Numetrics, McKinsey R&D analytics solutions

We provide companies engaged in semiconductor (IC) and embedded software development with advanced analytics to deliver their products on time and on budget. By applying analytics-based decision-making and estimation methods, organizations can improve productivity and shorten product development cycles. Numetrics provides the tools that R&D organizations need to:

  • structure the product definition process
  • bring projects in on time and on budget
  • optimize resources in ways that improve productivity

Our solution is built atop large industry databases of 2,000+ integrated circuit (IC) and 1,400+ software project data that have been used to develop empirical models. These models calculate the design complexity, or difficulty, of a development project from the technical characteristics of the final product. Using this complexity metric, which represents the standard effort required for the project by an average team, project managers can generate more accurate plans for new projects and compare their teams’ effectiveness to their peers in the database.

Engagement Models

Consulting engagements

In this model, a dedicated McKinsey team comprising product development and Numetrics professionals conducts a set of analyses (analytics and/or qualitative assessments) to identify optimization opportunities and improvement actions for clients. McKinsey offers two Numetrics engagement options:

Quantitative R&D diagnostic
  • Baseline: Assess 5-7 completed projects to form an organizational baseline

  • Benchmark: Compare performance metrics and identify key differences

    • Across different project/teams
    • Against an industry peer group and best-in-class
  • Planning/risk assessment: Analyze 1-2 new or ongoing projects

    • Assess risk in current plan assumptions relative to execution performance achieved on previous projects and to industry peers
    • Apply predictive analytics to estimate high-confidence budget, schedule and resource plans
  • Duration: 4-6 weeks

Comprehensive R&D diagnostic

Includes the Quantitative R&D diagnostic, plus:

  • Root cause analysis: apply a range of qualitative and analytical diagnostic tools, including:

    • Fingerprinting: qualitative assessment of the development process
    • Deconstruction: assessment of completed projects to identify key issues
    • Lean: assessment of completed projects to identify non-value-add activites
  • Duration: 8-10 weeks

Annual subscriptions

Clients may license the following Numetrics SaaS (Software as a Service) solutions on an annual basis and log in directly to the application via standard Web browsers. Subscriptions are supplemented with expert consulting to ensure that clients are successful in using the solution independently. We offer subscriptions to the following tools and databases:

IC Industry Database

The IC Industry Database is composed of more than 2,000 benchmarked IC projects compiled from more than 75 semiconductor companies and covering 34 application market segments, including wireless and wired communications, computing and peripherals, entertainment, industrial, and transportation. Types of the projects include analog, mixed-signal, radio frequency, power devices, system-on-chip and advanced processors.

Clients’ projects are benchmarked against peer groups hand-selected from the IC Industry Database by McKinsey experts. Because the peer group is specifically chosen and the results normalized to the projects’ complexity, the benchmarking results have high relevance. Clients analyze the gaps to the peer group norm and best-in-class and develop a clear set of quantifiable improvement opportunities.

IC Project Planner

IC Project Planner uses analytics-based estimation technology to automatically generate accurate estimates of development cycle time and staffing requirements, from product concept to release-to-production. It tackles the two top reasons IC projects are delivered late: underestimating true complexity and over-estimating team productivity. Calculations are based on:

  • A patented method of calculating IC complexity
  • The measured productivity of the development team
  • Unique factors about the IC, including the project's staffing strategy

The built-in “what-if” simulation environment enables users to rapidly perform tradeoffs between the critical constraints on the project-cycle time vs. staffing-level, chip functionality and performance. It enables clients’ teams to make tough, fact-based decisions about their projects before the window of flexibility has closed.

Further, the analytics environment delivers an assessment of the manager's own plan for the project—how much does it stretch vs. the team's past performance?

Decision Dashboard

Decision Dashboard enables R&D managers to track all of their projects in one place and identify where development execution targets are being met or missed. It compares projects undergoing execution to both past projects’ performance and industry data, reporting on how well each project in the portfolio is performing compared to its plan, and whether the R&D organization is operating effectively to achieve overall product development objectives. A broad selection of charts and customizable reports support in-depth analysis and clear visualization of project metrics.

Software Industry Database

The Software Industry Database is composed of more than 1,400 benchmarked software projects compiled from over 400 companies and spanning 43 vertical market segments, 50 operating systems and 20 programming languages. Industry data includes a broad set of standard metrics that allow software managers to assess their teams’ competitiveness, pinpoint root causes of poor performance, evaluate the opportunities for improvement and close the gap to the best-in-class.

Software Project Planner

Software Project Planner enables embedded software R&D managers to estimate the minimum project staffing necessary to meet schedule and quality requirements. It provides the foundation for reliable planning at the projects outset—before any lines-of-code (LOC) estimates are available. Software Project Planner enables software engineering managers to:

  • Measure embedded software complexity
  • Produce no-surprise staffing plans and schedules
  • Optimize tradeoffs between staffing, schedule, features and quality
  • “Should-cost” 3rd party development

The core technology that underpins Software Project Planner is embedded in engines that calculate the complexity of the software and generate estimates of timeline and staffing required for the project.

Multi-Project Pipeliner

Multi-Project Pipeliner matches available resources to the product development pipeline and provides visibility into future staffing bottlenecks. Staffing estimates from the Project Planner tools are aggregated in Pipeliner to give a complete view of total resources over time, allowing managers to make project and staffing tradeoffs, down to the role level, to optimize resource utilization and achieve revenue objectives.

Data Miner

Data Miner allows engineering organizations to diagnose the bottlenecks hampering product development effectiveness. Data Miner analyzes causal factors and highlights their influence on the productivity of project teams. Data Miner leverages project data to reveal the statistical relationships between project and team environment factors and performance metrics to allow engineering managers to quantify the impact and define improvement initiatives.