benchmarking-analyst

benchmarking-analyst

You are benchmarking-analyst - a specialized skill for benchmarking analysis including performance comparison and best practice identification.

Overview

This skill enables AI-powered benchmarking including:

• Internal benchmarking
• Competitive benchmarking
• Functional benchmarking
• Generic/best-in-class benchmarking
• Gap analysis
• Best practice identification
• Adaptation planning
• Performance tracking

Capabilities

1. Benchmarking Project Setup

from dataclasses import dataclass
from typing import List, Dict, Optional
from datetime import datetime
from enum import Enum

class BenchmarkType(Enum):
    INTERNAL = "internal"  # Compare within organization
    COMPETITIVE = "competitive"  # Compare with competitors
    FUNCTIONAL = "functional"  # Compare similar functions across industries
    GENERIC = "generic"  # Compare with best-in-class anywhere

@dataclass
class BenchmarkProject:
    title: str
    benchmark_type: BenchmarkType
    process_area: str
    metrics: List[str]
    partners: List[str]
    owner: str

def setup_benchmark_project(project: BenchmarkProject):
    """
    Set up benchmarking project structure
    """
    phases = {
        "1_planning": {
            "status": "in_progress",
            "tasks": [
                {"task": "Identify what to benchmark", "status": "complete"},
                {"task": "Identify benchmark partners", "status": "complete"},
                {"task": "Determine data collection method", "status": "not_started"},
                {"task": "Define metrics and calculations", "status": "not_started"}
            ]
        },
        "2_analysis": {
            "status": "not_started",
            "tasks": [
                {"task": "Collect current performance data", "status": "not_started"},
                {"task": "Collect partner performance data", "status": "not_started"},
                {"task": "Determine performance gaps", "status": "not_started"},
                {"task": "Identify enablers of superior performance", "status": "not_started"}
            ]
        },
        "3_integration": {
            "status": "not_started",
            "tasks": [
                {"task": "Communicate findings", "status": "not_started"},
                {"task": "Establish improvement goals", "status": "not_started"},
                {"task": "Develop action plans", "status": "not_started"}
            ]
        },
        "4_action": {
            "status": "not_started",
            "tasks": [
                {"task": "Implement improvements", "status": "not_started"},
                {"task": "Monitor progress", "status": "not_started"},
                {"task": "Recalibrate benchmarks", "status": "not_started"}
            ]
        }
    }

    return {
        "project_id": f"BM-{datetime.now().strftime('%Y%m%d')}",
        "title": project.title,
        "type": project.benchmark_type.value,
        "process_area": project.process_area,
        "metrics": project.metrics,
        "partners": project.partners,
        "owner": project.owner,
        "created_date": datetime.now().strftime("%Y-%m-%d"),
        "phases": phases,
        "status": "planning"
    }

2. Data Collection Framework

def create_data_collection_template(metrics: List[str], partners: List[str]):
    """
    Create template for benchmark data collection
    """
    template = {
        "data_points": {},
        "collection_guidance": {}
    }

    for metric in metrics:
        template["data_points"][metric] = {
            "our_performance": {
                "value": None,
                "unit": None,
                "period": None,
                "data_source": None,
                "confidence": None
            },
            "partners": {partner: {
                "value": None,
                "unit": None,
                "period": None,
                "data_source": None,
                "is_estimate": False
            } for partner in partners}
        }

        template["collection_guidance"][metric] = {
            "definition": f"Standard definition for {metric}",
            "calculation": f"How to calculate {metric}",
            "data_sources": ["System A", "Reports", "Surveys"],
            "normalization_notes": "Ensure comparable basis"
        }

    return template

def validate_benchmark_data(data: Dict):
    """
    Validate collected benchmark data
    """
    issues = []

    for metric, values in data['data_points'].items():
        # Check our data
        if values['our_performance']['value'] is None:
            issues.append(f"Missing our performance data for {metric}")

        # Check partner data
        partners_with_data = sum(1 for p in values['partners'].values()
                                 if p['value'] is not None)
        if partners_with_data == 0:
            issues.append(f"No partner data for {metric}")
        elif partners_with_data < len(values['partners']) / 2:
            issues.append(f"Limited partner data for {metric}")

        # Check comparability
        units = set()
        if values['our_performance']['unit']:
            units.add(values['our_performance']['unit'])
        for p in values['partners'].values():
            if p['unit']:
                units.add(p['unit'])
        if len(units) > 1:
            issues.append(f"Inconsistent units for {metric}: {units}")

    return {
        "is_valid": len(issues) == 0,
        "issues": issues,
        "completeness": calculate_completeness(data)
    }

def calculate_completeness(data):
    """Calculate data completeness percentage"""
    total_cells = 0
    filled_cells = 0

    for metric, values in data['data_points'].items():
        total_cells += 1  # Our data
        if values['our_performance']['value'] is not None:
            filled_cells += 1

        for partner_data in values['partners'].values():
            total_cells += 1
            if partner_data['value'] is not None:
                filled_cells += 1

    return round(filled_cells / total_cells * 100, 1) if total_cells > 0 else 0

3. Gap Analysis

import numpy as np

def perform_gap_analysis(data: Dict, higher_is_better: Dict = None):
    """
    Perform gap analysis between our performance and benchmarks

    higher_is_better: {metric: True/False}
    """
    higher_is_better = higher_is_better or {}
    gap_analysis = []

    for metric, values in data['data_points'].items():
        our_value = values['our_performance']['value']
        if our_value is None:
            continue

        partner_values = [p['value'] for p in values['partners'].values()
                         if p['value'] is not None]

        if not partner_values:
            continue

        # Calculate statistics
        best = max(partner_values) if higher_is_better.get(metric, True) else min(partner_values)
        worst = min(partner_values) if higher_is_better.get(metric, True) else max(partner_values)
        median = np.median(partner_values)
        mean = np.mean(partner_values)

        # Calculate gaps
        gap_to_best = best - our_value if higher_is_better.get(metric, True) else our_value - best
        gap_to_median = median - our_value if higher_is_better.get(metric, True) else our_value - median

        # Calculate percentile position
        all_values = partner_values + [our_value]
        all_values_sorted = sorted(all_values, reverse=higher_is_better.get(metric, True))
        our_rank = all_values_sorted.index(our_value) + 1
        percentile = round((1 - our_rank / len(all_values)) * 100, 1)

        gap_analysis.append({
            'metric': metric,
            'our_value': our_value,
            'best_in_class': best,
            'median': round(median, 2),
            'mean': round(mean, 2),
            'worst': worst,
            'gap_to_best': round(gap_to_best, 2),
            'gap_to_median': round(gap_to_median, 2),
            'gap_pe