#launchfaster | Explore Tumblr posts and blogs

andrecasal · 10 months ago

Text

How to reach a wide audience on social media

If you're new here, I'm André, a tech entrepreneur and founder of LaunchFast, a stack designed to help web developers significantly speed up their project development time. I post daily updates on my journey and progress.

Almost every day I share my journey in public. I do this on 24 different platforms. Here's how you can do it too:

Write an article in Markdown

Open the 24 browser tabs at once

Copy and paste the article into each tab

That's it! You're done. You've reached a wide audience on social media.

Here's my marketing tab:

And here's how you can open all the websites with one click:

Keep in mind that some of those, like the Smashing Magazine, aren't meant for daily shares but for in-depth web dev article submittions. But I like to keep them there to remind me to submit them!

I do go the extra mile and use invideo.io to create a video version of the article. It's the same process: copy and paste the article and it will spit out a video for you. I then download the video from invideo.io and upload it to Instagram, TikTok, and YouTube Shorts. Funny enough, the short videos get a lot of impressions (around 600 each).

"Sure André, but what did you do today? Where's the progress report?!"

Ah, right. To contexualize, I've decided to work on improving my landing page until I reach a 10%+ conversion rate, because being a tech entrepreneur, marketing has always been my Achiles heel. So here's the plan:

Improve the landing page until I reach a 10%+ conversion rate

Create a marketing strategy that I can implement in 2h/day

Improve the product

Growth hacking strategies

So here's what got done today:

Figured out what to insert on that green area of the hero for the landing page

As you know, the above the fold section of a landing page is extremely important. And the image much more so, because we process images considerably faster than text. So that visual (whether it is an image, video or an animation), needs to trigger immediate engagement and an emotional response, while reinforcing the product’s core value proposition. It's a bit of a tall order. But it's well worth the effort.

So today I've spent the entire day coming up with ideas and rating them based on that criteria and I settled on a "Commit & Push" simulation, so that web developers can feel how it is to push code to production with LaunchFast: ESLint, TypeScript, Vitest, and Playwright run and the code is deployed to production.

The difference between this visual being crappy or impactful depends on the execution and attention to details of the animation. So that's what I'm going to be doing tomorrow.

And that's it for today folks! I hope you enjoyed this one, cheers and have a great rest of your week! 👋

#buildinpublic #web development #marketing

0 notes

deandacosta · 10 months ago

Text

LaunchFast

http://dlvr.it/TCQdSW

0 notes

andrecasal · 10 months ago

Text

Improving the conversion rate 💰

After spending a few days with my family, I’m back developing LaunchFast.

I’ve been thinking about what to do next. Here’s the plan:

Improve conversion rate

Create marketing strategy

Improve the product

As a perfectionist, I promised myself I wouldn’t touch the product until I dealt with marketing: my Achilles heel.

Improve conversion rate

I’ve been thinking about how to make my landing page better. One of the things I realized is that I shouldn’t be comparing LaunchFast with any other boilerplate. LaunchFast stands on its own.

So I’m studying high-converting landing pages and working on a new landing page to boost conversions. This is the work in progress so far:

LaunchFast gets 40 users/day. I won’t stop working on this until I get at least a 10% conversion rate.

Create marketing strategy

I’ve already created a thorough 3-step marketing strategy:

Identify Ideal Channels & Tactics

Develop Testing & Validation Framework

Scaling Successful Channels

If you’d like to create a thorough marketing strategy, use this prompt on ChatGPT: "Act as a seasoned marketer and unconventional growth hacker to assist me in crafting a strategy for {Business Type} owners in {Location} on identifying, testing, and scaling new marketing channels and ideas to reach {Target Audience} with {Product/Service}."

Improve the product

I’m super eager to start working on the product again and delight my customers with awesome new automations and functionality. In particular:

React Email 3.0

Beautiful online video documentation 🎥✨

That’s it for now folks!

#buildinpublic #web development #marketing

0 notes

andrecasal · 10 months ago

Text

Public roadmap 🗺️

Here's the menu for today 📖

Asked customers for feedback

Add upvotes to the roadmap

Allow people to discuss roadmap features publicly on 𝕏

Add mailing list for product updates

Spoke to Jan Sulaiman, Global Director at 1NCE about database performance needs

Lisboa Innovation For All

Current metrics

Next steps

Let's get to it.

