R&D Concentration: A Leading Indicator for Development Efficiency

Background

In the field of R&D management, the industry has long been searching for metrics that can measure delivery efficiency. Common metrics include throughput (quantity), development cycle (speed), and resource utilization (efficiency). In practice, these three cannot serve as a guiding North Star metric.

1) Throughput counts the number of projects delivered in a period. It is a lagging indicator that only reflects the endpoint of a project, missing the opportunity for timely intervention. For example, a throughput of 0 in April does not mean production stopped, and a throughput of 1 in May does not guarantee the health of a five‑month‑long Project D.

Figure 1. Throughput counted in different months for multiple projects

2) Development Cycle is determined by the critical path of a single project. Personnel on the critical path are often distracted by ad‑hoc bugs or forced to wait for hand‑offs, which makes the cycle time an unreliable indicator of resource efficiency. In Figure 2, Person A leaves midway to handle external tasks, then waits for Person B to continue.

Figure 2. Project affected by out‑of‑plan work

3) Resource Utilization measures how saturated employees are with work. High saturation does not guarantee focus; if effort is scattered across many tasks, it does not accelerate project delivery.

Is there a single North Star metric that can provide real‑time feedback on R&D process efficiency and enable timely improvement?

Metric Introduction

The Youzan Efficiency Improvement team defined a new metric called R&D Concentration . It integrates throughput, cycle time, and resource utilization to reflect the decision benefit of “investing resources to shorten project cycles”. The calculation formula is:

R&D Concentration = (Project Workload in Person‑Days) ÷ (Development Cycle × Number of Participants) × 100%

Scenario A – Single Person Fully Loaded

When one person completes all work at full load, the concentration is 100% (10 person‑days ÷ (10 working days × 1) × 100%).

Figure 3. Single person fully loaded

Scenario B – Single Person Slowed by Distractions

If the same person is diverted by external tasks, the cycle extends (e.g., 15 working days) while workload stays 10 person‑days, yielding a concentration of 66.7%.

Figure 4. Single person slowed by distractions

Scenario C – Two People in a Tight Sequential Flow

When two roles (e.g., development then testing) work sequentially, each waits for the other. With 10 person‑days, a 10‑day cycle, and 2 participants, concentration drops to 50%.

Figure 5. Two people tightly coupled

Scenario D – Front‑Loading a Dependent Task

If Person B pre‑does part of the work that normally follows Person A, the cycle shortens (e.g., 8 working days), raising concentration to 62.5%.

Figure 6. Front‑loading a downstream task

Scenario E – Parallel Work with One Critical Path

When Person B works in parallel without depending on Person A, but Person A remains the critical path, concentration can reach 83.3% (10 person‑days ÷ (6 working days × 2) × 100%).

Figure 7. Two people working in parallel

Scenario F – Even Split of Work

If two people share the workload evenly, the cycle halves (5 working days) and concentration reaches 100%.

Figure 8. Two people each handling half the work

Practice Application

The histogram below shows the R&D Concentration distribution for a specific business line over a certain period. The highlighted red bars indicate that the most concentrated projects fall within the 12%–28% range.

Figure 9. R&D Concentration histogram

Case Studies

Positive case – Project A : 3 participants (frontend, backend, testing), 20‑day cycle, 45 person‑days workload → Concentration = 75%.

Figure 10. Gantt chart of Project A

Negative case – Project B : 8 participants, 43‑day cycle, 26 person‑days workload → Concentration = 8%.

Figure 11. Gantt chart of Project B

Analysis of Project B reveals several improvement opportunities: (a) large disparity in core developers’ time leading to a critical path; (b) participants with minimal workload that could be reassigned; (c) clear hand‑off between development and testing that could be shifted left or automated.

Conclusion

The R&D Concentration metric is a leading indicator that directly reflects development efficiency for any project. Because it can be measured continuously, problems can be detected early and corrective actions taken promptly. The authors hope readers will adopt this metric in broader measurement scenarios.

Further Reading

Measurement Practices in Efficiency Improvement

System Thinking in Project Management

Four Stages of Organizational Agile Transformation

Improving Meeting Efficiency

The Six Practices of Efficiency Improvement

Project Meeting Practices in Efficiency Improvement

🔔 If you are interested in efficiency improvement, feel free to join the Youzan Efficiency Improvement team by sending your résumé to [email protected] .