Many teams struggle with one persistent problem: delivery is not predictable. Plans are made with confidence, but outcomes consistently deviate—sometimes significantly. This becomes even more problematic in larger organizations where planning, budgeting, and stakeholder expectations depend on reliable forecasts.

In this talk, I share a real-world, data-driven approach to improving delivery predictability in a scaled environment. The approach was developed and applied in a large governmental ecosystem in the Netherlands, where multiple teams work in a complex, distributed setup.

Instead of relying on velocity or gut-feel estimation, the method combines:

• relative sizing (e.g., T-shirt sizing),
• available team capacity,
• and historical delivery data,

to create a simple but effective forecasting model.

I will walk through:

• how we built a normalized productivity metric based on historical data,
• how we used this to forecast delivery more reliably,
• and what actually happened in practice

The talk focuses on lessons learned, including:

• the impact of team composition changes,
• the hidden cost of non-delivery work (e.g., onboarding, knowledge transfer),
• and the importance of correctly modeling capacity.

Attendees will leave with:

• a practical way to improve predictability without adding heavy processes,
• insights into why their current forecasts might be unreliable,
• and concrete ideas they can apply immediately in their own teams.

This is not a theoretical model, but a field-tested approach based on multiple planning cycles and real delivery data.

Dr. Wasim Alsaqaf is an Agile Transformation Consultant, Agile Coach, and researcher specialized in large-scale agile software development,