I’ve engaged with my customers, asked how their experience had been, and asked for feedback

Today I’ve sent an email to all the people who bought LaunchFast.

I’ve asked for their feedback and haven’t received any replies yet, but I want to make them feel supported and that I’m here to help if they get into trouble or find any problems with the product.

Added upvotes to the roadmap

I’ve improved the current roadmap so customers can vote on their preferred features.

Non-customers can still see the roadmap, but cannot upvote.

This is how the roadmap looks at the moment.

(This is a screenshot from my local dev environment, that’s why there are no upvotes.)

Allow people to discuss features on 𝕏

You’ll also notice that every feature has a “Discuss on 𝕏” button. This isn’t in production yet, but it will be tomorrow.

Since the repo is private, users can click that button and discuss this feature, in public, on 𝕏. Each feature has a corresponding post with a small description, like so 👇

The downside is that users need an account on 𝕏, but I’ll try it like this for now and see how it goes.

Added a mailing list that users can subscribe to, for product updates

I’ve also added a newsletter subscription form for users who want to stay up-to-date with LaunchFast as new features are released.

If you’re one of them, feel free to subscribe!

Spoke to Jan Sulaiman, Global Director at 1NCE about database performance needs

I’ve spoken to Jan Sulaiman, Global Director at 1NCE, an IoT company, about their database performance needs. According to Jan, hitting the 500k writes/sec performance limitation of SQLite would “require hundreds of millions of devices.”

According to Jan (slightly edited for brevity): "[As] a very rough estimation, right now, we have around 5 Mio active devices. Our customers send, on average, one message per 15 minutes.

So that means we average 5556 messages/second.

This would also align with our overall Downlink/Uplink capacity. For our European Breakout, for example, we are currently averaging around 40 Mb/s downlink and 75 Mb/s uplink traffic. And that Breakout is handling around 2,2 Mio active devices.

Since you ask about write operations, we only need to look at the 75 Mb/s. Here I assume an average of 2 KB per message that needs to be written. If I use the bandwith, I also get roughly 4578 write operations per second.

So, it's pretty close to the first calculation.

Long story short - while we probably have quite a high number of operations we need to handle and millions of active devices - we still would never get to 500k+ transactions per second 😁"

This ties into first-principles thinking and my explanation for choosing SQLite over any other database (MySQL, Postgres, MongoDB, etc), even if hosted on the same machine - SQLite is a zero-configuration, zero-latency database, and it’s just a file, making it dead simple to manage. Other databases require you to manage a server, connections, and authentication (offering another attack surface for hackers), and you won’t benefit from their higher performance anyway.

Hosted databases like Firebase and Supabase solve this problem by managing the database for you, but you pay an even higher cost: your performance is now subjected and limited to the network’s bandwidth and latency.

In the best-case scenario, you add a 10 to 30ms overhead to every single query you make (this should be enough not to use them), and in the worst-case scenario, the database is being DDoS’d and you can’t connect to it, making your app dysfunctional.

But I digress…

So what’s the opportunity here?

Jan agreed to be my guest on one of the videos I will do as part of LaunchFast’s documentation 🎉

Lisboa Innovation For All

Lisboa innovation for all (https://lisboainnovationforall.com) is a social innovation prize from the Lisbon City Council, organized by the Unicorn Factory Lisboa and supported by the European Innovation Council, which aims to discover and support innovative and impactful solutions that can be applied practically in the city of Lisbon.

They’re offering 360.000€ for projects on education, healthcare, and migration, and now that LaunchFast has been released, it would be a perfect opportunity to show, in public, what a developer is capable of with a powerful tool like LaunchFast.

Current Metrics

LaunchFast will launch on @MicroLaunchHQ on the 1st of September: https://microlaunch.net/p/launchfastpro

MicroLaunch is a relatively new platform created by Said, and I’ve found a few errors, but I look forward to seeing how LaunchFast does on microlaunch and how much traffic it will bring.

At the moment, LaunchFast is hovering at around 40 users per day.

