Skip to content
Fast Company Executive Board

The Fast Company Executive Board is a private, fee-based network of influential leaders, experts, executives, and entrepreneurs who share their insights with our audience.

“Shift left” to start your journey to a cloud development model

The cloud operating model with infrastructure as code can help to dramatically improve productivity for all involved and pave the way for superior product development.

[Images: Adobe Stock/puhhha]

Profile photo of Ajay Singh

BY Ajay Singh

History buffs looking for a tale of speed and smarts winning out over imposing monolithic infrastructure need look no further than the Spanish Armada, which had its invincibility pierced by a more nimble British fleet, altering the course of what had felt like destiny.

For modern technology product leaders, the need for speed may not feel quite so dramatic. But it is a good reminder that moving a team building a classic product on-prem to a more agile, cloud-based development process carries significant benefits.

With a cloud development model, for example, you don’t overbuild features, enabling you to hit the mark more confidently. The market also perceives you to have significantly higher velocity with the same resource set. At the same time, your product quality can also improve as you increase the release velocity. Beyond that, you can adjust much faster to competitive changes in the market and customer needs, and the quality of the application you are delivering can increase.

Does it sound counterintuitive that moving faster can deliver all of that? It’s not. There are several components in the journey to the cloud model, but it all starts with engineering.

EMPOWERING YOUR ENGINEERING TEAM

Speeding up the engineering process using “shift left” techniques pushes tasks traditionally done later to earlier stages of the development process and brings development and testing into alignment. This can help identify problems before they get too far along in the development cycle.

As a leader, a strong first step to consider in the move to shift left is the creation of a separate unit for quality assurance in engineering services—one that reports to the highest level possible within the product team leadership. That sends a very strong message that engineering services and QA testing responsibilities are very important. It also sets up a team that is constantly thinking about how to improve engineering productivity. 

Then, establish a goal to dramatically cut the release cycle from a classic waterfall build of nine months to a year to a quarterly or monthly shippable build. With that, you should then have a team in place that wakes up every morning thinking about how to speed up and shift left, and engineering leadership that is on board with that objective.

DRIVING A CULTURE SHIFT

A key aspect of the shift-left principle is that developers own their quality assurance starting early on in the process. It is a mindset shift that requires a culture change. Developers who are not accustomed to doing their own testing will need some education around the idea that the modern developer is a “quality-accountable” developer. You not only develop, but you should also write your own test plan that demonstrates you have tested against the new capability you’ve added. This can ultimately help improve quality while at the same time increasing release velocity!

In this scenario, the job of engineering services is focused on tooling and the development pipeline so developers can do their self-testing. They’re doing more integration and customer-use scale-testing employing a ton of automation. Cloud-based development delivers the benefits of automation and makes the development infrastructure disappear—effectively rendering it invisible as nothing more than a line of code for a developer to run with.

This culture change can make engineering services engineers even more valuable. If they can get a 10% productivity improvement in a 300-person engineering shop, that’s like getting 30 more engineers.

THE OTHER LEGS OF THE STOOL

Customer input: If the first component is speeding up the engineering process using shift left techniques, the second is connecting that sped-up process to a near real-time customer input from practitioners that use your product. Your product needs to be intuitive, effortless, and user-friendly. Bringing customers into the fold during the design process can help your team hit their markers. Think of your customers as partners who can provide real-time feedback throughout development. If you hone in on feedback as you build your product, you can slice away the functions you don’t need and readjust quickly. Also, you do not need that big beta that is typically at the end of a waterfall process.

Modularity: After that, it’s making progress toward making the code base more modular, with clean API definitions. A modular development approach separates a product into smaller parts that can be developed individually. It’s a great way to decrease risk and improve performance by allowing developers to spot issues and isolate the problem without having to break the whole product apart. It is also a great way to onboard new members to the development team as they can start adding value by focusing on one module rather than learning the entire code base.

Upgrade: Finally, the last component is making significant progress on the cloud development process on self-service upgrade. As you speed up the development cycle, it is important that the customer can upgrade to the latest release with the push of a button. Manual requests for infrastructure and resources can stifle developers.

THE RACE FOR AGILITY

In any agility contest between super tankers and speedboats, the cloud operating model with infrastructure as code is the winning model, dramatically improving productivity for all involved and paving the way for superior product development.


Ajay Singh is Chief Product Officer at Pure Storage.