Ask HN: looking for fresh ideas, my startup is not working

Your examples explain it. “back hurts, anyone know a good chiropractor”, and "my phone cracked anyone know where to get it fixed.” People trend to trust other people, so why would they trust you? What curation and validation are you doing on the businesses? If you are rating them, what are you basing your ratings on? These are questions that I think about when I hear your business. Also, it sounds complicated, and non-scalable.

What you can do: if the people are already on twitter, go to them. Create a twitter-bot that people can follow, and it automatically follows you back. When someone says "my back hurts", you use natural language processing on their tweets, and recommend a solution to them, gluing together Yelp, Angie's List, or anything else that might be relevant. Offer three months free, if they like it, they pay a monthly subscription. If not, take them off a list. When successful enough with consumers, you MAY be able to turn around and start charging businesses, but then, you have the responsibilities of curation, and I think it's easier to let companies who specialize in that ( like Yelp or Angie's List) do that, and you piggy back off them. Check if you're allowed to use their data, and if they charge for it.

This is absolutely fantastic feedback and advice and it gives me a new possible road to take. Thank you. I might be misunderstanding something, but the main thing I wanted to do which a tweet-bot wouldn't is start conversations with the "bad back" person and the chiropractors. Even one or two messages back and forth could answer a question where a bot would come short, for instance "I hurt a muscle around my left shoulder blade lifting weights. Is this a normal athletic case for you to work on? If so would I need an initial x-ray? Do you take insurance and when is the earliest you can fit me in?" If the hurting back person was able to get a few chiropractors to answer these question then they could make a much better buying decision.

It seems like you're trying to apply a Business to Business (B2B) mentality to a business to consumer relationship. Is there a reason for that? Is that your background? The problem is, business to consumer is low margin, high volume. You need to find a way to make that work.

It sounds like you want to do a LawPivot, for general consumer application. You want to build a MARKET, which is the hardest thing imo to do in the startup world. Too many variables, and even with millions of dollars and a great founding team, you can fail (Airtime didn't fail, but for what they were hyped up as, I think it did). If you want to really do this, start with a small sector.

Btw, LawPivot would not work over tweets - it would ruin their professional presence (at least currently). Maybe you want to make a 411 tweet-bot of sorts? Where you tweet a question, and you answer it?

I feel as though you're hitting several different things at once, try to hit one thing, HARD.

Also, automate business to consumer, IN THE LONG TERM. Short term, anything goes. Mechanical Turk, you and a friend doing google searches after reading tweets, anything. Get paid, then automate. This is covered in "The Lean Startup", required reading.

That was just my stream of consciousness, thanks for reading!

I really want to thank you Manglav. You have helped me immensely. Your last reply is spot on. It is like LawPivot and most importantly I think I am failing because I am trying to build a MARKET. This really brings things into perspective for me. Building a MARKET is just not something I can do. Do you think there is way to find a specific niche in the market that I could start with? Also, I am curious as to your thoughts on Pinterest. It has no natural pre-existing market or at least I dont see how one reaches girls who like to create scrapbooks. Yet it succeeded. I am not saying I could do what they did. I am just curious as to how you would define their market.

To be clear, the market is already a niche (Taskrabbit, AirBnB, LawPivot, etc). I was advising you to find a smaller segment of the market to start from. This should be something you're familiar with, and should have a decent concentration of said niche. For example, I don't advise trying match Eskimos with Popsicles that they want (not enough volume). Finding "your niche" is sort of easy, try graphing your time weekly and see what activities you do. Then rank your activities based on the strength you have in them. The top one is your best niche to attack in your local market.

Pinterest pulled a Steve Jobs - it was their job to know what consumer's wanted a year in the future. They were able to see that the coming smartphone explosion would combine with people wanting to be social, and what evokes a lot of emotion? Looking at old photos with friends. Add some filters on top to hide any bad picture taking skills/catering to the masses, and you have a billion dollar idea. They also provided a billion dollar execution, by exiting within two years. Pinterest was a case when all the stars align in the startup universe, and they also worked really really hard to build the telescope to see them.

Another way of doing this might be to look at that persons friends on facebook and see if somebody has liked a page associated with 'chiropractor', then the referral essentially comes from friends.

If a human has to search, call and email, the business model won't scale quickly enough to ever fit the Silicon Valley idea of a startup.

Likewise if the business model is addressing unique customer needs one at a time.

There's not much value in leads for chiropractors, either. Their business grows largely by word of mouth or broadcast marketing.

If you're going to leads one at a time, the sales have to be big. The problems have to be non-trivial, and the actual research more valuable than the localized results from Bing.

On the point that is will not scale quickly. I wrote a bot that searches through various directories and pulls off email addresses, so its an automated process that does my work and emailing. I don't trust my bot yet so I still check over everything it does. The only reason I mention calling is if they want something done really fast and they are willing to pay for it. I understand completely that using a human to call is not scaleable. Its still a way to gauge my market and get customer feedback.

This is a great idea but as manglav said, people trust the opinions of their network so it will be hard to get users to use it when they can ask their friends/colleagues.

Also the solution manglav provides, perfect, if you can do it that would make the process much simpler but it would take a lot of work. Check out http://chirpify.com, they do something interesting with NLP on Twitter.

If you want to work on something else, I would be happy to talk to you more about what I am working on. Email is on my profile.

The one thing we have proven is that while people do trust their own network, they have been open to hearing from us. Off the top of my head, we convert maybe 40% of the time. By convert I mean getting someone to click on a link to our site from twitter.

I suggest doing something completely new. Twitter was a good way to get someone's attention in the early days. Now, people are wary of spam and will probably ignore your message. This alone should be enough to make you dump this idea. Even if they listen to your message, they still have to buy what you recommend. Too many if's.

Find a problem a core group of people have. Develop a solution to solve that problem, and find a way to distribute that solution.

Stop thinking of your ideas. Just pick a niche industry, and see what problems THEY are having. Only start building the solution when you see a convergence of the same problems.

Make some app that track the intake of diet by Kids to learn what kid loves or not and at what time. It might need a bit of mining, Machine learning stuff but worth making.

can I see the URL? I would love to check it out :) And perhaps we can brainstorm an idea or two?