Software development teams around the world face a recurring challenge: accurate project estimation under pressure to deliver fast. Even with frameworks such as Scrum or different agile approaches in place, missed deadlines and exceeded budgets are still common issues. Vaughn Vernon, a prominent expert in Domain-Driven Design (DDD) and author of “Implementing Domain-Driven Design” and “Domain-Driven Design Distilled,” presents an alternative: the “No Estimates” approach.
The Problem with Traditional Estimates
Vaughn doesn’t mince words:
“All software estimates are just guesses… Developers are the ones who pay for it because: ‘They were wrong. They don’t know what they’re doing,’” as executive management is known to say.
Traditional estimation methods often rely more on intuition than on data. As Vaughn explains, even well-meaning teams can fall into common traps:
“We sometimes feel good about what seems obvious… but there’s almost always some kind of trap lurking under what seems obvious. Those catch us off-guard when we reach them.”
As a result, teams often spend weeks estimating tasks that take less time to actually complete—a cycle that Vaughn aptly describes as an approach that’s as accurate as “licking your finger and sticking it into the wind.”
The “No Estimates” Solution: EventStorming and Steel Threads
Vaughn’s approach eliminates guesswork by providing a structured, scientific framework based on DDD principles and patterns. Central to this is the use of Design-Level EventStorming while applying the Steel Thread architecture pattern—an end-to-end workflow that models a specific business behavior by using a domain scenario.
“A steel thread represents a single behavioral scenario… It’s strong, important, and runs from user interaction to system outcome.”
By breaking projects into these threads, teams can apply metrics to individual components (e.g., user roles, user interfaces, commands, policies, aggregates, and events) using a spreadsheet with complexity-level metrics. Vaughn emphasizes this approach:
“The smaller the batches [of work], the more accurate [the estimate]. This is how our minds work best.”
This approach aligns with EventStorming, a collaborative modeling technique that visually maps business processes in domains. Teams use Big-Picture collaborative modeling to identify events, policies, user roles, and other high-level modeling elements at the outset, minimizing ambiguity and speeding up learning, and ultimately, delivery. Then they move on to Design-Level storming.
Why This Works
- Data Over Guesses: Assign a complexity metric (simple, moderate, or complex) to each component.
- Iterative Refinement: Small batches enable rapid adjustments.
- Business Alignment: Steel threads mirror real-world workflows, ensuring solutions evolve with business needs.
As Vaughn notes:
“This approach fits good design practices naturally… It’s based on what I do anyway.” This means there’s no wasted effort. Just pure continuous improvement.
Learn from Vaughn Vernon
Ready to ditch unreliable estimates? Vaughn’s Implementing Domain-Driven Design (IDDD) Workshops dive deeper into EventStorming, strategic design, architecture that includes Steel Threads, and tactical patterns for scalable systems. Participants gain lifetime access to resources, and practical toolkits.
Workshop Highlight:
- Hands-on exercises to model domains using EventStorming
- Strategies to transition from estimates to domain-driven planning
