I’ve been programming since 1983 when I wrote an accounting app for my first business (they were even called ‘apps’ back then – the holy grail was to write a ‘killer app’ – as it still is today).
I went from a BBC Model B with 32k RAM, to an Atari 502 ST, to a Windows machine, to Apple and finally to Linux.
I’ve created and run 2 software companies, the latter winning 3 Apple nominations at the UK Apple awards in London during the late 2000’s.
Every time I switch technology I research it thoroughly to the point of obsession.
For example, when we switched to Java in 1999, I had to figure out if it was the choice for the future. I used Amazon’s advanced search at the time to see what books had been planned for the future, and Java was the clear winner.
When creating a bunch of membership sites, I needed to find the best CMS, best protection plugin, best theme, best hosting, best autoresponder and best shopping cart. I trialled 3 separate combinations and found the best performing one.
I do all this research and testing because I know my future reputation depends on my delivery. And my reputation matters greatly to me.
As Peter Diamandis (XPrize founder) says “Perfection is not an option”. Good enough is never good enough if you truly care about what you’re doing (never listen to those who tell you perfection is something to avoid, they’re wrong).
In 2016, Google announced changes to their algorithm including the public revealing of their long-standing AI program they call RankBrain.
It’s quite obvious they will only remain king of search if they deliver up the best results (with or without ads).
So it makes sense they should use artificial intelligence to figure out the deep semantics involved in trying to deliver those results.
And now they can do it better than most humans.
Which means the nature of search is finally changing for the better.
I also noticed that my best articles were starting to rank higher without me doing a thing. They were doing this on merit.
So I had an idea for a new app. One that would talk directly with Google and deliver up the sort of content I would need to start doing in the future if I wanted to rank articles with ease.
As of right now, there’s no magic bullet software that will write innovative, unique and truly useful articles that will help us humans get the answers we need.
So there has to be a three pronged approach to this:
- Deep research into the subject matter
- Great writing to bring that subject to life
- An understanding of what Google thinks about it
And so SEORoadmaps was born. It’s been over 12 months in the making, but the beauty is, the remit has never changed.
The problem I had was choosing the right software and hardware stack to use for this.
Research is always first, so the programming language of choice was first on the bill.
All the usual suspects were present:
- Java
- C++
- PHP
- Python
- Ruby
I won’t bore you with why the decision to use PHP was made, but that was the obvious choice in the end (after extensive Railing with Ruby, which nearly won the day).
The IDE of choice was PHPStorm, which although still has many ‘quirks’ is still worthy enough to be called an IDE.
Bitbucket was the choice for a remote repository and Git for local.
Laravel was used for the framework, and Forge and Envoyer for deployment and management.
But the really tough decision was the hosting.
I’m a real fan of Amazon and everything they’ve done (except perhaps for employee relationships). An earlier company I had was chosen by Amazon UK as a vendor and they helped promote us through their channels many times.
But it meant bootstrapping a whole bunch of Amazon goodies together to get what I needed. That kind of overhead is the last thing I wanted for a new startup.
Which meant looking for a second tier option (because Amazon’s size means all others in the marketplace are minnows by comparison).
So to cut a long story short, I went with Digital Ocean (or DO as they’re often called).
It’s a 100% Linux cloud based hosting company with global penetration. When you need a new server, you ‘spin up’ a droplet. A droplet is a dedicated piece of “hardware” complete with memory and CPU with a wide choice of specification (even down to the Linux OS you want – I’m using Ubuntu).
You’ll need to know how to use a terminal to talk to your new server, but that’s easily mastered.
It also helps to get direct access to the NGINX config file, which is absolutely vital when you first set it up, and rather than going through the terminal, I decided on the Laravel Forge option to manage my droplets. They have a simple one-click button to view and edit it. Perfect.
As the app grows I can add more powerful servers and load balancing at the touch of a button.
In short, they offer everything I need, but the best bit is the cost.
The cheapest droplet, which is absolutely fine for a simple app, is just $5, but if you get referred by a friend, you get a $10 credit to get you going. That’s two months free.
And should you then continue to host with Digital Ocean and spend $25, your referrer is rewarded again. So you both win.
To do that, Digital Ocean give each user a referral link. Mine is below if you like the idea of that $10 credit.
[thrive_link color=’blue’ link=’http://www.digitalocean.com/?refcode=bd04870249f6′ target=’_self’ size=’medium’ align=”]Get Your Digital Ocean $10 Credit[/thrive_link]
You can delete droplets any time you want, so they’re great for testing, and then start new ones. You can SSH into them, you can add SSL, in fact you can do pretty much anything you want with them, and at a cost second to none.
Good luck with you next app. I hope it’s killer 🙂
Leave a Reply