W03 - Digital Workflow

Lately I've been planning for 2021. Jerry asked me a question: how do we measure everyone's ability to understand the business?

I sketched a few angles. From outcomes: it could be how well the architecture supports the business, the reasonableness of the technical roadmap, or how many business problems were resolved from an engineering perspective. From process: how much time was invested in learning the business domain, whether participation in requirement or planning discussions was sufficient, or feedback from colleagues in other functions. Clearly, these dimensions are not SMART. Translating such an abstract capability into concrete, measurable metrics is indeed difficult.

That led me to an amusing observation. We who work on digital transformation are not very digital ourselves. Without sufficient digitization, we can't effectively abstract and standardize, and therefore can't scale. We can't extract and disseminate the work experience of outstanding employees, nor easily build capability models like business understanding.

From an organizational management perspective, today we can only see data about how people allocate effort to requirements. Non-requirement effort allocation is largely not digitized.

Take the requirements dimension: each complete requirement goes through a long chain of actions — requirement review, technical design, coding, integration testing, code review, self-testing, environment deployment, staged release, and monitoring production. We have data for each part, but the data is not granular enough: outcome data is abundant, while process and behavioral data are scarce. And data between nodes is mostly not connected. For example, in Ones, the status of each requirement and task relies on manual updates, which are often missed or delayed, so the data only gives a rough picture. Ideally, as users we shouldn't even notice a requirements management tool like Ones in most scenarios.

This made me think of Cloud IDEs. I used to struggle to see the value of cloudifying the IDE, but from an organizational perspective it becomes significant. I believe its greatest value is a digitized workflow. RD work relies heavily on IDEs, yet such a core tool is fragmented and offline, generating isolated behaviors and data. If the IDE is cloud-based, it can act as a platform integrated with many upstream and downstream scenarios: branches directly linked to requirements, one-click test requests, one-click deployments, and so on. Even basic data would be sufficient to update Ones automatically. It can also address efficiency and security issues, such as providing a unified development environment and protecting source code off developers' local disks.

Imagine a future where designs are converted into editable code through intelligent code generation with one click, the code is imported into a Cloud IDE for further development, and one-click build and deployment produce cross-platform applications. All data and assets would be consolidated on the platform, and each node could be optimized precisely using abundant data.