How to Do a Startup in 2020?

Question

So I stumbled onto an idea for a 'app' that I think is pretty big. I've run it by a few people that are exceedingly cynical and the response was somewhat shocking (they both loved it). Having many years in this industry I understand that a good, new and original idea is pretty rare so I've naturally gotten pretty excited to develop this thing. However laying in bed last night with this all buzzing around I realized I don't know how to go about it in 2020. I did a number of startups in the early 2000's (non 'made it') and back then your options were basically to bootstrap it yourself or try an incubator.My question is this - in 2020 what's the best, most viable way to develop and bring to market a good/serious idea that you want to make a serious run at? It used to be you'd build a prototype or even the full website/app, etc and along with a pitch-deck you'd try to raise money or give an incubator a shot.A bit more about the idea... it's big enough where I think it would need the backing of something like ycombinator, techcrunch or techstars to really reach it's potential. Reason being, it would compete in an existing space where there are already huge, well funded players (I know this is typically the way things are). To put it into better perspective, it would be something on the level of a linked-in, indeed or maybe tinder, bumble are better examples. To be clear I'm not saying it's either of those niches, just saying it would be similar to those and I'd be dealing with breaking into that kind of established space.Just wondering what the best approach is for 'going for it' in 2020 for your dream-shot, retire and buy an island idea.Thanks!

slap_shot · Accepted Answer

It's the same as it's always been: build something people want and acquire customers.
> I think it would need the backing of something like ycombinator, techcrunch or techstars
That's a bad mindset, imo, and probably just a distraction. If it's such a great idea that you think it is, build it and acquire users.
I think it's worth glancing at https://www.ycombinator.com/resources/ - but I think more people use this as an alternative to building things and getting users.

muzani · Answer

Startups are still the same thing as they have been for 10, 20, even 100 years. You build something of value, and try to convert that value into assets. Acquiring companies won't look so much at your cashflow, but rather you as an asset that complements their existing processes. So the goal is basically to build as much asset as fast as possible, take it as far as you can, and sell that to someone who can take it further.
Bootstrapping is tough for tech, because the curve is usually exponential. If you were selling lasagna, you could easily make $1000 profit the first week, and bootstrapping is easy.
With tech, it's quite common to spend a year trying to get to $1000/week, and then make $30k/week the next year.
You also have to consider competition. There are many kinds of ideas. Making a Tinder for X might not have too much competition, but if you do something that Google/Amazon/Facebook is building, you'll get crushed if you move too slowly.
This is usually where venture backing comes in - investors are happy at getting good returns, and you're happy at skipping years of consulting, etc to pay the bills.
Don't spend too much time negotiating or optimizing; you're aiming for 3300% annual growth and a 20% better deal isn't worth your time. Paul Graham/Viaweb started at 10% for $10k. (http://www.paulgraham.com/ycstart.html) It seems unfair, but it was still a good deal for both sides.

1337n008 · Answer

well, you answered your own question - bootstrap it yourself. i know exactly where you're at, unfortunately.
my advice is - go with backend as api and separate frontend. define the api(protobuf+swagger plugin, not raw anonymous json because you will get lost in fluid schema later on and with 10+ routes you will go crazy without schema) and strict interfaces. so you know how data will flow. just pure basics, not all the functionality you have in your mind/vision. avoid DRY, copy code all over the place, go with kiss and most importantly YAGNI. use tech you know, not what is fancy right now.
then proceed to implement it in the fastest way possible. the end result should be a MVP with the worst code you can imagine. I myself have never been able to execute it, unfortunately. Each programmer strives for perfection which in this stage means time wasted not delivering the end product. I am certain that this way, no matter the size of the project, you should be able to produce something usable within 3 months.
You WILL be struggling with authN but mostly authZ(unsolved problem) so go for the simplest solution you can think of.
Next important thing is to do a monolith first. NO EXCEPTION. Forget microservices or SOA at this stage. I too fell into this trap and kinda keep repeating this mistake over and over. People like to use too much abstraction. Just brute-force it.
Then make sure you run one single VPS. No cloud, not kubernetes, no docker, nothing. Run it as close to metal as possible without any infra mumbojumbo fancy scheduler...tech..whatnot. And avoid cloud VPS. Look at OVH, Hetzner, DO or whatever. Just not cloud VPS because they are crazy expensive. Don't go for bare metal just yet(and avoid cloud altogether, it will bankrupt you in the long run, bare metal is where it's at).
Once you get the MVP up and running, look for first paid customer(s) and listen to their feedback. That feedback is free and no one else can give you better information than actual users. Also don't bother promoting the project on forums looking for people to test it for free. That is a dead end, trust me. Spend few bucks on some basic presentation page and some ads. Most notably it should be immediately clear on the website what your product does. I'd say 80% of companies these days fail at this horribly.
When you take payments, always upfront and annually will be better than monthly.
Once you get there, let us now :)
Good luck soldier.

funfunfunction · Answer

It really does depend on the idea. Is it a consumer product? B2B? In the last few years I&rsquo;ve seen a growing number of companies start with little or no seed funding and grow to a respectable size. If you&rsquo;re technical, I would start by building a prototype and trying to get a few paying customers to prove it something people are willing to pay for.I&rsquo;m curious to learn more. Reach out to me at sam@sellout.io if you want to chat.