7 Approach and Methodology

7.1 General Approach

Because the Business Understanding phase is pivotal to a project’s ultimate success, it needs to be inclusive, iterative, and have specific deliverables that must be met to be effective.

Inclusive Discovery

Here is a list of typical participants in the Business Understanding phase for a large-scale project. These are the people who will be directly affected by the project or have insights that will be used to inform the requirements and/or project design. Usually, there will be a multi-day kickoff to perform the Business Understanding discovery with the entire group present. This is important to ensure that business problem framing and solution framing are aligned and that roadblocks are identified quickly and addressed inline. While this appears to be a large number of participants, some of the roles outlined here can be filled by a single person. Many projects have ultimately been derailed because key participants were not included in the Business Understanding meetings, leading to unidentified roadblocks or business objectives.

Strategic stakeholders

Generally, executives at the SVP level or higher focus on the strategic objectives of the company. In some organizations, Directors also fill this role. Higher-level executives want to ensure that projects align with strategic objectives.

C-level executives (including CEO, CFO, etc.)
Executive Vice Presidents (EVP-level) executives
Senior Vice Presidents (SVP-level) executives
Directors (in some organizations)

Tactical stakeholders

Managers
Team members

Tactical stakeholders are the process owners. They perform the “day-to-day” process work. Often, tactical stakeholders are the teams that will use the data science solutions to bring about change. They ensure continuity with current business processes and provide valuable feedback on how the data science-driven solution should work.

Data subject matter experts

Business analysts
Data engineers

Data subject matter experts understand the structure, semantic meaning, and business importance of data in corporate data stores. They identify potential issues with data availability, quality, privacy, and other concerns.

Information technology (IT)

Enterprise architects
Software developers

Enterprise architects understand the topology of hardware infrastructure and design of software systems and how they integrate to deliver business solutions. They ensure that the data science solution follows architectural and security standards. Software developers create the necessary middleware (software that contains business logic) and user interface elements of the data science solution.

Data science team

Data scientists
Business analysts

Data scientists build predictive and prescriptive models that drive both tactical and strategic decisions. Business analysts deliver business intelligence reporting, dashboarding of Key Performance Indicators (KPIs), and visualizations.

Specific Deliverables: Business Understanding Document

The Business Understanding document serves as the foundational blueprint for any data science project, providing a comprehensive framework that connects business objectives with technical implementation. Its primary purpose is to ensure all stakeholders have a clear, shared understanding of what the project aims to achieve, how it will be executed, and how success will be measured. This document acts as both a strategic compass and an operational roadmap. Starting with strategic alignment and stakeholder analysis, it establishes the “why” and “who” of the project. The requirements documentation and feasibility assessment sections then build upon this foundation to define the “what” and “how possible.” The business process impact analysis ensures we understand the organizational changes needed, while the KPIs and success criteria establish how we’ll measure progress and success. Finally, the project plan transforms all these insights into actionable steps. Most importantly, this document serves as a single source of truth that teams can reference throughout the project lifecycle to maintain alignment, resolve disputes, and make informed decisions. It should be treated as a living document that can be updated as new information emerges or business conditions change while still maintaining its core purpose of keeping the project aligned with business objectives.

Overview of strategic alignment

Strategic alignment focuses on ensuring the data mining initiative directly supports business objectives and strategy. This includes documenting current business challenges, identifying key stakeholders and their needs, and understanding how the project’s outcomes will drive business value. It also involves mapping organizational capabilities, constraints, and assumptions that could impact project success.

Stakeholder analysis

A detailed mapping of:

Key stakeholders and their roles
Communication requirements and preferences
Decision-making authority levels
Expected involvement and responsibilities
Impact assessment for each stakeholder group

Requirements documentation

A comprehensive documentation of both functional and technical requirements, including:

Data requirements and sources
System integration needs
Security and compliance requirements
Stakeholder access and reporting needs
Performance and scalability requirements
Data governance and privacy considerations

Feasibility assessment

A thorough analysis that evaluates:

Technical feasibility (available tools, data, expertise)
Economic feasibility (ROI analysis, cost-benefit breakdown)
Operational feasibility (organizational readiness)
Legal and ethical considerations
Data availability and quality assessment
Resource availability and constraints

Business process impact analysis

Documentation of how the data mining project will affect:

Current business processes
Workflow changes needed
Training requirements
System integrations
Organizational change management needs
Dependencies on other systems or processes

KPI and metrics

The KPI and metrics deliverable establishes quantifiable measures that will be used to track project progress and success. This includes identifying both business KPIs (like revenue growth, customer retention, or cost reduction) and technical KPIs (model accuracy, precision, recall). Each metric should have a clear baseline, target value, and measurement methodology. The document should also explain how these metrics connect to broader business objectives.

Success criteria

Success criteria define the specific thresholds and conditions that must be met for the project to be considered successful. This deliverable includes both business and technical success criteria:

Business criteria might specify required improvements in operational efficiency, revenue targets, or cost savings
Technical criteria outline required model performance levels, data quality standards, and system performance requirements

The document should also detail how success will be measured and validated throughout the project lifecycle.

