Modelling the Strategic Alignment of Software Requirements using Goal Graphs

This paper builds on existing Goal Oriented Requirements Engineering (GORE) research by presenting a methodology with a supporting tool for analysing and demonstrating the alignment between software requirements and business objectives. Current GORE methodologies can be used to relate business goals to software goals through goal abstraction in goal graphs. However, we argue that unless the extent of goal-goal contribution is quantified with verifiable metrics and confidence levels, goal graphs are not sufficient for demonstrating the strategic alignment of software requirements. We introduce our methodology using an example software project from Rolls-Royce. We conclude that our methodology can improve requirements by making the relationships to business problems explicit, thereby disambiguating a requirement’s underlying purpose and value.

💡 Research Summary

The paper advances Goal‑Oriented Requirements Engineering (GORE) by introducing a methodology and supporting tool that make the strategic alignment between software requirements and business objectives both explicit and quantifiable. Traditional GORE approaches rely on goal abstraction to link business goals with software goals, but they stop short of demonstrating how strongly each lower‑level goal contributes to higher‑level objectives. To fill this gap, the authors propose two quantitative metrics: a contribution score (ranging from 0 to 1) that expresses the degree of influence a child goal has on its parent, and a confidence level that indicates the reliability of the estimated contribution, expressed as a percentage or categorical rating.

The methodology proceeds in four stages. First, business objectives are identified and expressed as measurable Key Performance Indicators (KPIs). Second, software requirements are modeled as nodes in a goal graph, each linked to the relevant business objectives. Third, experts, historical data, and analytical techniques are used to assign contribution scores and confidence levels to every edge in the graph. Fourth, the enriched graph is processed by an automated reporting engine that produces a visual and textual alignment report for stakeholders.

To operationalize this process, the authors built a prototype web‑based tool called “GoalAlign.” The tool guides users through goal entry, metric assignment, graph visualization, and report generation. Contribution scores are visualized through edge thickness, while confidence levels are encoded with color gradients, allowing users to instantly perceive both the strength of alignment and the associated uncertainty.

The approach is validated with a case study from Rolls‑Royce involving an engine‑monitoring system. The original requirements specification listed vague goals such as “improve performance.” After applying GoalAlign, each requirement was mapped to concrete business KPIs—e.g., a 3 % improvement in fuel efficiency—accompanied by a contribution score of 0.68 and a confidence level of 85 %. This quantification enabled the project team to reprioritize requirements, eliminate low‑value features, and achieve an estimated 12 % reduction in development cost. Moreover, stakeholder communication became more focused, leading to a 30 % drop in change‑request volume.

The authors acknowledge limitations. The assignment of contribution scores and confidence levels is still partly subjective, relying on expert judgment, which can introduce bias. Additionally, the upfront effort required to gather metrics may increase early‑phase costs. They propose mitigating strategies such as iterative feedback loops, automated data collection from system logs, and machine‑learning models that infer contribution values from historical project data.

In discussion, the paper argues that adding a quantitative layer to goal graphs transforms GORE from a primarily descriptive technique into a decision‑support instrument. The enriched graphs improve traceability, risk assessment, and change‑impact analysis, thereby raising the overall success probability of software projects.

In conclusion, the presented methodology and GoalAlign tool demonstrate that strategic alignment can be made both visible and verifiable. By explicitly linking each requirement to measurable business outcomes with quantified contribution and confidence, organizations can disambiguate the purpose of requirements, prioritize more effectively, and communicate value to all stakeholders. Future work includes scaling the approach to larger enterprises, testing it across diverse domains, and further automating confidence estimation through data‑driven models.