Digital experiences are deeply ingrained into our everyday lives. When we shop online, bank with mobile apps, or collaborate virtually, the digital interaction is becoming as second nature as breathing. Expectations for intuitive, engaging experiences are greater than ever, and product development teams are feeling the pressure to deliver.

Now more than ever, there’s a need to innovate faster while maintaining data-driven certainty and value with every release. That’s why leading development teams are striving to increase their release cadence, and it doesn’t end there. They’re now being relied upon to bring optimized experiences to customers and stronger measurement capabilities to key stakeholders involved.

It’s time to move forward fearlessly. The best practices of yesterday are no longer relevant. Data needs to play a crucial role at all levels of the decision-making process. 

At Split, we partner with some of the most forward-leaning product development teams around the world. As a result, we have a unique vantage point into what’s working and what’s not across agile transformation. Interestingly, the teams benefiting from Split are now laying the foundation for a new set of best practices. They’re also quickly emerging as tomorrow’s digital winners. 

Leaders understand that they’re not shipping a version of an application that’s the same for all users at every moment. Instead, they’re providing thousands of versions of an application. Each defined by its own collection of features, woven together to create unique, controlled experiences.

Because features are the new tool of measurement, it requires us to change the way we develop. We must reinvent the rollout process and make smarter, data-driven decisions. As we move forward with a new set of best practices, here are a few truths that every product leader should follow. 

Truth 1: Speed Creates Resiliency 

As development teams embrace feature management, they’re able to move faster and with greater flexibility. Big application releases are instantly streamlined into controlled feature releases. Development and testing can occur simultaneously. Plus, every feature is paired with a defined set of users and is measured for optimal impact. This is what progressive delivery looks like.

At Ancestry, the world’s leading genealogy site, feature management is helping development teams increase their release cadence. It’s also empowering resilience across their organization. Read about their new culture of change here. 

Truth 2: Feature Management Empowers Continuous Delivery

Feature management might even be the number one enabler of continuous delivery. Armed with the ability to decouple a release from deployment, code can now be delivered as continuously as possible. Thanks to feature flags, developers can merge finished features in production environments. They’re empowered to improve velocity, push fewer lines per commit, and achieve continuous delivery throughout the process.

Experian, a credit reporting agency that aggregates information for over 1 billion people, is now leveraging feature management. Since partnering with Split, it has seen its release cadence skyrocket from 2 to 100 releases per month. That’s a considerable increase.

Truth 3: Rollouts Are Now Agile Experiments, Too

Understanding the impact of every feature unlocks new possibilities. Each one should be carefully monitored with precision. Otherwise, it’s a missed opportunity. 

David Zarruk, the head of data science at Rappi Bank, believes in the value of data-driven decisions. He stresses the importance of feature-level impact data at the São Paulo-based digital bank. By setting up an experiment for every new feature delivered, success metrics aren’t just for comparison purposes. They’re for rolling out intuitive design choices and creating flawless experiences for online banking customers. 

Hear from David Zarruk himself during his session at Split’s Flagship 2022 event.

Truth 4: Feature Context Isn’t Just Important, It’s Imperative

As features become the new units of measure, the data behind them becomes infinitely valuable. Feature context can help tell us who saw what feature and at what time. Did key customers see your new feature? If so, what was the result of their experience? These questions can now be answered, giving us an unprecedented level of understanding. 

Say goodbye to muddied waters, hello to clean data. Feature context can provide a clearer picture to help navigate service delivery to performance management, to product analytics, and more. 

It’s no longer enough for software developers and product teams to track performance at an application level. They need to view things through the lens of feature performance. When troubleshooting issues, the availability of all relevant contextual information can decrease the time between detection and remediation. 

Bringing feature-level context from Split into application performance management systems like Datadog or New Relic, allows teams to save time.  Correlating a code change or feature roll-out with a shift in performance metrics accelerates investigation into the true cause of an issue. Before feature-level context, issues could be identified, but there was no easy way to uncover the feature that was at the root of the problem. Now we can triage with precision. 

Likewise, customer support teams must understand feature context as well.  They must understand the feature context when a ticket is created. This way they can more effectively assist a customer if they know which feature they have turned on. It’s imperative these teams are provided feature context in the tools they operate daily.  

Truth 5: Data-Driven Cultures Take Commitment (and Humility)

The truth is, we don’t always know what the data is going to tell us. Therefore, we need to be open about what gets uncovered. We also need to be comfortable with being wrong. Failure is not the end of the world; it’s a new realm of opportunity. 

The good news is that feature data platforms with experimentation capabilities like Split can help us fail fast. This means we can adjust failed features quickly without the need for severe rollbacks or major loss of customer engagement. 

Sure there’s a steep learning curve, but it doesn’t have to be an uphill battle as more companies embrace experimentation and data-driven decision-making. Especially when there’s a culture of openness and humility.

Ronny Kohavi, former Vice President and Technical Fellow at AirBnB, discusses this in detail at Split’s Flagship 2022 keynote. Be sure to watch it here. 

New Best Practices Begin With Experimentation

To achieve success in a new era of best practices, embrace a culture of experimentation. Armed with unprecedented data around end-user experiences, development teams will excel as the ultimate knowledge source for senior leadership. The decision-making authority will shift away from leaders to the people on the front lines, directly managing and releasing features. This is happening among the digital winners in software development and data analytics as we speak. 

Don’t fall behind, stay ahead with a feature data platform that helps you meet the new best practices of today. See how Split can help you empower a culture of experimentation and data-driven decision-making. Schedule a demo!