Aggressive idea execution 🚀

Francisco Donate 📝 Edit 🕑

Back in 2014 after winning a NASA hackathon everyone was amazed at how fast I was making things, but I wasn't happy so I spent the last 5 years consciously improving on this.

The most important things I learned about executing ideas at high speed with an estimated of 5-10x speed boost each are: the right abstraction, a fast iterative process and a winner team.

Previous to all of this is learning to be brutally honest with yourself. You need to have a very clear understanding of where your weaknesses are and leave your ego apart to focus on what really matters to you. *"If you don't know yourself you will succumb in every battle."*

You might develop impostor syndrome by doing just that. Luckily, being honest with yourself also helps here! I feel like I'm a normal-to-mediocre programmer, but I know I'm biased so I can look at my progress from a more objective point of view.

The right abstraction

First you have to really know what you want to achieve. There are too many perverse incentives and distractions working against you! Sit down and really consider this.

Then work only for your goal and review this as often as needed. You might have several goals and that is fine, but be very conscious about your real goals and remove any unconscious goal. Ask yourself how to best reach your goal and ask knowledgeable friends about this.

Choose the right tool and path for the job, which will depend greatly on your personal situation. If you are an engineer with a business partner you'll work towards your goal better as a CTO, even if that doesn't give you all the fame (hopefully assuming fame is not your real goal).

This also has personal implications for learning about technology (Node.js? React?), different fields (design? marketing?) and about life itself (friends? hobbies?).

Let me give you some counter examples of how not to do things:

We are writing custom software because we want to launch a successful online store.

Too often business people think they need it and devs want to make it, so new ecommerce software is written. I'd suggest to instead trying these and only skip one if there's a strong real reason:

  1. Sell your products in online stores like Amazon, eBay, etc. This allows to start selling a product in few hours and see how the market reacts.
  2. Use one of the ecommerce solutions like Shoppify, Magento, etc. This allows for custom branding and better marketing techniques in few days.
  3. Customize wordpress with woocommerce or similar. This allows for lower fees and very custom functionality if your business really needs this and you can launch in few weeks.
  4. Integrate with 3rd party solutions like Stripe or Paypal. You would do this when adding payments into existing software, but I would not recommend it for an ecommerce.

We need to support 10 million concurrent users because everyone will love our idea.

Let me tell you a very controversial opinion: it's great to have scalability issues! While you might not believe that, I'll rephrase it as a question for you: what is a better risk to have?

Of course no one likes those late nights, but in the grand schema of things and with those two being at each end of the scale, I'd choose the first one any day.

We are discussing what pattern is better because we want to win a hackathon.

After half a day of debates, I can already tell you that your team is not going to win. Throw dirty code at it as fast as possible, change your ideas as you go and adapt. Absolutely no one will care about code quality when evaluating your competition project.

Fast iterative process

Great, after much consideration you have decided you can use a pre-made solution! Close this post, go there and get it done! Really!.

But if you decide to actually write software, you want to "fail fast, fail cheap". You want to be able to make great progress at every stage as fast as possible. My general process is this, with times for personal ideas spent in each stage (multiply by 2-3x for work/3rd party ideas):

  1. An idea comes. It is there and I'm thinking about it during my day to day down time.
  2. Many paper mockups (1-10 hours) with alternative designs help to materialize it. Once it's on paper I can attack it for weakness and change it fast.
  3. A quick prototype (1-10 days) is perfect to see if the core idea is useful and how people work with it. I make this quick and dirty with the idea of throwing the whole thing away.
  4. A Minimum Viable Product (10-100 days) that has the core functionality with the needed support systems like login, storage, interface, etc. It will reuse the core idea from the prototype, but not the rest. It helps to discover new business cases and features.
  5. Production (100+ days) with a proper deploy pipeline, testing and documentation. It has been improved from the MVP but it's flexible enough to accept new features.

The time will decrease with a team but not linearly because of combination effectiveness, so the points 1-3 are normally better done by a single person. Point 4 is debatable and point 5 should definitely be made by a team.

Once again the main issue here is people and their egos trying to go straight to point 5. "Why would my ideas be wrong? I'm an expert! It's the world/market/etc that is wrong!" Normally it gets funnier (or, sadder) the more "expert" someone believes they are.

As a practical example of how these steps might work for my personal project Core Cards:

  1. Anki interface felt so 1999, how would Spaced Repetition look like if it was made today?
  2. Made paper prototypes with a swipping interface which look like something I'd love using everyday.
  3. Made a prototype to test the spaced repetition algorithm since I'm not 100% convinced. Used a Node.js server with Google Spreadsheets for the data. Added Mongodb so that I can test it between server restarts.
  4. I rewrote it in React and localStorage so that I can use it anywhere. Changed it to Preact and IndexedDB from what I learned here.

Never "released for production*", but I've used it 5.000+ times and my Japanese has improved greatly because of it! *Mischief managed. Some day my objectives might change and I might take it to production and try to sell it.

A winner team

Even if you are great, a team of talented people will reach much further. But while two people can do much more than a single person, it's still not double.

There are communication and complexity flows to take into account! It's perfectly possible for some people not only to be very low on team throughput, but even be negative by bringing some of the productive members down!

That's why 3-6 highly skilled individuals can do much more than 10-20 normal ones. How many people with an effective productivity of 0 do you need to get the same output?

I've been lucky in life and made great friends in life. There was a moment where I wanted to make something amazing to go to another NASA hackathon.

First I called my friend from the previous competition and we joined together on Friday night. We brainstormed for ~1 hour to decide what to do, building a working hardware prototype in ~36 hours for the Sunday night.

It was going to be tough, but luckily we could get all of our first choices of each:

After convincing everyone we all met on that same friday night, they loved the idea and we changed a few details quickly. Time was tight so we made a rule: if you get stuck over 1h, throw it away and try something different.

Now I know that this was a hackathon. With this hyper-talented team it took a weekend to design, build, deploy, 3d-print and present an amazing prototype:

See Project Details

We were lucky that we could literally pick the best of each field. I don't think we could have gotten anywhere with a team made of your average person, and I shudder when thinking that half of the people are below average!

Final notes

I'm totally wrong, please feel free to correct me and I'll read and consider every single one of your opinions. Special things I'd love to learn: