In a recent independent report, sponsored by Copado and executed and analysed by Beagle Research, Salesforce researched customers using DevOps to accelerate and improve the speed and quality of their implementations. The report was based on thousands of data points collected from over 200 global customers. Andrew Leigh, SVP of marketing, Copado spoke to DevOps Online about the findings of the report and what it means for DevOps.
What is the biggest discovery from the report?
The State of Salesforce DevOps report had several key insights and findings but if we had to pick one to highlight first, it would be that the highest performing companies are building a DevOps strategy around Salesforce to maximize their returns. More than 17% report an ROI of over $5 million.
Is there anything you found that you were surprised by?
Yes, we were surprised by the level of velocity and stability the teams using [certain cloud firms] experience compared to the broader IT industry. Teams with relatively immature DevOps processes compared to the broader IT industry performed better on average through the power of the platform.
What were the biggest challenges survey respondents faced while trying to implement DevOps?
As we looked at the overall demographics of our respondents, we noticed some surprising results and challenges that came with it. The one metric that stood out was the correlation between the size of the delivery team and the time it takes to get new functionality into the market. As teams reached over 50 people or more, the velocity of innovation dramatically slowed from less than 10 days to several weeks. We measured this in terms of Lead Development Time. We also saw a major dip in Release frequency and quality rates. The lesson is that large teams require disciplined orchestration to compete in today’s fast-paced digital economy.
How do you overcome these issues?
The first step is to understand that DevOps is not a technology. DevOps is a revolution that is sweeping through IT departments to allow them to manage the new scale and speed of modern development and delivery teams. Large teams need to look at the ways they are organised and challenge traditional thinking. Traditional silos need to be broken down and disparate teams need to be realigned according to transparent innovation pipeline optimizes the entire organization. All processes must be metrics-based to establish baselines and measure both improvements and weaknesses. Take a “no-blame” approach to the transformation to ensure teams are bought in and momentum is not lost. Finally, tooling is important. Use what meets both your current and future needs. Be careful not to over-engineer or become dependent on point solutions for every process.
When carrying out the report, why did you decide to focus on release frequency, development lead time, change failures and restore time?
We aligned with industry standards to measure the innovation performance metrics. Nearly ten years of DevOps data science has been done to identify the right metrics to be measured. DORA leads the space with their ground-breaking work that is documented in detail in their annual reports and the book Accelerate.
What advice do you give to firms struggling to involve DevOps practices?
Overall, business, no matter their comfort level or how much they’re struggling, need to have a solid footing in the following as they think about their future DevOps strategy. First, it’s best to be mindful of the size and makeup of your delivery team, strive for daily releases especially as your team gets bigger, implement continuous integration across all developer, security and admin teams, and design and automate your continuous delivery process. Don’t forget to build a DevSecOps centre of excellence into your strategy.
As part of the results, you found that the size of team matters. Why do you think this was the case?
Yes, that was such a critical finding to the research, which found that the size of the delivery team is the number one factor affecting the throughput of innovation. The organisation and discipline of these heterogeneous teams is the key to maximizing performance of Salesforce projects. In addition, it’s important to note that 80% of larger teams used Salesforce-specific commercial tools to aid in their development lifecycle. We saw smaller teams rely on generic DevOps tools or tools built in-house but none of the larger teams that elite performers (the businesses that are doing well, seeing strong ROI), build in-house.
Why do you think it is important to integrate changes daily?
Integrating developers’ code changes daily is a hallmark of continuous integration. Integrating smaller changes more frequently reduces the complexity of merges, and allows teams to refactor and keep their codebase more maintainable. Elite performers were 173% more likely to integrate changes between developers at least daily.
What trends do you think are having the biggest impact on DevOps right now?
As we enter 2020, we anticipate a few trends in DevOps. As DevOps has proven to be such a critical part of business, we anticipate more companies will restructure their traditional processes and departments – like finance or marketing – to accommodate increasing flexibility and agility from their IT departments. We also expect to witness Value Stream Management become front and centre as enterprises implement tools to deliver metrics and analytics to help them manage its complexity.