Process Analysis and Value Stream Mapping

Introduction

Lean manufacturing and service operations seek to create customer value with the least necessary consumption of time, material, information, space and human effort. Process analysis provides the analytical foundation for this aim. It makes work visible, identifies the sequence and interaction of activities, distinguishes value-adding work from waste, and establishes a fact-based basis for improvement.

A process is a set of related activities that transforms inputs into outputs for an internal or external customer. In manufacturing, the transformation may convert raw material into a finished product. In services, it may convert a customer request and information into advice, approval, treatment or payment. Although the physical characteristics differ, both settings contain flows, queues, handoffs, decisions, defects and capacity constraints.

This chapter develops an integrated approach using SIPOC diagrams, process maps, service blueprints and value stream maps (VSMs). It also explains takt time, cycle time, lead time, throughput, capacity, work-in-process (WIP) and process efficiency. These measures should not be interpreted in isolation: improving one may worsen another unless the complete flow and customer requirement are considered.

Learning Objectives

By the end of this chapter, the reader should be able to:

1. Define process boundaries, customers, suppliers, inputs and outputs using SIPOC.
2. Construct and interpret manufacturing process maps and service blueprints.
3. Develop current-state and future-state value stream maps.
4. Calculate and distinguish takt time, cycle time, lead time, throughput, capacity, WIP and process efficiency.
5. Diagnose bottlenecks, queues, information delays and non-value-adding activity.
6. Prioritise improvement opportunities and evaluate the managerial limitations of process maps and lean metrics.

1. Process Thinking and Customer Value

A process has three essential properties. First, it has a purpose, normally expressed as a customer requirement. Second, it has a sequence of activities and decisions. Third, it produces measurable outputs. Process analysis therefore begins with the question: what outcome does the customer value, and what conditions must the process satisfy to produce it reliably?

In lean thinking, an activity is value-adding when it transforms the product or service in a way the customer requires, performs the transformation correctly the first time, and does not violate applicable obligations. Activities that consume resources without creating customer value are non-value-adding. Some are pure waste, such as unnecessary movement or avoidable rework. Others are currently necessary non-value-adding work, such as legally required documentation. The latter should be simplified or automated, but must not be removed without considering compliance and risk.

The principal forms of operational waste are defects, overproduction, waiting, unused human capability, transportation, inventory, motion and overprocessing. In services, these often appear as incomplete applications, premature work, customer waiting, poor use of employee knowledge, unnecessary transfers, case backlogs, repeated data entry and excessive approvals.

A process boundary must be explicit. A map that begins at “customer need” and ends at “customer satisfaction” may be conceptually attractive but operationally difficult to measure. A useful boundary might instead be “receipt of a complete loan application” to “communication of the lending decision.” The boundary determines which delays, resources and performance measures are included.

2. SIPOC: Establishing the Process Architecture

SIPOC stands for Suppliers, Inputs, Process, Outputs and Customers. It is a high-level scoping tool, not a detailed workflow. It prevents a team from beginning with local activities before agreeing on the process purpose and its upstream and downstream relationships.

A SIPOC should normally be developed at a level where the process can be represented by approximately four to seven major stages. Excessive detail defeats its scoping purpose. The team should validate definitions with process owners and customers, particularly the meaning of “complete input” and “acceptable output.”

SIPOC is valuable at project initiation because it exposes external suppliers, hidden customers and information dependencies. It does not show queue length, handoff timing, resource loading or the exact path of exceptions. Those require a more detailed map.

3. Process Mapping

A process map represents the logical sequence of activities, decisions, inputs, outputs and feedback loops. Common symbols include rectangles for activities, diamonds for decisions, arrows for flow and documents or information symbols for records. The symbol convention is less important than consistency and clarity.

A rigorous mapping exercise should:

• observe the actual process rather than rely only on procedures;
• include normal and exceptional paths;
• identify the responsible role at each handoff;
• record waiting and rework explicitly;
• distinguish information flow from physical flow; and
• validate the map with people who perform and receive the work.

A “work-as-imagined” map taken from a procedure manual often differs from “work-as-done.” The difference is analytically important. Informal workarounds may reveal system defects, but they may also be essential controls. Improvement should therefore investigate why variation exists before standardising or eliminating it.

Process Measures

Let:

• D = available operating time per period;
• Q = customer demand per period;
• t = processing time per unit;
• C = cycle time, the elapsed time between completed units under stated operating conditions;
• R = throughput, completed units per unit of time;
• N = average WIP.

The measures are related but not interchangeable. Processing time is the time a unit is actively worked on. Cycle time describes output spacing or completion rate under a defined operating condition. Lead time is the elapsed time from entry to exit, including processing, waiting, movement, inspection and delays.

