You haven't mentioned what style of business you want to start. That makes a huge difference in what your approach should be.
My experience is starting more traditional small businesses (i.e., bootstrapping without VC money). If that's not what you have in mind, my comments may not be valuable to you.
First, forget LinkedIn. It's worthless for this sort of thing.
When I've done with good results is to go where my potential customers are. Identify the demographic you're wanting to address ("everybody" is a wrong answer -- in general, the more specific and narrow, the better) and find out where they hang out. I don't mean just online, but if they have physical gatherings, go to those too. Don't go there to sell or overtly interview "prospects". Go there to become part of that community -- you'll learn about all the invisible pain points they have, you'll develop a presence in that community, and you'll learn what it values. All of that is critical to success.
By the time you're actually trying to find customers, you'll be "one of them" and finding them will be much, much easier. You'll at least know where they are and how to talk to them.
>We earned non-zero revenue but very small and not enough to say they were living expenses. We did some amount of marketing but not enough. My introspection tells me that this happened because we built without good ICP or any feedback from users.
You say that you developed without "good ICP" or feedback from users, but you still managed to have "some" revenue. You may be onto something provided more people become customers. In other words: enough people hear about your product and of those, enough people make a purchase, and of those enough people keep paying you.
The problem may or may not be ICP related, and may or may not be feature-set related, but I'm willing to wager that it somehow is not enough customers purchasing.
It is unclear from the "not enough marketing".
Tinkering with these variables is not mutually exclusive: you can have "more marketing", and have more feedback starting with your existing customers, and get more marketing from your existing customers in the form of recommendations or getting other people to use the product, or introduction into their organizations, etc.
>How and where do I find users to talk to? Where do I find these cohort of people that can at least respond to user interviews or MVP feedback. I don't have good network and that hurting most.
You solve a problem for users who are aware of the problem, are aware it is a problem, have spent coins to fix it or have a say in some coins being spent on fixing it, are unsatisfied.
One of the ways to speed this up is to go higher up the abstraction scale and target a vertical/sector, but not too high or you'll spend so much time you'll never ship, and not too low without being careful not to get locked-in. This way, you get the distribution of someone stronger. Call it piggy-backing, gravity assist, leverage, whatever you want.
What a mouthful of abstract new-age mumbo-jumbo... Let's do an example:
Imagine you're solving a problem for an airline. Going up the abstraction layer is to work directly with their ticketing system provider, say Amadeus. Now, instead of just one airline, how many airlines use Amadeus? The next step is to talk with Amadeus. The benefits is that its distribution is much, much, stronger than you trying to get with multiple airlines, and your solution will stay mostly as is because you're using the same data model.
If you're solving a problem for hotels, go one notch up the abstraction layer and look-up software vendors for that industry/vertical. Get compatible with their software/data model. Distribute. Does it have a marketplace? An ISV page (Independent Software Vendor)?
What's also beneficial is if there's a conference for that industry: do they all get together at some point somewhere? Is there a standardizing or regulating body, an authority of some sort, mandatory registration with an organization? Is there a ledger somewhere? A trade association they belong to? A union? Where do they congregate?
>As a sidenote, it has become miraculously difficult to get a job as a developer in late 30s.
For every "software" position at a "software company" that knows it is, there are thousands at companies that don't know they are. Oracle and Microsoft didn't get to that size selling to "tech companies". There are huge organizations running on VB macros or FileMaker or WinDev, factories with so many PLCs and no love for humans, etc. "You may find in the sea what you don't find in the river."
You haven't mentioned what style of business you want to start. That makes a huge difference in what your approach should be.
My experience is starting more traditional small businesses (i.e., bootstrapping without VC money). If that's not what you have in mind, my comments may not be valuable to you.
First, forget LinkedIn. It's worthless for this sort of thing.
When I've done with good results is to go where my potential customers are. Identify the demographic you're wanting to address ("everybody" is a wrong answer -- in general, the more specific and narrow, the better) and find out where they hang out. I don't mean just online, but if they have physical gatherings, go to those too. Don't go there to sell or overtly interview "prospects". Go there to become part of that community -- you'll learn about all the invisible pain points they have, you'll develop a presence in that community, and you'll learn what it values. All of that is critical to success.
By the time you're actually trying to find customers, you'll be "one of them" and finding them will be much, much easier. You'll at least know where they are and how to talk to them.
>We earned non-zero revenue but very small and not enough to say they were living expenses. We did some amount of marketing but not enough. My introspection tells me that this happened because we built without good ICP or any feedback from users.
You say that you developed without "good ICP" or feedback from users, but you still managed to have "some" revenue. You may be onto something provided more people become customers. In other words: enough people hear about your product and of those, enough people make a purchase, and of those enough people keep paying you.
The problem may or may not be ICP related, and may or may not be feature-set related, but I'm willing to wager that it somehow is not enough customers purchasing.
It is unclear from the "not enough marketing".
Tinkering with these variables is not mutually exclusive: you can have "more marketing", and have more feedback starting with your existing customers, and get more marketing from your existing customers in the form of recommendations or getting other people to use the product, or introduction into their organizations, etc.
>How and where do I find users to talk to? Where do I find these cohort of people that can at least respond to user interviews or MVP feedback. I don't have good network and that hurting most.
You solve a problem for users who are aware of the problem, are aware it is a problem, have spent coins to fix it or have a say in some coins being spent on fixing it, are unsatisfied.
One of the ways to speed this up is to go higher up the abstraction scale and target a vertical/sector, but not too high or you'll spend so much time you'll never ship, and not too low without being careful not to get locked-in. This way, you get the distribution of someone stronger. Call it piggy-backing, gravity assist, leverage, whatever you want.
What a mouthful of abstract new-age mumbo-jumbo... Let's do an example:
Imagine you're solving a problem for an airline. Going up the abstraction layer is to work directly with their ticketing system provider, say Amadeus. Now, instead of just one airline, how many airlines use Amadeus? The next step is to talk with Amadeus. The benefits is that its distribution is much, much, stronger than you trying to get with multiple airlines, and your solution will stay mostly as is because you're using the same data model.
If you're solving a problem for hotels, go one notch up the abstraction layer and look-up software vendors for that industry/vertical. Get compatible with their software/data model. Distribute. Does it have a marketplace? An ISV page (Independent Software Vendor)?
What's also beneficial is if there's a conference for that industry: do they all get together at some point somewhere? Is there a standardizing or regulating body, an authority of some sort, mandatory registration with an organization? Is there a ledger somewhere? A trade association they belong to? A union? Where do they congregate?
>As a sidenote, it has become miraculously difficult to get a job as a developer in late 30s.
For every "software" position at a "software company" that knows it is, there are thousands at companies that don't know they are. Oracle and Microsoft didn't get to that size selling to "tech companies". There are huge organizations running on VB macros or FileMaker or WinDev, factories with so many PLCs and no love for humans, etc. "You may find in the sea what you don't find in the river."