HN Explorer

Ask HN: I built before I sold, Am I Doomed?

1 points by prakhar897 ↗ HN
I built a personalized cold message writer for LinkedIn. It uses ChatGPT + LinkedIn Profile to write messages.

I was so sure that this idea is needed that I jumped the validation part and built the whole product. Now the product has very slow sales and is very unfocused. Should I quit the whole thing and start again or do something else. Any advice would be appreciated.

9 comments

[ 3.9 ms ] story [ 29.2 ms ] thread
jimnotgym
You need to see how people are using your software, how they are struggling, what they would like. See if you can meet with both some of your customers and watch them use your software. Try and meet some of the people who have tried but not bought, find out why?
prakhar897
Thanks!! I'll contact all existing trial users for an interview.
PaulHoule
Can’t you get it to write cold messages to sell itself?
prakhar897
1. I've put it on subreddits but haven't got the traction I wanted. I think the product might not be solving the correct problem. Hence, I didn't try to cold message people with it.

2. The price point is 5$/month which really not feasible to contact customers individually. If I do 100 cold outreaches a day, 10 try the product and one converts, it's 5$/day salary.

PaulHoule
i think that is your problem there. The product is too cheap. How much value do you think a customer could get if it were successful at selling?
jimnotgym
Go back to the classic YC advice, do things that are not scalable. Cold call a few a day and use it as a way to get users and feedback. It will also give you more users, some of which may have networks that they may share it with.
verdverm
If you quit at the first point it gets hard, maybe founding a company is not for you.

My advice, stick with it, but understand there are probably 100s of these already. You need to think about what makes your offering unique, especially with ChatGPT + <anything>. It's trivial to wrap an LLM around a small API or block of text.

Have you done any prompt engineering for your product? That will make a world of difference in output if you do it well.

Here's my example:

1. We are trying to address the code gen issues in LLM by combining it with our deterministic code gen tech

2. We spent time to reshape the problem and craft a prompt that works well for this integration

https://twitter.com/verdverm/status/1652504163635347456

fuzzfactor
Not doomed, now you know exactly what you have to offer before building your sales pipeline.

Not everybody has this.

OTOH if you had first identified an eager market before bulding your offering, then you would have known exactly what to build before you started the tech project. You can always do that too, then you'll have twice as much to offer.

The way it is now there could be a strong chance that your sales efforts for what you already have, will turn up some good leads about what to build next time that will be easier to sell.