Next Steps

This was the plan yesterday:

Engage more with Product Hunters ahead of the next launch (after payment and AI integrations potentially)

Create the documentation for LaunchFast, which includes video format that will also serve as content for social media

Integrate payments and AI into LaunchFast

Allow customers to suggest and prioritize items in the roadmap ✅

Engage with current customers to assess their experience and potentially fix pain points ✅

Add a newsletter component to the landing page to allow users to get notified of updates to the stack ✅

As for the next steps, I don’t know in which order I will do them, but this is the general plan:

Engage more with Product Hunters ahead of the next launch (after payment and AI integrations potentially)

Create the documentation for LaunchFast, which includes video that will also serve as content for social media

Integrate payments and AI into LaunchFast

Improve the current directory (https://launchfast.pro/launch-directories)

Possibly apply to “lisboa innovation for all”

That’s it for today, folks!

Have a great weekend and see you tomorrow!

P.S.: If you’re interested in LaunchFast, feel free to discuss and vote (https://x.com/andrecasaldev/status/1829538090135982455) on the features you’d like to see come onto the product!

#buildinpublic

0 notes

andrecasal · 10 months ago

Text

Product Hunt aftermath 🔥

Ok, so today is the day after launching LaunchFast on Product Hunt.

I don’t have anything to compare this launch against, to know how successful it was, so I’m taking it as my baseline and figuring out what I can improve for the next one.

Let’s get into it. Here are the metrics for this launch:

271 different people visited the landing page

Made a few sales

Ended in 11th place for the day

Things I will do better for the next launch

Getting votes on Product Hunt is much like a political campaign. You have to get in front of people and give them good reasons to vote for you.

Exposing your product doesn’t matter much if the people you expose it to aren’t Product Hunt users with accounts that are at least 2 weeks old, because new accounts can’t upvote products.

So, on my next launch, I will engage much more and much earlier with people who are active on Product Hunt.

This is also my target audience for LaunchFast, so win-win.

Missed opportunities

The other thing I feel was a missed opportunity was not having a newsletter component on the landing page to capture emails from users who might be on the fence about buying LaunchFast before it has payments or AI integrations.

The other one was that I could have pushed it a little further to get 2 more votes and ended in the top 10. The benefit of ending in the top 10 is that LaunchFast would have been featured on Product Hunt’s Daily Digest, which is read by millions? (I forget of exact value) of readers.

Moving forward

Now that the product has officially been launched, it’s time to look forward.

Here’s the plan:

Engage more with Product Hunters ahead of the next launch (after payment and AI integrations potentially)

Create the documentation for LaunchFast, which includes video format that will also serve as content for social media

Integrate payments and AI into LaunchFast

Allow customers to suggest and prioritize items in the roadmap

Engage with current customers to assess their experience and potentially fix pain points

Add a newsletter component to the landing page to allow users to get notified of updates to the stack

And that’s it folks. Overall, I’m feeling very positive about this experience and I think I’ll do a considerably better job on the next launch.

That’s it for today folks. Have a great rest of your week 🙌😄

#buildinpublic

0 notes

andrecasal · 10 months ago

Text

LaunchFast, a tool for devs, launches today on Product Hunt 🔥

#buildinpublic

0 notes

andrecasal · 10 months ago

Text

Building an NPM package

Today I worked to improve the “npx -y launchfast” command. Here’s what got done:

Improved text colors

Added TS support

Added a stepper

Absorbed the input for folder input

Explained why the script needs a GitHub Private Access Token

Add support for multiple package managers (npm, yarn, pnpm)

Show the number of steps

Improved text colors

One of the CLI UI components is called a note. A note has a title and a description. The description used to be faded out. I've improved the contrast ratio to make sure developers don't skip that section.

Added TS support

I caught myself running the script 3 times in a row and having basic type errors, so I decided to use TypeScript.

Developing with TypeScript is such a breeze ♥️

You avoid running invalid code because all type errors are gone, you get auto-completion, and GitHub Copilot’s suggestions are better.

Added a stepper

When you’re building a CLI, it’s often the case you need a stepper. This is a simple technique.

