Reminded by Iren's post, I'd like to share a piece written by my co-founder about another software development methodology. We learned it a couple of years ago in rspective and we still use it to build Voucherify today. It's called Sparta and it's super satisfying for devs. Here's why.

--

It is unlike any other project, we have ever had our hands on. In every aspect. Forget everything you ever knew about creating software, just do it better, whatever it takes, btw I need it in 3h, can you make it?

It’s been a few months now, since we started working with a startup. They are a raising star on the market of professional services in Germany. They prove to be heavily successful, with the numbers skyrocketing one day after another. They’ve done it before and they will do it again. They know exactly what they are doing, why, and which direction they are going. They think big and go big, laughing at the big-fish competitors with way more funding.

We were hired soon after they had launched the service. Before us, all their “Operations used to be carried out by one-man-army with an excel spreadsheet and lots of manual work, like validating orders, matching resources, confirming bookings, calling, emailing, arranging stuff. As rspective, our main goal is to automate the “Operations”, try new things and build the entire service-exchange platform in the long run.

We are thinking to launch today, surprise! :)

We can hardly call it a project, feels more like a process, to tell you the truth. There is no such thing as planning for more than a few days ahead, usually a few hours ahead being uncertain. It’s a unique soft-dev process, called Sparta ;). There were times like “Guys, please stop everything now, we’ll be on air on RTL anytime tomorrow, we gotta do load-testing, full-power on blitz.io and loader.io!“. Other times we would launch the service in another city – “we are thinking to launch today, surprise! :)“, set up online payments in a few hours, migrate the entire platform to Salesforce in one weekend.

Things will break. If nothing breaks, it means we’re doing it wrong or the speed is too low

Thanks to this, things are happening literally in no time, the pace of development is unbelievable. Features which in “normal project environment would take man-days / -weeks (research, planning, estimating, splitting user stories, implementing, QA testing, CI testing, documentation, …), here are deployed live to production after a few hours. There is no overhead for communication on planning far ahead, super clean code, high code coverage, continuous integration, moving issues from one column to another, whatsoever. We get things done, go straight to the point, avoiding all that unnecessary hassle.

We were used to building whole full-blown solid IT solutions with tons of custom software, with “help of databases, frameworks, elaborate design patterns, shared modules, preparing VMs, taking care of next-to-impossible corner-cases. Now, we’ve learnt a lot on being time- & resource-efficient – in many situations one can achieve the same thing by putting a handful of SaaS providers together, wiring it up with a few lines of code (the less custom code the better, actually – higher reliability and lower maintenance costs). Delivering a naive solution that will do for today and tomorrow and not any longer is just fine – who knows, maybe tomorrow we’ll throw it all away as it won’t be needed anymore?

Once, I counted the SaaS providers that we’ve been integrating with one way or another (incl. proofs of concepts) – total of 20 SaaSes within 20 days working with that company. As of now, ~35 wouldn’t be far from reality. Just to name a few: salesforce, zoho, postmark, mailchimp, mandrill, contentful, parse.com, nexmo, paypal, zendesk, google drive, dropbox, apostle.io, newrelic, mixpanel, ironmq, logentries, heroku, cloudcontrol, openshift, blazemeter, blitz.io, loader.io, dataloader.io…

The possibilities to learn new things are endless. It’s not only about tech, it’s also the lean approach – measuring impact, advancing with small steps to deliver minimum viable product. And the Think Big mindset – where nothing is impossible. On top of that, the joy of delivering a real business value and seeing it run (or fail) on production the same day is priceless – instant feedback.

When it comes to time pressure and stress level – it’s really high – scheduling “asap like really asap”, very quick decisions, often resulting in frequent context-switching. The responsibility is overwhelming – very thin safety net, new features often deployed straight to production, operating on live data. You break the critical flow – whole company operations is blocked. Simple as that.

It’s definitely not a place for everybody. Sometimes it’s really hard, disappointing or even harsh. Sometimes it’s working double shifts. Other times it’s just fun. But it’s a truly amazing adventure and a huge load of experience for both parties.