When Forrester updates a best practice report, it’s rarely just a technical refresh — it’s a signal to the industry about what to pay attention to next. The newly updated “Modern Development Metrics That Really Matter” by Diego Lo Giudice and team (November 2025) is no exception. The report reframes how modern software development teams — powered by Agile, DevOps, and increasingly, AI — should measure their success.
It replaces output obsession with outcome intelligence. But while the report is rich in frameworks and logic, the question remains: Can the measurement models it proposes truly scale across hybrid, AI-assisted, and low-code/no-code ecosystems?
Let’s unpack that.
Escaping the Iron Triangle — or Bending It Differently
The report begins by revisiting Martin Barnes’ classic “iron triangle” — time, scope, and resources — and suggests that Agile, aided by AI, has allowed teams to “flex” it. Forrester rightly points out that modern development success is about value delivery within those constraints, not just hitting deadlines or sticking to a fixed scope.
This is a needed evolution. Traditional metrics like “velocity” or “number of commits” often reward motion, not progress. But in a world where teams use AI copilots and no-code platforms to build applications faster than ever, scope itself becomes fluid. It’s not just about what gets built, but whether what’s built matters.
Still, the critique here is that Forrester’s report — while modern in framing — underestimates how platform abstraction is changing the triangle itself. When citizen developers use no-code tools or AI agents to automate workflows, time and resources aren’t just flexed — they’re redefined. Scope is no longer a constraint but an adaptive variable. Metrics, therefore, must capture adaptability, not just agility.
The Shift to Business Value as the “Center of Gravity”
Forrester’s biggest conceptual shift is placing business value at the center of its MAD (Modern Application Development) metrics framework. Surrounding it are four supporting metric types: progress, quality, efficiency, and engagement.
It’s a strong move. For too long, software measurement frameworks treated value as a byproduct — a lagging indicator revealed in revenue charts. Forrester brings it front and center, urging leaders to measure not just how fast teams deliver but how relevant that delivery is.
Yet, there’s a practical dilemma. Measuring business value in software development is complex and often subjective. How do you quantify “value” in internal tools, process automation, or AI-assisted content workflows?
Here’s where the report could have gone further — by acknowledging the role of agentic AI platforms and no-code environments in making value measurement more visible. Today’s AI-infused systems can trace a feature’s impact in real time — from workflow efficiency gains to reduced rework cycles — by correlating data streams across the delivery pipeline.
No-code dashboards, meanwhile, enable non-technical managers to visualize these metrics without dependency on engineering teams. This democratization of visibility could very well be the next evolution of Forrester’s MAD framework — value made transparent through autonomous systems.
Beyond Velocity: When Speed Misleads
“Forrester’s warning that velocity means nothing without quality” might sound obvious, but it’s still an industry trap. The report’s inclusion of real-world anecdotes — teams gaming metrics or reclassifying defects to meet KPIs — underscores how easily data becomes theater.
The critique here is less about Forrester’s insight and more about what it implies: our measurement culture is reactive. We still measure after the fact.
Modern AI-enabled delivery doesn’t have to. Agentic AI, embedded in development pipelines, can act as a real-time evaluator — flagging quality issues, testing AI-generated code, and predicting defect density before release. These systems aren’t passive dashboards; they learn, intervene, and optimize flow metrics automatically.
Forrester hints at this evolution when it discusses “code accepted from AI” and “rework of AI-generated code” as new measures. But this addition feels incremental rather than transformative. The next generation of metrics should capture not just how teams react to AI, but how AI acts as a teammate — augmenting efficiency, reducing friction, and accelerating continuous learning.
Engagement Metrics: Measuring What Machines Can’t
One of the most thoughtful inclusions in Forrester’s updated framework is engagement metrics — reflecting the human side of software delivery. The pandemic and remote work era exposed how culture, morale, and collaboration directly affect output quality. Forrester suggests tracking participation in hackathons, NPS surveys, and even “digital mood marbles” to monitor developer sentiment.
It’s an important shift — from seeing teams as delivery units to viewing them as dynamic systems. But engagement measurement can’t remain anecdotal. Here again, AI and no-code ecosystems can play a role. By combining behavioral analytics (commit frequency, review participation) with feedback automation, organizations can move from measuring engagement to cultivating it.
Imagine agentic AI copilots nudging developers when feedback loops slow down or auto-summarizing team retrospectives into actionable patterns. Or no-code tools that let team leads create morale dashboards tied to productivity flows. Engagement isn’t just a sentiment to measure — it’s a lever to engineer.
From Flow Metrics to Flow Intelligence
Forrester dedicates a substantial portion of the report to flow metrics — measuring how work moves across the software value stream. CTOs like Mik Kersten (Planview) have long championed this approach, arguing that flow visibility is essential for aligning DevOps and business outcomes.
The critique? Flow metrics alone can’t explain why flow slows down. They’re descriptive, not diagnostic.
The next frontier is flow intelligence — where AI agents actively interpret bottlenecks and autonomously trigger process adjustments. A slowdown in pull requests could, for example, prompt an AI agent to recommend code modularization or suggest team redistribution. With no-code/AI orchestration, even non-developers can intervene, optimizing delivery without writing a line of code.
Forrester’s model sets the stage for this — but doesn’t quite get there. The industry’s real evolution isn’t just in measuring flow but in closing the feedback loop automatically.
The Seven Habits: Practical but Conservative
The report concludes with seven “habits” for effective measurement — from tailoring metrics by context to correlating telemetry through ML. These are sound and actionable recommendations, especially the call to correlate leading and lagging metrics to predict value outcomes.
However, they’re still rooted in human-led interpretation. In reality, organizations are moving toward systems that self-measure, self-correct, and self-report. Platforms that combine AI-driven analytics with no-code automation already enable this. For example:
AI copilots can automatically benchmark team performance and alert leaders to anomalies.
No-code workflow builders can integrate telemetry from DevOps tools into business dashboards without manual scripting.
Agentic AI systems can link quality metrics to customer outcomes — turning correlation into continuous improvement.
The “seven habits” framework, therefore, feels like an important bridge — but one that will soon give way to autonomous measurement ecosystems where context, health, and value evolve together.
Toward a More Adaptive Metrics Culture
Forrester’s Modern Development Metrics That Really Matter delivers a crucial message: success in modern development isn’t about counting commits or sprint velocity — it’s about understanding the value and flow of outcomes. The framework is robust, research-backed, and much-needed in an era of AI-assisted delivery.
Still, it’s not the final word.
The next wave of measurement will go beyond defining metrics — it will redefine the act of measuring itself. No-code platforms will make metric orchestration accessible to every business unit. Agentic AI systems will interpret and act on data in real time. Developers and non-developers alike will move from reporting performance to engineering performance.
When that happens, the question won’t be “What metrics really matter?” — it will be “What metrics measure themselves?”