The steps variable is just an array of functions you call with a shared context, so they share data between themselves.

This allows me to isolate each step into its function, making the code more modular and manageable.

Absorbed the input for folder input

When npx -y launchfast is done getting the info from the user, it runs npx create-remix@latest. Unfortunately, the create-remix command doesn’t have the best error validation for the folder name, failing when the user inserts an absolute path (without ./).

Luckily, create-remix accepts a folder name as an argument, so I ask the user (with proper validation) the name of the folder he wants to create and pass that down to the create-remix command.

Here’s what that looks like at the moment: const command = npx --yes create-remix@latest ${ctx.folderName} --package-manager ${ctx.packageManager} --git-init --install --init-script --template andrecasal/launch-fast-stack --token ${ctx.privateAccessToken}

I’m happy my users don’t have to type that in manually 😄

Explained why the script needs a GitHub Private Access Token

I was afraid of making the prompts too long.

I sacrificed clarity because of it.

But after hearing almost every developer wonder what the GitHub Private Access Token was for, I changed the prompt to something more obvious:

“We need a Private Access Token (PAT) to download the template. Open GitHub to create one (use the default read-only scope, scroll down, and press "Generate token")?”

Add support for multiple package managers (npm, yarn, pnpm)

Adding support for multiple package managers was surprisingly easy.

It’s as easy as running npm --version and catching an exception.

Now the CLI supports installing LaunchFast with npm, yarn, or pnpm.

Show the number of steps

A nice tip I’ve got from testing with users was to show the user where he was in the process. So I’ve added a “Step n/total: ”. This makes navigation and managing expectations considerably easier.

Next Steps

Tomorrow I’ll try to make Fly’s CLI installation automatic. The challenge with this is not installing Fly’s CLI itself, but reloading the environment variables so that the fly --version command is recognized. I’ll let you know how I solve that.

Then I’ll apply all the other minor improvements to the CLI and test with new developers to see how they feel about it.

After that’s done, I’ll record a demo video to put on the landing page 🚀

That’s it, folks. I hope this post is useful to you. Feel free to dive into the open source code or ask me questions on 𝕏 in case you’re building an NPM package or a Remix Stack initialization script.

Cheers, and have a great weekend ♥️

#buildinpublic

0 notes

andrecasal · 10 months ago

Text

The Pareto Principle or the 80/20 rule 📏

I spent over 7 hours today watching video footage from test drives conducted by developers using my product. Interestingly, I observed that 90% of the feedback was generated during the first two test drives. The subsequent test drives identified the same issues repeatedly. Here, we see the Pareto Principle in action, where 80% of results came from just 20% of the efforts. To benefit from this, make sure to refine your product after each test drive with actual users. By doing so, you'll enhance your product much more rapidly. ## Improvements to the CLI Here's a list of key improvements for the CLI: 1. Package Manager Choice: Let users select their preferred package manager like npm, pnpm, yarn, or bun. 2. Automatic CLI Installation: Set up the Fly CLI automatically and refresh environment variables. 3. Clarify Requirements: Explain the necessity for Fly’s CLI and a GitHub Private Access Token, highlighting that LaunchFast deploys to Fly and uses these for setup. 4. User-Friendly Prompts: Modify the prompts from `create-remix` to be more forgiving, particularly addressing issues where the installation fails without "./" in the folder name. 5. Progress Indicators: Include step numbers (e.g., Step 5 of 12) to show users how far along they are in the process. 6. Enhanced Instructions: Use more vibrant colors in instructions to ensure they catch the user's attention. 7. Streamline Setup: Simplify by not asking for the App name upfront; users can change this later. 8. Editor Detection: Automatically detect installed code editors like VS Code or Cursor, and offer to open the project in their preferred editor. 9. Error Handling: Implement a try-catch block to display errors clearly if the script fails, aiding in quick debugging and recovery. 10. Additional Tweaks: Incorporate various smaller enhancements to polish the user experience. ## Next Steps The plan for tomorrow is to implement all these into the CLI. There's another session lined up for tomorrow afternoon. I'm excited to see how these changes will be received. That wraps up today's update. I hope you have a fantastic Friday ahead! ♥️

#buildinpublic

0 notes