Discovery: Learning What's in Your Process

Discovery is the foundation of all process improvement. Before you can improve something, you need to understand what it actually is—not what people think it is, or what it's supposed to be, but what really happens.

What is Discovery?

Discovery involves observing and researching what's in a system or process. It identifies the "nouns and verbs" of your operations:

• Nouns: The things involved—facilities, equipment, people, documents, materials, data
• Verbs: The actions taken—transforming, routing, deciding, storing, communicating

Discovery answers fundamental questions:

• What enters the process?
• What transformations happen?
• What decisions get made?
• Where do things go?
• What comes out?

Discovery vs. Data Collection

These two activities are related but distinct:

Discovery says: "There's a review step where a supervisor approves orders over $1,000."

Data collection says: "The review step takes an average of 2.3 hours and happens 47 times per week."

You need discovery first so you know what to measure.

Discovery Techniques

1. Subject Matter Expert (SME) Interviews

The people doing the work know things that aren't written down anywhere. Structured conversations with SMEs are often your richest source of process knowledge.

Tips for effective SME interviews:

• Prepare questions but allow conversation to flow naturally
• Ask "show me" rather than just "tell me"
• Listen for exceptions - "Well, usually we do X, but sometimes..."
• Verify understanding by summarizing back what you heard
• Follow up on references to other people, systems, or documents

Sample interview questions:

2. Process Walk-throughs

Nothing beats seeing the process in action. Walk-throughs can be guided or unguided.

Guided walk-throughs:

• An expert leads you through the process
• They explain what's happening at each step
• You can ask questions in context
• Good for understanding intent and nuance

Unguided walk-throughs:

• You observe independently
• Follow the work as it flows
• Notice what actually happens vs. what's described
• Good for catching workarounds and actual behavior

3. Document Review

Existing documentation provides context and history, even if it's incomplete or outdated.

Documents to seek out:

• Process documentation - Procedures, work instructions, flowcharts
• System documentation - User manuals, data dictionaries, interface specs
• Training materials - What do new employees learn?
• Forms and templates - What information gets captured?
• Reports - What metrics are tracked? What exceptions are flagged?
• Organizational charts - Who's responsible for what?

What to look for:

• Gaps between documentation and reality
• Steps that lack clear ownership
• Decision points without clear criteria
• Handoffs between groups or systems

4. SIPOC Analysis

SIPOC provides a high-level framework for understanding any process:

Example: Expense Reimbursement SIPOC

Discovery Contexts

Existing Systems

Most discovery focuses on processes that already exist. You're learning about established operations to improve, adapt, or automate them.

Approach:

1. Review existing documentation (however incomplete)
2. Conduct SME interviews
3. Walk through the process physically
4. Observe actual operations
5. Compare findings to documentation
6. Document gaps and variations

New Process Design

When designing something new, discovery works backward from desired outcomes.

Approach:

1. Define what the process must produce
2. Identify what inputs are available
3. Work backward to determine necessary steps
4. Test assumptions through prototyping
5. Iterate based on what you learn

Real-World Discovery Examples

Manufacturing Facility

"I reviewed the P&IDs (process and instrumentation drawings) before visiting. During the plant tour, I could see how the actual equipment mapped to the drawings, ask questions about operations I didn't understand, and identify where the documentation was out of date."

Discovery activities:

• Document review: P&IDs, equipment specs, maintenance logs
• Guided walk-through with plant engineer
• Observation of shift operations
• Interviews with operators and maintenance staff

Insurance Claims Processing

"The policy manual described a linear process, but the adjusters showed me all the loops—how claims bounce back for more information, how supervisors get pulled in for judgment calls, how special cases get routed differently."

Discovery activities:

• Review of policy and procedure manuals
• Shadowing adjusters through multiple claim types
• System demonstrations showing screens and workflows
• Interviews with supervisors about exception handling

Software System Integration

"Before we could connect the new system, we needed to understand exactly what data the old system produced, what format it used, and how downstream processes depended on it. We interviewed data consumers we didn't even know existed."

Discovery activities:

• Technical documentation review
• Database schema analysis
• Interviews with system users across departments
• Data flow tracing from source to destination

Documenting Discovery

What you learn needs to be captured in ways that are useful for analysis and communication.

Process Maps

Visual representations of process flow are essential discovery outputs.

Entity Lists

Catalog the "nouns" you discover:

Decision Tables

Document the "verbs" of decision-making:

Common Discovery Pitfalls

Accepting Documentation at Face Value

Written procedures describe what should happen. Discovery reveals what actually happens. Expect gaps.

Stopping at the "Happy Path"

Don't just learn the normal flow. Exceptions, errors, and edge cases often reveal the most about how a process really works.

Forgetting Informal Processes

Much real work happens through workarounds, shortcuts, and tribal knowledge that never gets documented. Look for these.

Not Validating Findings

Always verify what you learn with multiple sources. What one person describes may not match what others see or do.

Discovery Checklist

Before moving to data collection, ensure you can answer:

• What triggers this process to start?
• What are all the inputs required?
• What are the major steps or phases?
• Who performs each step?
• What systems support the work?
• What decisions get made, and by whom?
• What are the possible outcomes?
• Who receives each output?
• What exceptions or variations exist?
• How is the process currently measured?