Project plan

The project plan provides a comprehensive roadmap for executing the data science initiative. It includes:

Detailed timeline with major milestones and dependencies
Resource requirements (technical, human, and infrastructure)
Risk assessment and mitigation strategies
Budget allocation and tracking mechanisms
Communication plan for stakeholders
Quality assurance approach and checkpoints
Change management strategy
Contingency plans

The plan should be realistic, accounting for available resources and organizational constraints while maintaining alignment with business objectives.

7.2 Methodology

Performing Business Discovery and Framing the Problem

Here are some requirements for effectively performing discovery with business stakeholders:

Start with why

Begin every discovery session by understanding the fundamental business drivers
Ask “Why now?” to understand urgency and priority
Connect the initiative to strategic business objectives
Define clear, measurable success metrics aligned with business KPIs

Stakeholder mapping

Create a comprehensive stakeholder map early in the process
Identify both direct and indirect stakeholders
Include domain experts who understand the business context
Document each stakeholder’s interests, influence, and concerns
Map dependencies between stakeholder groups

Current state analysis

Document existing business processes and workflows
Identify pain points and inefficiencies
Map the current data landscape and systems
Understand existing solutions and why they’re insufficient
Analyze current performance metrics

Problem definition

Use root cause analysis to identify underlying issues
Quantify the business impact of the problem
Define clear boundaries and scope
Distinguish between symptoms and core problems
Identify interconnected challenges

Value proposition

Calculate potential ROI and business benefits
Identify both tangible and intangible benefits
Estimate resource requirements and costs
Set realistic timeline expectations
Consider long-term strategic value

Data science goals

Translate business objectives into specific data science goals
Define clear success criteria for the analytical solution
Identify technical objectives and requirements
Document constraints and limitations
Align with data governance policies

Risk assessment

Identify potential technical and business risks
Assess data quality and availability risks
Develop mitigation strategies
Consider regulatory and compliance implications
Evaluate security considerations

Solution exploration

Explore multiple potential approaches
Conduct feasibility analysis for each option
Identify quick wins and long-term solutions
Analyze trade-offs between different approaches
Consider scalability requirements

Change management

Develop a stakeholder communication plan
Identify training and support needs
Plan for organizational adoption
Consider cultural impacts and resistance
Create feedback mechanisms

Documentation

Maintain clear documentation of requirements
Record assumptions and dependencies
Create a glossary of business terms
Document decisions and their rationale
Establish version control

Each of these steps should be applied iteratively throughout the business understanding phase. Remember to:

Validate findings with stakeholders regularly
Update documentation as new information emerges
Keep focus on business value rather than technical solutions
Maintain clear communication channels with all stakeholders
Foster collaborative problem-solving

Framing/Presenting a Data Science Based Solution Back to the Business Stakeholders

Here are some requirements for effectively presenting data science solutions to business stakeholders:

Executive summary focus

Begin with clear business value proposition
Highlight key findings and recommendations
Present ROI and business impact metrics upfront
Keep technical details in appendices
Address strategic alignment

Story structure

Start with the business problem context
Narrate the analytical journey
Show the solution evolution
End with concrete impact and results
Include real-world examples

Business metrics first

Lead with improvements to key business KPIs
Quantify cost savings and efficiency gains
Highlight new revenue or growth opportunities
Compare results to initial success criteria
Show trends and projections

Visual communication

Use clear, business-focused visualizations
Create interactive demonstrations when possible
Develop intuitive business dashboards
Avoid complex technical plots
Ensure accessibility

Technical translation

Convert technical concepts into business language
Use relevant industry analogies
Provide real-world examples
Layer technical details based on audience
Create relatable metaphors

Implementation plan

Outline clear deployment steps
Detail resource requirements
Present realistic timelines
Include key milestones and dependencies
Address operational impacts

Risk management

Be transparent about limitations
Clarify key assumptions
Present mitigation strategies
Address potential failure modes
Include contingency plans

Adoption strategy

Detail user training requirements
Present change management approach
Outline ongoing support plan
Include feedback mechanisms
Consider cultural factors

Future roadmap

Suggest next steps and phases
Present scaling opportunities
Identify potential enhancements
Link to long-term business strategy
Consider emerging technologies

Stakeholder engagement

Prepare for common questions
Identify key decision points
Define ongoing success metrics
Plan for different audience levels
Build consensus

Key Presentation Principles:

Tailor the depth based on audience
Focus on outcomes over methods
Use progressive disclosure for technical details
Encourage interactive discussion
Prepare multiple paths through the material
Have technical details ready but not front-loaded
Maintain engagement through storytelling

Remember to:

Prepare backup slides for detailed questions
Have specific examples ready
Know your audience’s technical level
Be ready to pivot based on stakeholder reactions
Document key decisions and next steps
Follow up on action items promptly