I was recently at a public talk by Alan Jones at the UNE Smart Region Incubator where Alan talked about launching startups and developing ideas.
Alan put it quite eloquently that “With change comes opportunity” and we are all very capable of building the next best thing as technological barriers and costs are a lot lower than 5 years ago but Alan also mentioned 19 start-ups-ups fail but “if you focus on solving customer problems you have a better chance of succeeding”. Regions need to share knowledge and you can a learn from others peoples mistakes.”
I was asked after this event to share thoughts on “how do I learn to develop an app” and “how do you get the knowledge”. Here is my poor “brain dump” on how to develop software ideas (It’s hard to condense 30 years experience developing software). I will revise this post over the coming weeks so check back often.
If you have never programmed before check out this programming 101 guide here.
I have blogged on technology/knowledge things in the past at www.fearby.com and recently I blogged about how to develop cloud-based services (here, here, here, here and here) but this blog post assumes you have a validated “app idea” and you want to know how to develop yourself. If you do not want to develop an app yourself you may want to speak with Blue Chilli.
Find a good mentor.
True App Development Quotes
- Finding development information is easy, following a plan is hard.
- Aim for progress and not perfection.
- Learn one thing at a time (Multitasking can kill your brain).
- Fail fast and fail early and get feedback as early as possible from customers.
- 10 engaged customers are better than 10,000 disengaged users.
And a bit of humour before we start.
(click for larger image)
Here is a funny video on startup/entrepreneur life/lingo
This is a good funny, open and honest video about programming on YouTube.
Follow Seth F Samuel on twitter here.
Don’t be afraid to learn from others before you develop
My fav tips from over 200 failed startups (from https://www.cbinsights.com/blog/startup-failure-post-mortem/ )
- Simpler websites shouldn’t take more than 2-3 months.You can always iterate and extrapolate later. Wet your feet asap
- As products became more and more complex, the performance degrades. Speed is a feature for all web apps. You can spend hundreds of hours trying to speed of the app with little success. Benchmarking tools incorporated into the development cycle from the beginning is a good idea
- Outsource or buy in talent if you don’t know something (e.g marketing). Time is money.
- Make an environment where you will be productive. Working from home can be convenient, but often times will be much less productive than a separate space. Also its a good idea to have separate spaces so you’ll have some work/life balance.
- Not giving enough time to stress and load testing or leaving it until the last minute is something startups are known for — especially true of small teams — but it means things tend to get pretty tricky at scale, particularly if you start adding a user every four seconds.
- It’s possible to make a little money from a lot of people, or a lot of money from a few people. Making a little money from a few people doesn’t add up. If you’re not selling something, you better have a LOT of eyeballs. We didn’t.
- We received conflicting advice from lots of smart people about which is more important. We focused on engagement, which we improved by orders of magnitude. No one cared. Lesson learned: Growth is the only thing that matters if you are building a social network. Period. Engagement is great but you aren’t even going to get the meeting unless your top-line numbers reach a certain threshold (which is different for seed vs. series A vs. selling advertising).
- We most definitely committed the all-too-common sin of premature scaling. Driven by the desire to hit significant numbers to prove the road for future fundraising and encouraged by our great initial traction in the student market, we embarked on significant work developing paid marketing channels and distribution channels that we could use to demonstrate scalable customer acquisition. This all fell flat due to our lack of product/market fit in the new markets, distracted significantly from product work to fix the fit (double fail) and cost a whole bunch of our runway.
- If you’re bootstrapping, cashflow is king. If you want to possibly build a product while your revenue is coming from other sources, you have to get those sources stable before you can focus on the product.
- Don’t multiply big numbers. Multiply $30 times 1.000 clients times 24 months. WOW, we will be rich! Oh, silly you, you have no idea how hard it is to get 1.000 clients paying anything monthly for 24 months. Here is my advice: get your first client. Then get your first 10. Then get more and more. Until you have your first 10 clients, you have proved nothing, only that you can multiply numbers.
- Customers pay for information, not raw data. Customers are willing to pay a lot more for information and most are not interested in data. Your service should make your customers look intelligent in front of their stakeholders. Follow up with inactive users. This is especially true when your service does not give intermediate values to your users. Our system should have been smarter about checking up on our users at various stages.
- Do not launch a startup if you do not have enough funding for multiple iterations. The chances of getting it right the first time are about the equivalent of winning the lotto.
