How to determine if a DevOps transformation is on track
Article by Micro Focus A/NZ director of southern region Amish Prajapati.
To compete effectively and thrive in the current business environment, organisations must innovate quickly and build software that generates high business value. When implemented successfully, DevOps practices offer many benefits to organisations, letting teams pinpoint potential issues earlier and innovate faster with less risk.
The 2021 World Quality Report found that DevOps adoption is rising steadily, with DevOps-enabled teams regarded as the orchestrators of quality within the enterprise. By using DevOps practices, organisations can deliver high-performance applications in a shorter timeframe than their competitors.
Getting DevOps right is easier said than done. While many organisations have decided to undertake a DevOps transformation, scaling this in hybrid IT environments without compromising quality, security and availability can present significant challenges.
Here are five key ways organisations can monitor whether their DevOps transformation is on track.
1. Manage the ‘J curve’ effect
From the outset, it is important to remember that deploying a DevOps strategy requires completely changing current approaches to building and delivering software. Key stakeholders should recognise that DevOps does not just involve transforming the technology but also the organisation’s culture.
A common theme for early adopters across their DevOps journey is the ‘J curve’ effect. This describes a pattern of initial quick wins, followed by a slower and, at times, frustrating stage where teams are tasked with solving more complex problems. In many cases, this can negatively impact team morale, particularly as new obstacles arise in a constantly moving environment.
Effectively managing the J curve effect depends on a ‘Kaizen’ approach. Kaizen is a Japanese term that refers to the practice of continuous improvement and a DevOps principle. This involves having a clear picture of the organisation's current state, establishing the desired future state, and making improvements until the gap has been bridged. To ensure teams have a clear view of each stage of the transformation, these improvements need to be incremental.
2. Understand the current state of the technology value stream
When a culture of continuous improvement has been established, businesses need to determine exactly what needs to be enhanced. This requires a thorough understanding of the current state of the technology value stream — that is, transitioning a business idea to a technology solution.
A value stream can be split into two sections: product design and development/product delivery. Many organisations start with the DevOps practice of continuous delivery. However, by focusing on the product delivery stage, these organisations fail to factor in potential hold-ups in the product design and development process. For this reason, making improvements earlier in the value stream can accelerate time to market and cut costs.
As such, it is crucial that the full value stream, not just product delivery, is managed and optimised from the start.
3. Align outputs with outcomes
When it comes to measuring the performance of software teams, many organisations make the mistake of focusing solely on outputs rather than considering outcomes. This approach fails to capture the relationship between the product and the effect it has had on the end-user.
Building a product and bringing it to market doesn’t guarantee economic value for the business. Likewise, creating the best product does not necessarily equate to business success, particularly if the market is already flooded with cheaper, ‘good enough’ alternatives.
Therefore, gauging the success of a DevOps strategy should involve looking at both productivity and business outcomes. This could mean looking at the outcome of reducing a cycle time, such as faster feedback, improved customer satisfaction or fewer outages, or exploring the link between increased deployment streams and new online revenue streams.
By aligning team outputs and the value delivered to end-users, organisations will be better positioned to determine if their DevOps implementation is on track and make any necessary adjustments.
4. Use key performance indicators (KPIs) as a learning tool
To create a learning organisation successfully, KPIs should be regarded as a goal. This means moving away from the mindset that performance can be assessed by measuring it against set numerical targets and simplified metrics. Not only does this method stifle creativity, but it also removes a team’s sense of self-determination.
Tracking performance at the end of a delivery cycle/release or annually is of limited value given the fast-moving nature of modern delivery. Adding a culture of ongoing review and feedback increases the value of all KPIs.
Management teams should see metrics as a learning tool, helping them find barriers to progress and implementing processes to overcome them. This way, they can better motivate DevOps teams and effectively foster a culture of innovation. With ongoing review, these barriers can be identified more readily and earlier in the delivery cycle.
5. Smart approaches
Like any transformational journey, implementing DevOps will not be straightforward. While there are many positive outcomes, there will also be multiple challenges. Putting in place smart approaches that encapsulate both technology and culture is fundamental for managing the highs and lows.
To understand the challenges inhibiting value-based delivery, organisations must disassemble silos of value and see the end-to-end process as a single supply chain. Ideas and concepts feed the start; business value is delivered at the end.
Enterprises can make the most of their DevOps strategy by focusing on the entire technology value stream, applying continuous improvement practices and continuously reassessing measurement, including the role of KPIs.
This will ultimately let them develop high-quality software that satisfies the needs of the business, and at a level of quality that exceeds the demands of the end-user.