How do you engineer for a startup unicorn?

Want to turn your startup into a unicorn? The next Apple? Google? or Facebook? Netflix? or the next Amazon? What all these high-performing organisations have in common is that they all leverage a set of successful patterns and practices learned over decades to continuously transform and improve their software development operations which directly helps in scaling their business. A startup is all about growth, and you won’t be able to sustain this growth if your software development operations is unable to keep pace with your fast growing business needs.

The good part is we do have all these best practices and learning from across these high-performing organisations distilled down for you. DevOps is a collective term we use for these patterns and best practices. Now you too can leverage DevOps, to shorten your software development lifecycle, get more of the agility and become better at the quality of your product delivery. All of these finally resulting in a software operation that is fully aligned and that can scale with your business.

DevOps is “Lean” applied to software delivery

A lot of what we do in DevOps is derived from Lean Manufacturing. DevOps brings the promise and possibility of realising Toyota Motors level efficiency and reliability to the software production pipeline at your startup. Here are the three basic principles of DevOps, otherwise known as the “Three Ways” outline the values and philosophies that guide DevOps processes and practices:

1. Left to right flow – The value stream

This is a set of principles and practices that accelerate the delivery of IT services. For example, Continuous Delivery and the extended principles and practices that lead to an accelerated flow. Amazon, Google, Facebook, Etsy, and Netflix routinely deploy hundreds of times in a day to production. A lot of these companies have their employees deploy to production on the day of joining. Facebook brags about their culture where new employees ship code that make it to production before they have finished their paperwork – the employee onboarding paperwork. Compare and contrast this to organisations that struggle deploying maybe more than twice a year and have kind of a waterfall deployment model. And, most of those, we know, tend to be low-performance organisations. Nicole Forsgren, a PhD in Statistics and Psychometrics, reported in her survey report (2015) that high performance organisations had the following characteristics:

  1. They were deploying code 30x more frequently
  2. Have 200x shorter lead time (idea to market)
  3. 60x less failures
  4. Mean time to recover (MTTR) is 166x faster

2. Right to left flow – The Feedback loop

This is a set of principles and practices that amplify the feedback loop. That shortens the feedback loop. Creating a problem-solving culture, as well as understand monitoring, as it applies to DevOps. Includes monitoring business metrics and change management as it applies to DevOps.

3. Continuous learning – Kaizen

The unique trait of high-performing companies is that they are good at learning fast. Organisational learning and safety culture. Practice items like blameless postmortems, resilience engineering, and systems thinking, as they apply to DevOps.

Unicorns and Horses

So how does the improvement really look like at these unicorn startups? Here is run down of some publicly shared stats (2015):

  1. Google has stated that they do somewhere about 5,000 code commits a day, 75 million test cases per day
  2. Amazon, 136 thousand code deploys per day, 15 million per year
  3. Netflix, 500 code deploys per day
  4. Etsy, hundreds plus a day

It’s not just the unicorns though, that have these kind of examples of improvement. Companies like Ticketmaster, Nordstrom (has been around for a hundred years), Target, USAA, ING are also moving fast and achieving better resilience by leveraging these best practices.

  1. Ticketmaster witnessed 98% reduction in MTTR and this was based on them moving fast.
  2. Nordstrom achieved 20% shorter lead time across the board.
  3. Target improved on their full stack deployment from three months down to minutes.
  4. USAA improved their releases from every 28 days down to 7 days.
  5. CSG went from 200 incidents per release down to 18.

And all these stories are examples of moving fast with high reliability. ING has got 500 application teams doing DevOps. At Target, they’ve created this thing called the DevOps Dojo. A 15,000 square feet floor space at their headquarters in Minnesota and they actually bring groups in from all over the organisation, to immerse them into this kind of DevOps mindset.

No matter what industry you are in, or what product or service your organisation provides, this “scientific way” of thinking is paramount and necessary for survival for every startup if it aims high for itself.

At Opscale, we take a DevOps first approach to everything we do. We help startups pick the right tools and technology that is needed for further success and execution of their business so they are able to move fast and deploy safe.

%d bloggers like this: