Only got a minute? Here’s the 5 things we learnt:
- Continuous delivery can cut down costs and deployment time while also helping you deliver true customer value
- However, in order to implement continuous delivery you need to shift your organisation’s mindset towards production environments and open everyone up to live experimentation
- A key component of continuous delivery is decoupling releases from deployment and turning production deployments into a non-event
- Make sure to set your continuous delivery north star and to measure your efforts to see your journey
- The tech for continuous deployment should be focused on automating the path to production (“no humans allowed”)
Dive in for more details and learnings:
Continuous delivery has been a hot topic in our industry for a long time, yet despite this, many organisations still aren’t doing it. As enabling as it is, it’s hard to implement as it goes beyond your tech department and impacts many parts of an organisation. In order to work, continuous delivery has to pervade every part of the process, from users to designers.
As such, on Thursday 16th May, we came together (along with a special guest from Nando’s) to discuss continuous delivery; the key to launching faster, cheaper, and more effectively.
Kicking off the event, Reg Meyer, Head of Technology Solutions at Nando’s, took us through what Nando’s does (it’s more than just chicken), how they have begun to move towards continuous delivery and how they measure this movement. Overall, he noted that by shifting to continuous delivery, as opposed to manual deployment, an organisation can drastically cut down on delivery times - as well as continuously bring quality to customers by shortening the loop on incorporating customers’ feedback. However, in order to implement such benefits, you need will need time to transition to such a methodology. He noted that throughout this journey measurement frameworks (centred around maturity of practices) are key to keeping you on track and that having the right attitude is vital to this change.
Sean Gabriel, Delivery Director at Red Badger, echoed this sentiment, showing how mindsets need to change to make continuous delivery a reality. He noted that a big part of this mindset shift is ending the ‘big deployment’ moment. Instead, he says deployment should be both an inbuilt part of testing (i.e. in production), and a means to ‘shift testing left’ and push quality upstream. However, doing this requires an organisation to be open to ongoing live testing, controlled experimentation, and the idea that you don’t know everything (nor can you) until code meets the customer in production. With this new mindset, a project can move forward more nimbly, creating the right changes that benefit users even more.
Building on the discussion so far and talking us through the enabling technologies for continuous delivery, was Red Badger Chief Scientist, Stu Harris. Based on his years of experience with delivery and deployment, Stu’s tech recommendations, such as Kubernetes containers, and the use of monorepos and microservices, demand a softness that allows you to create flexible programs that can adapt to your needs. He also noted that with the right tech, you should be able to fully automate the process with a human on the sidelines providing clear declarative intentions for the program, as well as being on hand to jump in on the issues that invariably arise when testing.
Overall, continuous delivery is a journey that requires a cultural shake up and is a process that never really ends. To implement this methodology, it’s best to measure where you are, reassess your release strategy and decouple these from deployment, and make sure you are supporting all teams (be they tech-focused or otherwise) with the right tools.
Keen to view the full talk? You can watch the recording here.
Are you looking to get the continuous delivery journey started or learn more about it? Get in touch at firstname.lastname@example.org