I have seen quite a few projects fail because:
- The wrong technology was mandated.
- The software was not documented (by the developers).
- The software was shelved because new developers hated it or did not want to support it.
Project Roles (hats)
It is important to understand the roles in a project (project management methodology aside) and know when you are being a “decision maker” or a “technical developer”. A project usually has these roles.
- Sponsor/owner (usually fund the project and have the final say).
- Executive/Team leader/scrum master (manage day to day operations, people, tasks and resources).
- Team members (UI, UX, Marketers, Developers (DevOps, Web, Design etc) are usually the doers.
- Stakeholders (people who are impacted (operations, owners, Helpdesk)).
- Subject Matter Experts (people who should guide the work and not be ignored).
- Testers (people who test the product and give feedback).
It can be hard as a developer to switch hats in a one person team.
How do you develop and gain knowledge
First, document what you need to develop (what problem are you solving and what value will your idea bring). Does this solution exist already? Don’t solve a problem that already exists.
Developing software is not hard, you just need to be logical, research, be patient and follow a plan. The hardest part can be gluing components together.
I like to think of developing software like making a car if you need 4 wheels do you have 4 wheels? If you want to build it yourself and save some money can you make wheels (make rubber strips with steel reinforced/vulcanized rubber, make alloys and add bearings and have them pass regulations) or should you buy wheels (some things are cheaper to make than other things)? Developing software can be easy if you know what your are doing and have the experience and are aware of the costs and risks. Developing software can lead you down a rabbit hole of endless research, development, and testing if you don’t know what you are doing.
I “need a webpage”:
- Research: Will Wix, Shopify or a hosted WordPress website do (is it flexible or cheap enough) or do I install WordPress (guide here) or do I learn and build an HTML website and buy a theme and modify it (and have a custom/flexible solution)?
I “need an iPhone and Android app”:
Research: You will need to learn iOS and Android programming and you may need a server or two to hold the apps data, webpage and API. You will also need to setup and secure the servers or choose to install a database or go with a “database as a service” like cloud.mongodb.com or google firebase.
Money can buy anything (but will it be flexible/cheap enough), time can build anything (but will it be secure enough).
Developing software can be easy if you know what your are doing and have the experience and are aware of the costs and risks but developing software can lead you down a rabbit hole of endless research, development and testing if you don’t know what you are doing.
Almost all systems will need a central database to store all data, you can choose a traditional relational SQL database or a newer NoSQL database. MySQL is a good/cheap relational SQL database and MongoDB is a good NoSQL database. You will need to decide on how your app talks to the database (directly or via an API (protected by OAuth or limited access tokens)). It is a bad idea to open a database directly to the world with no security. Sites like www.shodan.io will automatically scan the Internet looking for open databases or systems and report this as an insecure site to anyone. It is in your interest to develop secure systems in all stages of development.
CRUD (Create, Read, Update and Delete) is a common group of database tasks that you can do to prove you can read, write, update and delete from a database. While performing CRUD operations is a good to benchmark to also see how fast the database it. if a database is the slowest link then you can use memory to cache database values (read my guide here). Caching can turn a cheap server into a faster server. Learning by doing can quickly build skills so “research”, “do” and “learn”.
Most solutions will need a website (and a web server). Here is a good article comparing Apache and Nginx (the leading open source web servers).
Stacks and Technology – There are loads of development environments (stacks), frameworks and technologies that you can choose. Frameworks supposedly make things easier and faster but frameworks and technologies change (See 2016 frameworks to learn guide and 2017 frameworks to learn guide) frequently (and can be abandoned). Frameworks supposedly make things easier and faster but be careful most frameworks run 30% slower than raw server side and client code. I’d recommend you learn a few technologies like NGINX, NodeJS, PHP and MySQL and move up from there.
Apps can be developed for Apple platforms by signing up here (about $150 AUD a year) and using the XCode IDE. Apps can be developed for the Android Platform by using Android Studio (for about $20 (one off fee)). Microsoft has a developer portal for the Windows Platform. Google also has an online scalable database as a service called Firebase. If you look hard enough you will find a service for everything but connecting those services can be timely, costly or make security and a scalable solution impossible so beware of using as-a-service platforms. I used the Corona SDK to develop an app but abandoned the platform due to changes in the vendor’s communication and enforced policies.
