretty interesting. Seems to scratch my itch for some real objective (in)validation and also prevents me from writing a bunch of unnecessary code.