No matter what industry you work in, the evolving demands of an “agile” promise is a challenge to live up to. For developers, that requires constant pivots, faster response times, and a graceful recovery from failure. Still, against the enormous pressure to continually improve, the best software teams deliver the goods safely and on time. Touché!
But what’s the barometer for true success? How can we advance among the digital leaders and gauge our progress along the way?
As it turns out, there’s an important set of metrics all software teams should use as their benchmark. They’re the DORA metrics.
Not familiar? Need a quick refresher? Here’s a brief recap:
DORA stands for The DevOps Research and Assessment Team. The group was established at Google. After researching the opportunities of the DevOps practice, they put together a bulletproof way to track software development teams’ performance. Identified after surveying over 31,000 software people, they designed four very helpful DORA Metrics. If you don’t know them by heart already, here they are:
The Four Metrics
- Lead Time for Changes (MLT): The time it takes between a change commit and its deployment
- Deployment Frequency (DF): How often your team deploys software to production
- Mean Time to Recover (MTTR): The time between a system failure and its full recovery
- Change Failure Rate (CFR): How often your teams’ deployments lead to failures in production that require remediation
Simply stated, the DORA Metrics show you how often, how quickly, how consistently, and how reliably your team delivers software. This isn’t to compare for bragging rights. There isn’t a trophy to celebrate a winner. DORA metrics are just a reminder that there’s always room for improvement (no matter where your software practices sit).
HINT: If you’re not already using feature flags, do it! They’re an easy tool to help you quickly boost your DORA metrics.
The Math Adds Up With Feature Flags
Feature flags (also known as feature toggles) are tools software teams use to safely test features in production. You can turn flags ON and OFF to experiment with functionality without pushing new code or disrupting experiences. In the context of DORA metrics, think of them as cheat codes that take you to the next level of software delivery.
Here’s how feature flags can help you improve your DORA metrics:
Go Straight to Production & Improve Your Lead Time
With feature flags, you can cut back on the number of staging environments, as well as, unnecessary rounds of peer reviews. Instead, send your feature changes straight to production—right to the trunk wrapped up in feature flags! Not only does this significantly reduce time (and costs), it’s safer for all parties involved. For one, they’ll vastly reduce the chance of a merge conflict, and that will surely boost your MLT.
Switch Off Flags, Switch On Your Deployment Frequency
The OFF function of a feature flag is just as crucial as the ON, because it lets you deploy features and validate them before releasing them to customers. This means you can introduce features little by little with frequent dark deployments hidden behind flags, toggling ON only once they’ve passed the necessary tests. As you commit smaller changes like this on a regular basis (even during peak traffic times), you’ll soon be embracing CI/CD and improving your DF metric. Not bad!
When It Comes to MTTR, You’ll “Kill It”
If what you turned on suddenly breaks something, feature flags are a fast response solution. It’s why they’re sometimes referred to as “kill switches,” allowing you to roll back a problem causing feature with the flip of a toggle. Plus, if you’re using premium feature flag platforms, you’re granted visibility of every flag across departments. That means your entire ops team is alerted to the location and cause of the issue for a faster resolution. In fact, MTTR for teams using Split’s feature flagging platform is less than one-minute.
Fail Before Your Customers See It
Because feature flags let you test directly in production, any failure can be identified before all of your end users are exposed. Maybe it’s a latency problem that’s discovered in the deploy stage. Maybe it’s a throughput issue that only reaches a subset of users in a gradual rollout. Either scenario, the disruption is minor in comparison to a full traffic scenario. Ultimately, it’s a less painful failure: falling like a feather versus falling like a log. Improve your CFR metric with feature flags.
Feature Flags Can Improve Your Metrics
If performance matters to your team, then you’ve got to have the DORA metrics to back up your software delivery performance. Feature flags are a sure way to help you accelerate your delivery and boost your numbers quickly. Stay tuned for further strategies on the subject, a deep dive into each metric, plus insight into how feature flags will change how you think about the DORA metrics.