If you are not sure don’t be afraid of ask for help on Twitter.
Twitter is awesome for finding experts
Recent twitter replies to a problem I had.
Learning about new Technology and Stacks
To build knowledge you need to learn stuff, build stuff, test (benchmark), get feedback and build more stuff. I like to learn about new technology and stacks by watching Udemy courses and they have a huge list of development courses (Web Development, Mobile Apps, Programming Languages, Game Development, Databases, Software Testing, Software Engineering etc).
I am currently watching a Practical iOS 11 course by Stephen DeStefano on Udemy to learn about unreleased/upcoming features on the Apple iPhone (learning about XCode 9, Swift 4, Whats new in iOS 11, Drag and drop, PDF and ARKit etc).
Udemy is awesome (Udemy often have courses for $15).
If you want to learn HTML go to https://www.w3schools.com/.
https://devslopes.com/have a number or development related courses and an active community of developers in a chat system.
I would recommend you use Twitter and follow keywords (hashtags) around key topics (e.g #www, #css, #sql, #nosql, #nginx, #mongodb, #ios, #apple, #android, #swift, #objectivec, #java, #kotlin) and identify users to follow. Twitter is great for picking up new information.
Companies like https://www.civo.com/ offer developer friendly features with hosting, https://www.pebbled.io/ offer to develop for you and https://serverpilot.io/ help you spin up software on hosting providers.
What To Develop
First, you need to break down what you need. (e.g ” I want an app for iOS and Android in 5 months that does XYZ. The app must be secure and be fast. Users must be able to register an account and update their profile”).
Choosing how high to ensure you development project scales depends on your peak expected/active concurrent users (ratio of paying and free users). You can develop your app to scale very high but this may cost more money initially, it can be bad to pay to ensure scalability early. As long as you have a good product and robust networking/retry routines and UI you don’t need to scale high early.
Once you know what you need you can search the open source community for code that you can use. I use Alamofire for iOS network requests, SwiftyJSON for processing JSON data and other open source software. The only downside of using open source software is it may be abandoned by the creators and break in the future. Saving you time early may cost you time later.
Then you can break down what you don’t want. (e.g “I don’t want a web app or a windows phone or windows desktop app”). From here you will have a list of what you need and what you can avoid.
You will also need to choose a project management methodology (I have blogged about this here). Having a list of action item’s and a plan and you can work through developing your app.
While you are researching it is a good idea to develop smaller fun projects to refine your skills. There are a number of System Development Life Cycles (SDLC’s) but don’t worry if you get stuck, seek advice or move on. It is a good idea to get users beta testing your app early and seek feedback. Apple has the TestFlight app where you can send beta versions of apps to best testers. Here is a good guide on Android beta testing.
If you are unsure about certain user interface options or features divide your beta testers and perform A/B or split testing to determine the most popular user interfaces. Capturing user data and logs can also help with debugging and user usage actions.
Develop smaller proof of concept apps in new technologies or frameworks and you will build your knowledge and uncover limitations in certain frameworks and how to move forward with confidence. It is advisable to save your source code for later use and to share with others.
I have shared quite a bit of code at https://simon.fearby.com/blog/ that I refer to from time to time. I should have shared this on GitHub but I know Google will find this if people want it.
Get as much feedback as you can on what you do and choose (don’t trust the first blog post you read (me included)).
Most companies offer Webinars on their products. I like the NGINX webinars. Tutorialspoint have courses on development topics. Sitepoint is a good development site that offers free books, courses, and articles. What are API’s information by Programmable web.
You may want to document your application flow to better understand how the user interface works.
I use the Sublime Text 3 text editor when editing websites locally.
(image courtesy of https://www.sublimetext.com/ )
I use the Mac Paw app to help test API’s I develop locally.
(image courtesy of https://paw.cloud )
Snippets is a great application for the Mac for storing code snippets.
I use the Cornerstone Subversion app for backing up my code on my Mac.
I use the Forklift file manager on OSX.
Siege is a command line HTTP load testing tool.
http://loader.io/ is a nice web-based benchmarking tool.
Bootstrap is an essential mobile responsive framework.
Atlassian Jira is an essential project tracking tool. More on Agile Epics v Stories v Tasks on the Atlassian community website here. I have a post on developing software and staying on track here using Jira.
Jsfiddle is a good site that allows you to share code you are working on or having trouble with.
