No stack: a leaner tech startup approach
The days when building a new technology meant finding thousands of dollars and full-stack programming knowledge are over. Tech entrepreneur Stoyan Mitov welcomes us to the era of no-stack API-based startups with shorter development times and less expenses.
Do you remember the 90s when you were supposed to build everything from scratch and spend tens of thousands of dollars for servers, storage, software licenses and connectivity? What about the early 00s – when you needed at least $1,000 for a minimal server set-up? Or the last five years with the cloud infrastructure where you pay less, but still need to know Java, Ruby, Node.js and MongoDB to start with.
If the answer is yes, one can easily realise what an impediment to start something it was if you didn’t have the money or the needed tech skills. Nowadays, it’s different. The no-stack, API-based startups are here and will rule the near future. Today, you don’t need thousands of dollars or full-stack tech skills to build a product from scratch. You can start directly selling on Facebook and communicate and accept payments through WeChat. You can use Twilio and Stripe and have a ready mobile sales solution. In this article I’ll share a personal story from my startup, a story from my great Java team at Dreamix and some pros and cons of the no-stack dev.
The stack-free team
Six months ago I co-founded a mobile app called Grajdanite (“Citizens”) with which we want to solve the problem of bad road behaviour by encouraging and enabling responsible and polite citizenship. We’re four co-founders and due to limited resources, we didn’t wanted to spend a lot of time coding the entire app, so we integrated a ready-built platform for the backend part. Its stack includes Ionic, Cordova and AWS. It saved us a lot of dev time which we invested in creating content, testing and building a better UI and UX. Moreover, it greatly helped us in handling the load, once we had 1,300 new registration in less than a minute. So we were freed from worrying whether it may go down, and also the platform provided a great administration panel with nice charts and statistics. The no-stack approach here proved useful for us and allowed us to release a stable mobile app really fast, that made it to #1 in the Apple Store Bulgaria.
Another story is from a project we had with my great Java dev team at Dreamix. We help startups out with tech expertise, and the last one we partnered with called Twibble.io is using Angular.js, Spring, Elasticsearch and out-of-the-box APIs for its payments subscriptions using Stripe, for sending mails through Mandrill, for collecting user feedback with UserVoice, for security and web optimisation via Cloudflare, hosted on Rackspace and fully distributed on Twitter using its APIs. We were also using tools like Mixpanel and Optimizely for tracking user behaviour.
These API solutions allowed the team to concentrate on the content, product development, user engagement, UI and UX improvements, etc. Twitter gave them instant access to more than 300 million potential customers with a ready distribution channel. In both cases for Citizens and Twibble, the no-stack approach was beneficial and added great value to these startups, such as a vast decrease in development time and expenses.
The last story is how Dreamix helped a Bulgarian startup called Vitrinaapp.eu used by Bulgarian artists to sell their crafts worldwide via a mobile store. They had a great designer and front-end developer, but no backend expertise. Shortly after they contacted us, we started by spending some hours researching the best way to provide a backend exposed via REST API. Thinking lean, we discovered that using Mongolab.com for storing the data, Cloudinary.com for image processing and Mandrill.com for sending mails with orders would do the job.
Right after that, we discovered Moltin.com which is a ready-to-use backend-as-a-service for e-commerce. Basically everything we needed was there and the front-end developer could use the REST services right away. The no-stack approach worked extremely well for building the prototype of the application. We were able to integrate the backend very quickly, and the feature list it provided was something we could not achieve if we had started implementing a custom solution.
The limits of no stack
However, the no-stack approach has its limitations and impediments that are not so important in the beginning, but that are annoying after time. First of all, whenever a company uses a third-party solution, it loses its nimbleness. To give you an example – we wanted to release Citizens for Windows Phone users, but it happened that due to the Ionic in our backend service we used, it did not support windows phones, meaning we still don’t have Citizens for our windows users.
Additionally, what happens if Twitter closes its API – it’s easy, Twibble will die in a day because it’s fully dependent on it for distribution. Twibble’s business model is based on Twitter’s open API. Moreover, the margins when a third party solution is used will be less once your company achieves certain level of growth – which isn’t good for the evaluation. There are other negatives as well, but the above are the most important.
This platform-based APIs no-stack approach wasn’t here 10 or even 5 years ago. More than 90% of the companies that base their business model on services they provide through open APIs are startups. This is the current trend and it has a lot of benefits and also some negatives. However, it’s great because the no-stack approach removes many limitations such as high costs and great dev skills usually needed to start a tech company. If you consider China, where you just need to list your merchandise at Taobao and install WeChat to communicate and accept payments, it’s a great opportunity for rural, poor people with family to grow a small business.