This is the first of a three-part blog series about:
- The U.S. Federal Government’s practice of implementing hybrid-Agile/”Agilish” software development projects
- The perils of Agilish projects
- How we can make Agilish projects more Agile
You are likely reading this because you are looking for a way to make Agile work within the byzantine world of government/public sector Information Technology (IT). My perspective comes from more than a decade of working with the U.S. Federal Government implementing software systems using Agile. I have spent a good portion of that time working in “hybrid-Agile”, “Scrumfall”, “Agilefall”, or “Agilish” projects. My advice: Do your best to avoid Agilish projects.
Reasons to Avoid Agilish Projects
Look, I get it. You are likely getting conflicting direction from your government agency or client. They want Agile’s speed and flexibility with traditional project management or “Waterfall” project controls. They want end users to have greater say in shaping solution features during development, but award firm-fixed-price contracts and require adherence to inflexible program baselines. Even when agencies use “Agile-friendly” contracting rules, contract language and oversight requirements often impose Waterfall processes and reporting requirements.
The problem is that you can’t have your cake and it too. There is no free lunch. <Insert another hackneyed phrase here> In trying to be both Agile and Waterfall, most Agilish government projects toss agility aside to accommodate established project management processes. I now discuss concrete reasons why Agilish projects are a bad idea.
Agilish isn’t a Real Thing
Hybrid Agile or Agilish is not a recognized software development approach. It is an project management approach to Agile without roots in proven Agile or software engineering practices. There is no standard framework (like Scrum) or body of knowledge (like SAFe or PMBOK) for Agilish implementations. Without standards or best practices, you’ll be making up your own Agilish implementation as you go. The same goes for contract firms that claim expertise in hybrid-Agile.
Predictive vs. Adaptive Planning
In traditional project management, the bulk of project planning happens before and immediately after the start of a project. This is a predictive planning approach. Predictive planning assumes the next project will be similar to previous projects. Thus, it is possible to estimate (predict) the amount of work (scope), necessary resources, timeframes (schedule), and cost (budget) of the next project based on historical data and experience. This assumption is fine for routine and relatively simple projects but breaks down for more complex, technically challenging, or innovative projects.
Agile planning is adaptive planning. In Agile, planning and management of risks and opportunities occur throughout the entire effort. Decisions are based on lessons learned and insights gained during development. This allows Agile teams to, “Welcome changing requirements, even late in development.”
The Worst of Both Worlds
Mixing Waterfall with Agile yields the worst of both worlds. You introduce predictive planning and project controls (e.g., Earned Value Management (EVM)) that sap the Agility out of Agile while lacking detailed program baseline data necessary to measure project progress and apply project controls a la Waterfall.
Agilish from the Start
In government institutions, the shift away from traditional project management to Agile is rarely, if ever, a bottom up initiative. Most likely, your agency or government client decreed Agile as its preferred or official software “project management” approach. Such pronouncements usually mark the beginning of Agilish projects.
Agile, and its most popular implementation, Scrum, are not “project management” approaches at all. Agile is an umbrella term covering multiple approaches (e.g., Scrum, XP, Kanban, Crystal) to building and delivering complex technical systems. These approaches share the tenets of the Agile Manifesto and its foundational 12 Principles of Agile, which I paraphrase and distill below:
- Teams develop and deliver system capabilities iteratively and incrementally within short timeframes
- Empowered teams manage their own work and work at a sustainable pace
- The measure of progress is delivery of valuable capabilities rather than alignment to plans
- Direct and ongoing collaboration between and among technical, organizational, and customer stakeholders largely replaces formal documentation
Agile is a vision for a way of working, not a way of managing work or workers. Scrum is a lightweight framework of roles, rules, and best practices that embody the tenets of the Agile Manifesto and the principles of Agile. Characterizing Agile or Scrum, or any other flavor of Agile, as a “project management methodology” virtually guarantees Agile adoption failure.
Software Development Has Changed
Most enterprise software solutions are complex systems that change and evolve for as long as they are operational. Agile and DevOps enable this ongoing evolution by:
- Decreasing the time it takes to develop, test, and deploy new software capabilities
- Increasing the frequency of new software capability deployments
- Automating Operations and Maintenance (O&M) activities
Agile and DevOps allow organizations to move away from deploying large software releases periodically or infrequently towards continuous flows of software feature releases. A strong DevOps culture fosters:
- Close collaboration between developers and systems operations personnel
- Automation of system deployment, monitoring, alerting, and troubleshooting activities
Agile does away with discrete Waterfall phases for planning, analysis, design, testing, and deployment followed by a long O&M phase. All of these activities happen in Agile, but occur continually and in rapid cycles, thanks to DevOps. Information and insights gained during these cycles inform how solutions evolve and how teams improve technical quality and team processes.
Agile and DevOps open the door to a different way of working. DevOps enables software development teams to deliver increments of valuable functionality sustainably and when needed rather than on a set schedule. As the speed and frequency of releases increase, scheduled project start and end dates lose their significance.
We move into what some call a “Software Development as a Service (SDaaS)” model. Instead of standing up new projects, agencies augment available software development teams or form new ones. The additional software development capacity adds to the existing software development flows. Agencies have the choice of adding technical contributors as agency hires or contractors. By treating software development as a service capability, organizations are able to ramp software development capacity up or down depending on mission needs and budget constraints.
Conclusion
Acquisition of software systems by the U.S. Federal Government is transitioning from a traditional project management approach to Agile. Over the last five years, the Federal government has enacted significant statutory and policy changes to align acquisitions and contracting processes with Agile and DevOps. Unfortunately, this transition will take time due to the complexity of Federal acquisitions and contracting and the resulting decades-old predictive planning culture.
In the part two of this series, I’ll talk about the reasons why Federal software systems projects become Agilish and the resulting problems. Part three will cover ways to make Agilish projects more Agile.