Dribbble is a “show and tell” site for designers and creatives.
Stackoverflow is the go-to place to ask for help.
Things I care about during development phases.
Concentrating too much on one facet can risk exposing other facets. Good programmers can recommend a deliver a solution that can be strong in all areas ( I hate developing apps that are slow but secure or scalable and complex).
Use my link to get a free Digital Ocean server for two months by using this link. Read my blog post here to help setup you VM. You can always use Ubuntu on your local machine to use Ubuntu (read my guide here). Don’t forget to use a GIT code repository like GitHub or Bitbucket.
Locally you can install Ubuntu (developers edition) and have a similar environment as cloud platforms.
- Deploy servers close to the customers (Digital Ocean is too far away to scale in Australia).
- Accessibility and testing (make things accessible from the start).
- Backup regularly (Use GIT, backup your server and use Rsync to copy files to remote servers and use services like backblaze.com to backup your machine).
- Transportability of technology (Use open technology and don’t lock yours into to one platform or service).
- Cost (expensive and convenient solutions may be costly).
- Buy in themes and solutions (wrapbootstrap.com).
- Do improve what you have done (make things better over time). Thing progress and not perfection.
There is no shortage of online comments bagging certain frameworks or platforms so look for trends and success stories and don’t go with the first framework you find. Try candidate frameworks and services and make up your own mind.
A good plan, violently executed now, is better than a perfect plan next week. – General George S. Patton
Sometimes cost is not the deciding factor (read my blog post on Alibaba cloud). You should estimate your apps costs per 1000 users. What do light v heavy users cost you? I have a blog post on the approx cost of cloud services. I started researching a scalable NoSQL platform on IBM Cloudant and it was going to cost $4,000 USD a month and integrating my own App logic and security was hard. I ended up testing MongoDB Cloud where I can scale to three servers for $80 a month but for now, I am developing my current project on my own AWS server with MongoDB instance. Read my blog post here on setting up MongoDB and read my blog post on the best MongoDB GUI.
Here is a great infographic for viewing what’s involved in mobile app development.
You can choose a number of tools or technologies to achieve your goals, for me it is doing it economically, securely and in a scalable way that has predictable costs. It is quite easy to develop something that is costly, won’t scale or not secure or flexible. Don’t get locked into expensive technologies. For example, AWS has a user pays Node JS service called Lambada where you get Million of free hits a month and then you get charged $0.0000002 per request thereafter. This sounds good but I prefer fixed pricing/DIY servers better as it allows me to build my own logic into apps (this is more important than scalability).
Using open source software of off the shelf solutions may speed things up initially? Will It slow you down later though? Ensure free solutions are complete and supported and Ensure frameworks are helping. Do you need one server or multiple servers (guide on setting up a distributed MySQL environment )? You can read about my scalability on a budget journey here. You can speed up a server in two ways Scale Up (Add more Mhz or CPU cores) or scale out (add more servers).
Start small and use free frameworks and platforms but have a tested scale up plan, I researched cheap Digital Ocean servers and moved to AWS to improve latency and tested MongoDB on Digital Ocean and AWS but have a plan to scale up to cloud.mongodb.com if need be.
Remember outsourcing work tasks (or complete outsourcing of development) can buy you time and or deliver software faster. Outsourcing can also introduce risks and be expensive. Ask for examples of previous work and get raw numbers on costs (now and in the future) and concurrent users that a particular bit of outsourcing work will achieve.
If you are looking to outsource work do look at work that the person or company has done before (if is fast, compliant, mobile scalable, secure, robust, backup up, do you have rights to edit/own and own the IP etc). I’d be cautious of companies who say they can do everything and don’t show live demos.
Also, beware of restrictions on your code set by the contractors. Can they do everything you need (compare with your list of Moscow must haves)? Sometimes contractors only code or do what they are comfortable with that can impact your deliverables.
Do use a private Git repository (that you own) like GitHub or BitBucket to secure your code and use software like Trello or Atlassian JIRA to track your project. Insist the contractors use your repository to retain control.
You can always sell equity in your idea to an investor and get feedback/development from companies like Bluechilli.
Monetization and data
Do have multiple monetization streams (initial app purchase cost, in-app purchase, subscriptions, in-app credit, advertising, selling code/components etc). Monthly revenue over yearly subscription works best to ensure cash flow.
