Gatsby

About these notes

The benefits of a dynamic website

# Gatsby is inventing a new category of websites
Imagine a static site generator that can rebuild a website with fresh data in less than five seconds (or even one second).

Combined with a CDN that deploys to hundreds of locations around the world in ~200ms ([Fastly](https://www.fastly.com/network-map)), new data would be live for all visitors in less than ten seconds.

That would make it possible for many (now) dynamic sites to be built as static ones, with all [[The benefits of static websites]] (performance, scalability, uptime, cost,…) included! 

While it’s technically a set of static HTML files on a CDN in the end, the architecture is neither truly static or dynamic, but something in-between.

That is what [[Gatsby]] is working towards—and [we are getting close](https://www.gatsbyjs.org/blog/2020-01-27-announcing-gatsby-builds-and-reports/). 🤯

----

I was sold on this idea when I realised that we could have architected [[Spectrum]], a public real-time chat app, like this. Here is what I wrote in our Slack at the time:

> _If build + deployment of new content is fast and stable enough (< 1s?) new content is available super quickly everywhere around the world, no matter where the db is—it’s just a bunch of static HTML files after all. In theory, we could’ve built Spectrum with that architecture and made it much more scalable, performant and stable!_

 🤯🤯🤯




Spectrum

# Getting acquired was the most stressful time of my life
The GitHub acquisition of [[Spectrum]] was the most stressful time of my life. I am generally a calm person and rarely let work tension affect me, however I had a lot of sleepless nights during that process.

Our entire future depended on this working out. We were at the end of the road.

Get acquired by GitHub… or die.

There were seemingly infinite questions about the company, the product, the people, the metrics, the users, the code, the architecture, the services. We obsessed over every single answer, as any one of them could have been a dealbreaker. (I say “we” but really Bryn and Brian handled most of it)

I felt like I was a fraud for how bad Spectrum was technically ([Tech Choices I Regret at Spectrum](https://mxstbr.com/thoughts/tech-choice-regrets-at-spectrum)) and feared I might end up being the reason the acquisition falls through.

Even if the acquisition would go through, would GitHub value the product and the communities using it like we do? Would we be able to champion it internally?

It felt like nothing mattered and we were in the hands of fate. I had no motivation to work on anything. My brain was spinning every night with doubts and fears and I slept worse than I ever have.

This kept going for months.

Until one day I had an email in my inbox with the acquisition contract. The relief I felt when we signed it was indescribable. A mountain was lifted off my shoulders.

---

For a great portrayal of the feelings you go through when being acquired, I’d highly recommend listening to the [“Exit” season of the StartUp podcast](https://gimletmedia.com/shows/startup/xjhwn8). I listened to it and it viscerally describes the same feelings I went through.




# My filters for potential business ideas
_Inspired by [Derrick Reimer’s](https://www.derrickreimer.com/essays/2019/05/28/finding-my-next-bootstrapped-business-idea.html) excellent post._

[[I want to build an small business]] eventually, however I find coming up with ideas for one rather difficult. In order to quickly evaluate potential, I thought about a list of filters that help me focus on the right things.

## Are people searching for the problem on Google?
With [Changefeed](https://changefeed.app) we made the mistake of building a useful product that nobody knows they need. It requires sales to be successful (a skill I do not have) and is harder to sell than a product in an already existing market.

For my next business, I want to build something in a large market, something that people are already searching for on Google. Justin Jackson calls this “[Selling ice cream on a busy beach](https://justinjackson.ca/beach)”.

## Can I ship a first version in a few weeks?
Since I am bootstrapping, I cannot spend years (or really even months) on any given idea before figuring out if it’s valuable. While this can be hard to judge, I need to discard really hard problems, and ones outside of my expertise.

_Interestingly, this means the more I [[Invest in the best tools]] the more potential options are available to me, as it should take me less time to build things._

## Is my audience going to use it?
Ideally, my product should be useful for my existing audience. I spend a lot of time building it, so it would be great if that overlapped with the product I was building. (We bootstrapped [[Spectrum]] that way) 

Since my audience is largely like me, this would also mean I’m following [[Solve your own problems]].

## Bonus: Would it make my users money?
If my product (directly or indirectly) makes my users money, they are more likely to pay me a portion of that. A great example would be egghead.io: they make their instructors money directly, and their viewers money by helping them level up and get better jobs. Win-win!




# My life’s mission

[[🌱]]
Reflecting on my career, the common thread throughout my work seems to be that I make developers’ lives easier.

Obviously, all of my open source work is focused on this but even my most successful product, [[Spectrum]], ended up implicitly being focused on this! (and, in hindsight, we should have explicitly focused on it)

This is probably a byproduct of the fact that I [[Solve your own problems]]. Because I am a developer, it is easy for me to spot and solve problems developers have and I love doing it.




# Non-technical regrets at Spectrum
In hindsight, these are the things I wish we had done differently at [[Spectrum]].

## We should have focused on our niche
- Wanted to build for our use case, open knowledge communities (e.g. open source)
- Needed to make money, so built business features
- Businesses were not interested, and had we focused on our core niche use case we would have grown much faster
- Easier to raise money if necessary, more satisfying, ultimately more momentum and higher chance at success

**It's better to have 100 true fans who love the product than 10,000 barely interested users.**

## We should have done way more marketing
- Spend too much time on building new, shiny features
- Should have spent more time on marketing, typical engineering mistake
- Not just social media, but also in-product growth marketing, marketing pages on the website, etc.

**Do way more marketing and sales than you think you need to do.**

## We should have told our story
- Branding was good and unique
- Brand did not have minimum viable personality https://avc.com/2011/09/minimum-viable-personality/
- It differentiates you from the "big guys" as they cannot have a story as personal as you in most cases
- For example, The Practical Dev (dev.to) is doing a great job at this

**Tell your story, it differentiates you from the big players in a crowded market**

[[Sell your story]]

## We should have launched all the time
- Launches are great for marketing, every new release excites customers
- Shipped tons of stuff without ever announcing it
- Missed opportunity to showcase a new feature and get more users at the same time

**Launch before you think you are ready and then again and again and again.**

[[Always be shipping]]  [[Iterate as fast as possible]]

## We should have stuck with the MVP technology
- Had an early MVP built with minimal technology (Firebase)
- Saw some tiny traction, thought we needed proper setup for custom functionality

**Stick with the minimum viable technology setup until you feel product-market fit. You will know in your heart if you have it.**

## We Did Not Talk Enough With Users
- Went into a hole and built a shiny feature for months (e.g. WYSIWYG editing) without talking to users about their priorities
- Ended up wasting valuable time and adding a lot of unnessary tech debt

**Talk with your users, they should be your guiding star.**




Getting acquired was the most stressful time of my life

Non-technical regrets at Spectrum

Gatsby is inventing a new category of websites

My filters for potential business ideas

My life’s mission

# Spectrum
> _The community platform for the future._ – [spectrum.chat](https://spectrum.chat/)

I co-founded Spectrum in February 2017 with [Brian Lovin](https://twitter.com/brian_lovin) and [Bryn Jackson](https://twitter.com/superbryntendo). It is a public chat app for large online communities. Spectrum was acquired by GitHub in November 2018.

Here are some things I have written about that time:

- [Tech Choices I Regret at Spectrum](https://mxstbr.com/thoughts/tech-choice-regrets-at-spectrum) (blog post)
- [[Non-technical regrets at Spectrum]]
- [[Getting acquired was the most stressful time of my life]]




@mxstbr's working notes

Spectrum

Links to this note

What do you think?