Thread by @QuinnyPig

"I hate my job, I'm going to quit and start my own company. How hard can it be?"

Oh so tremendously hard. A thread.

I said this a lot when I was consulting for an agency. "Hey they're billing the client WAY more than they're paying me; I should go direct and capture the margin!"

It's nice work if you can get it, except you can't.

Sales and marketing are actual skills. Customers don't generally fall out of the sky.

If you found a customer to go full time consulting with, you're basically an FTE without a raft of employee protections.

The valuable skill you have, the thing you do? Awesome! You can do that way less if you go independent. Sales, marketing, accounting, chasing customers to pay, benefits administration, insurance discussions, etc. etc. etc.

So you get to either charge way more per hour to fill in the gaps, work 90 hours a week, or... stop billing by the hour.

(Yes, alternately you can go for an agency model and expand, but add a pile of HR stuff to your workload. You are now Overhead.)

Or maybe you build a SaaS product. That's damned close to real estate; you spend a lot of time building before the first dollar comes in.

I've talked before about how @mike_julian and I started the Duckbill Group. "Find an expensive problem people would pay to make go away, charge a fair price, and exceed their expectations."

Great plan, sound model, easy product. OH SO MUCH WORK.

When dealing with businesses beyond a certain scale, there's no such thing as an easy sale. The buyer, champion, person who benefits from what you sell, and person who cuts the check are all very organizationally distant. It takes time.

I mean our model is about as easy to understand as it gets. "Pay us a fixed fee. We come in and help you save many, many, many times that fee on your @awscloud bill."

Again: there is no such thing as an easy sale.

I mean think about it for a second. Who in a company feels that pain acutely?

When do they feel it?

How do you reach out to them at the right time, and via what method?

Things to consider:
* Engineering is skeptical of paying for expertise, generally.
* Finance is befuddled by the AWS bill and doesn't understand what engineering says
* "Fix the AWS bill" is on nobody's yearly goals; it's an unwelcome distraction.

Also consider: how much would you hate me if I was in your inbox weekly with a "Hi, want me to fix your AWS bill yet?" cold outreach email?

So what's the answer?

My answer was to go very loud. The newsletter, the podcasts, the twitter nonsense... It's fun, it's engaging.

And it means you don't have to ask a lot of people "what should I do about my AWS bill" before my name comes up.

That's right! "Marketing" only sucks because so few people market via shitposting!

It's hard to do *well*, but look at cloud billing for a second--how *boring* would it be without a certain snark applied?

Okay. *ALMOST* nobody markets via shitposting. There are exceptions.

This went in some weird directions, but okay, to sum up!

* Running a business is hard, don't do it unless you're tired of getting fired all the time.
* Shitpost more, spam people less.
* Let me negotiate with @awscloud on your behalf and cut a bloody swath through your bill.

I will now field questions.

Remember, once upon a time I was an engineer! "Rearchitect to save money" is a terrible idea, so we don't suggest doing it. "Change this small thing in Terraform and get back to what you were doing" is the goal here.

https://twitter.com/threddyrex/status/1359367688162668547?s=20

Precisely! To absolutely *ANYONE* else, "I was looking at my @awscloud bill and thought of you" is the deadliest of insults!

https://twitter.com/0xMatt/status/1359375679519789059?s=20

Managing my own psychology, hands down. The problems you deal with the most aren't a fit for public forums, so it's isolating and strenuous to deal with.

https://twitter.com/JimAndrakakis/status/1359376104780349442?s=20

Ballpark a million dollars a year in spend. There are exceptions, but it’s a good rule of thumb.

https://twitter.com/jkhelgerson/status/1359377216979820545

I’m increasingly convinced that my “I run @awscloud Marketing” joke isn’t really a joke.

https://twitter.com/GDayJason/status/1359377468952465408

Yes. For example, both Oracle and the Duckbill Group clearly have a budget line-item labeled "Spite."

https://twitter.com/nixgeek/status/1359378545475133440?s=20

1. 15-20% first pass unless we find something... interesting.
2. Sure, but not via us...
3. ...because we know the AWS knobs and dials super well, whereas Azure etc. are radically different for negotiation purposes.

https://twitter.com/jkhelgerson/status/1359378746327900160?s=20

The fee varies based upon complexity, but is directionally "less than an engineer's annual salary," which leads to hilarity when we find $20 million a year in savings.

We have a fair bit of tooling, but it's not what you think.

And most of it's easy.

https://twitter.com/nixgeek/status/1359380328515661826?s=20

(Well "easy" in the sense of "not significantly disruptive to engineers to implement.)

...more than that. Hey @mike_julian, we need to update the counter on duckbillgroup.com again.

https://twitter.com/MoreScooters/status/1359382463768915970?s=20

That was fun. I should do it again soon. It was a complete lark except for the fact that it worked SUPER well.

https://twitter.com/0xMatt/status/1359381143447998467?s=20

Almost entirely the former. Negotiations are special engagements and hard to describe in too much detail in public, unfortunately.

https://twitter.com/jkhelgerson/status/1359381179531743232?s=20

A bit over three years after I started. If the person who starts the company can't sell the offering, no one else can either.

https://twitter.com/nnanna/status/1359383353670193152?s=20

We've done a few of these--usually migrations. "Help us validate our cost structure so that when we scale up the bill doesn't blow holes in our collective faces" style.

https://twitter.com/closemindedjerk/status/1359385006926405632?s=20

Oh, two points as to why this model of ours works.

1. We have *no partnerships* with any vendor, and we don't do implementation. There's no interest conflict here.
2. We're not VC backed. We don't need to capture "all the value" or grow obscenely.

I didn't even try. "Hey, I'll come in as a contractor and Do the DevOps" is invariably a losing game.

https://twitter.com/neilatherley/status/1359387111879802882?s=20

This is why we're not VC backed. We don't have to capture the entire market, nor do we want to. I really hope someone spins up our equivalent for Azure, because I won't do it.

So far every similar business we've seen deviates from what we've done.

https://twitter.com/aHev/status/1359386787051831298?s=20

(And that's okay! But they invariably lead to conflicts of interest. "We'll implement fixes for what we find!" being the most common...)

No, because if your @awscloud bill isn't painful, why would you pay us? You always *can* save money, but people are expensive. You're spending dollars to chase pennies at some point.

https://twitter.com/MehranJal/status/1359386531295813633?s=20

There are a few I've worked with in the past that I can recommend, but there's no business relationship between us. There are others I've worked with that I would NOT recommend.

https://twitter.com/AbstractCode/status/1359389925146202113?s=20

To be clear, I have sponsors for the newsletter, podcast, and some video things, but that's not the same thing as having a partner.

I like and recommend @ChaosSearch constantly, but that's because they're good. I don't get paid anything different if you buy or not.

I was recommending ChaosSearch before they were a sponsor because they solved a problem. If they stopped sponsoring me I'd still recommend them to solve that problem.