Thread by @QuinnyPig

"So I want to start a business" you think. You're wrong, but you won't figure that out until later.

A thread on how I would think about it these days, updated for 2021.

Consulting? SaaS? Something else? You're skipping ahead. The first step is to find an expensive problem that people would cheerfully pay you to make go away.

Consulting is quick-to-revenue. You can get a check signed in a couple of weeks at most for your first few "friend network" deals.

SaaS requires a lot of upfront initial investment.

The former is easier; the latter is more lauded in our society today.

"Consulting doesn't scale" is a common trope. It's inaccurate. *You* personally consulting won't scale past a certain point; the agency model clearly works well into the billions a year territory.

So let's focus on the consulting path. Products are hard.

"Find a niche." If I had done "cloud architecture" instead of "@awscloud billing," it would have been way harder to find clients (despite the fact that they're one and the same).

@awscloud Assume you'll spend 30% of your time on servicing clients. The rest will be things that are not billable.

Unless you relish the thought of 100 hour weeks... don't bill by the hour. "I will do X for you, and you will pay me Y to do it" style projects (carefully scoped!) win.

You don't want to constantly be weighing doing things with your family vs. "I could be billing this hour."

And you don't want your customers to view every time they call you as an investment decision.

If you want to argue the point and insist on billing hourly, consider that there are ~2K work hours a year.

Further consider that your hourly bill rate x 2000 = "well within tech company target comp ranges." Get a job; it's way easier.

You'd be surprised; find a narrow enough niche and dedicate time to actually understand it, and you'll be a world-class expert in that niche in surprisingly little time.

Think less "databases" and more "migrating specific workload types to specific engines."

Talk less about how you do what you do, and more about the outcome you deliver.

I assure you, my clients could not care less about whether I use Tableau or Microsoft Excel or Python. Speak in terms of business outcomes, not technology.

If you're a consultant, you're in the business of charging people for advice. Why on earth would you not value good advice from people who are better at things than you?

Pay for legal, accounting, insurance, and other forms of advice willingly. Or you will hate yourself later.

You've got 12-18 months generally before your personal network runs dry of word-of-mouth referrals. You want to be constantly growing your presence.

Did you think my Twitter shitposting was somehow just me slacking off all this time? :-)

What you're really after here is the "rolodex moment" of someone talking about the expensive problem you solve, and someone else busting in with "Oh! I know exactly who you need to talk to."

Spend money cautiously. Have enough starting capital to see you through a few months. Have a "failure point" at which you decide it's time to wind things down and get a real job instead.

I'm a strong advocate of not writing custom code for clients; as soon as you do that you get dragged into their implementation cycles. It may not be avoidable for some niches; I would avoid those particular niches.

Find ways to use things you have to do anyway along the course of doing your business.

Example: I turned "keeping up with AWS releases" into a newsletter and then podcast. They're profitable on their own!

A bunch of randos will attempt to waste your time, talking about "partnership opportunities" and the rest. I wish I'd been better early on at shooting those discussions down; I'm way better at it now.

No partnership is likely to change the game for you.

At some point you'll want to hire salespeople, but *you* have to be able to sell the offering first. You can't abdicate that responsibility.

If you can't, they can't either.

By far the easiest path is to get a job somewhere and let the rest of the company worry about the things that aren't in your wheelhouse. If that option is open to you I strongly suggest taking it.