Capture usage data and determine trends around successful engagement, Improve what works. Use A/B testing to roll our new features.
I like Backblaze post on getting your first 1,000 customers.
Maintenance, support risk and benefits
Building your own service can be cheaper but also riskier, if you fail to secure an app you are in trouble if you cannot scale you are in trouble. If you don’t update your server when vulnerabilities come out you are in trouble. Also, Google on monetization strategies. Apple apps do appear to deliver more profits over Android. Developers often joke “Apple devices offer 90% of the profits and 10% of the problems and Android apps offer 90% of the problems and 10% of the profits”.
Do inform you users with self-service status pages and informative error messages and don’t annoy users.
Use Free Trials and Credit
Most vendors have free trials so use them
https://aws.amazon.com/free/AWS have 12 month free tiers.
Use this link to get two months free with Digital Ocean.
Don’t be afraid to ask.
MongoDB Cloud also gives away free credit if you ask.
Sites like Shodan.io will quickly reveal weaknesses in your server (and services), this will help you build robust solutions from the start before hackers find them. Read https://www.owasp.org/index.php/Main_Page to know h0w to develop secure websites. Listen to the SecurityNow podcast to learn how the technology works and is broken. Following TroyHunt is recommended to keep up to date with security in general. @0xDUDE is a good ethical hacker to follow to stay up-to date on security exploits also @GDI_FDN is a good non-profit organization that helps defend sites that use open source software.
White hack hackers exist but so do black hat ones.
I use the ASafaWeb site to test your sites from common ASP security flaws. If you have a secure certificate on your site you will need to ensure the certificate is secure and up to date with the SSL Labs SSL Test site.
Once your websites IP address is known (get it from SSL Labs) run a scan over your site with https://www.shodan.io/ to find open ports or security weaknesses.
Shodan.io allows you and others to see public information about your server and services. You can read about well-known internet ports here.
Anyone can find your server if you are running older (or current) web servers and or services.
It is a good idea to follow security researchers like Steve Gibson and Troy Hunt and stay up to date with live exploits. http://blog.talosintelligence.com is also a good site for reading technical breakdowns of exploits.
Do share and talk about what you do with other developers. You can learn a lot from other developers and this can save you loads of time and mistakes. True developers love talking about their code and solutions.
Quite a lot of time can be spent on deciding on what technology or platform to use, I decide by factoring in cost, risk and security over flexibility, support and scalability. If I need flexibility, lower support or scalability then I’ll choose a different technology/platform. Generally, technology can help with support. Scalable solutions need effort from start to finish (it is quite easy to slow down any technology or service).
Don’t be afraid to admit you have chosen the wrong technology or platform. It is far easier to research and move on than live with poor technology.
If you have chosen the wrong technology and stick with it, you (and others) will loath working with it (impacting productivity/velocity). Do you spend time swapping technology or platforms now or be less productive later?
Intellectual property and Trademarks
https://namechk.com/ is also a good place to search for your app ideas name before you buy or register any social media accounts.
Using https://namechk.com/ you can see “mystartupidea” name is mostly free.
And the name “microsoft’ is mostly taken.
See my guide on how to get useful feedback for your ideas here.
- Use Git Source Control systems like GitHub or Bitbucket from the start and offsite backup your server and environments frequently. Digital Ocean charges 20% of your servers costs to back it up. AWS has multiple backup offerings.
- Start small and scale up when needed.
- Do lots of research and test different platforms, frameworks, and technologies and you will know what you should choose to develop with.
You will know when you are a developer when you have gained knowledge and experience and can automatically avoid technologies that will not fit a solution.
Don’t be afraid to share what you know (read my blog post on this here). Sharing allows you to solidify your knowledge and get new information. Shane Bishop from EWWW Image Optimizer WordPress plugin wrote Setting up a fast distributed MySQL environment with SSL for us. If you have something to share on here please let me know here on twitter.
It’s never too late to do
One final tip is knowledge is not everything, planning and research is key, a mind that can’t develop may be better than a mind that can because they have no experience (or baggage) and may find faster ways to do things. Thanks to http://zachvo.com/ for teaching me this during a recent WordPress re-deployment. Sometimes the simplest solution is.
Donate and make this blog better
Ask a question or recommend an article
DRAFT: 1.85 added white hack info (Saturday 22nd July 6:01pm AEST).