4. Takt Time, Cycle Time and Capacity

Takt time translates customer demand into a required production rhythm. If available time and demand are measured over the same period:

Its units are time per demanded unit. For example, if a line has 420 productive minutes per shift and demand is 105 units per shift, takt time is 4 minutes per unit. This does not mean that every unit must be processed in exactly four minutes; it means that the line must complete one unit on average every four minutes to satisfy demand.

The calculation assumes that the selected available time excludes planned breaks, shutdowns and other unavailable periods, and that demand is expressed for the same shift or period. If demand is variable, a representative planning horizon or a demand distribution should be used rather than an arbitrary daily value.

Capacity for a stable repetitive operation can be expressed as:

Here effective cycle time includes the constraints of the resource, such as changeovers, planned downtime or quality losses, according to the capacity model being used. If multiple identical resources operate in parallel, ideal capacity may be multiplied by the number of resources, but only when work can actually be divided and demand is available.

For a single resource, the basic condition for meeting demand is:

This condition is necessary but not sufficient. A line may have adequate average cycle time and still miss demand because of downtime, variation, poor balance, defects or material shortages. In a multi-step line, the bottleneck is the step with the lowest effective capacity, or equivalently the longest effective cycle time after considering available resources.

Worked Example 1: Takt and Line Capacity

A packaging cell operates 7.5 hours per day. Breaks and planned meetings total 45 minutes. Daily demand is 102 cases. The cell has three sequential operations with observed cycle times of 3.2, 4.5 and 3.8 minutes per case.

Available time is 450 − 45 = 405 minutes per day. Therefore:

The second operation has the longest cycle time, 4.5 minutes, and is the bottleneck. Its theoretical daily capacity is:

Thus the cell cannot meet demand of 102 cases per day in its present configuration. Adding effort to the first or third operation will not remove the constraint. Management should investigate work content, staffing, equipment, method and quality losses at operation two.

5. Lead Time, Throughput, WIP and Little’s Law

Throughput is the rate at which units leave a process. In a stable process, average throughput is approximately equal to average demand served. Lead time is the elapsed time a unit spends in the system. WIP includes all units started but not yet completed, including units being processed, waiting, inspected or transported.

For a stable system with a consistent definition of WIP and lead time, Little’s Law states:

Equivalently:

The units must be consistent. If WIP is 240 cases and throughput is 60 cases per hour, average lead time is 4 hours. Little’s Law is a long-run relationship, not a promise that every individual case takes four hours. It assumes a stable system over the measurement period, a defined population, and no systematic accumulation or depletion of WIP.

Little’s Law is particularly useful because it shows why reducing WIP can reduce lead time when throughput is maintained. It also reveals a difficult trade-off: increasing WIP may protect utilisation or service availability, but usually increases lead time and hides process problems.

6. Process Efficiency and Flow Efficiency

A common process-efficiency indicator is the ratio of value-adding processing time to total lead time:

If a service request receives 18 minutes of actual analysis but takes 180 minutes from receipt to completion, process efficiency is 10%. The measure is meaningful only when value-adding time and lead time have compatible boundaries and definitions. It should not be interpreted as “employee efficiency.” A low ratio may result from queues, batching, approvals, information defects or demand variability.

A second useful comparison is schedule or service attainment, such as the proportion of orders completed by the promised date. Efficiency should be balanced with quality, safety, employee well-being, resilience and customer outcomes. Pursuing a high ratio by removing necessary controls can increase risk and total cost.

7. Service Blueprinting

A service blueprint adapts process analysis to the simultaneous and often intangible nature of service delivery. It commonly separates:

1. customer actions;
2. frontstage or visible employee actions;
3. the line of interaction between customer and provider;
4. backstage employee actions;
5. support processes and systems; and
6. physical or digital evidence experienced by the customer.

For example, in an online insurance claim, the customer submits a claim and receives updates; a visible service agent requests clarification; a claims analyst evaluates evidence backstage; and document management, payment systems and fraud controls provide support. The blueprint can show where the customer waits, where information is transferred, and where a failure becomes visible.

Blueprinting is especially useful for identifying fail points: locations where errors, ambiguity or delay can damage customer experience. It also distinguishes customer waiting from internal processing, which a conventional departmental flowchart may conceal. Service blueprints should include alternative paths for incomplete information, escalation, complaints and accessibility needs.

8. Value Stream Mapping

Value stream mapping examines the end-to-end flow required to deliver a product or service, including both material or customer flow and information flow. A current-state map records how the system operates now; a future-state map describes a proposed flow based on lean principles and operational constraints.

A typical VSM contains customer demand, supplier and customer information, process boxes, inventory or queue triangles, shipment or service delivery, data boxes and a timeline. Process data may include cycle time, changeover time, uptime, number of operators, batch size, WIP, defect rate and available time. In service operations, inventory symbols may represent cases waiting in a queue, while transportation may represent digital or physical handoffs.

Constructing a Current-State Map

1. Select a product family or service family with similar process steps.
2. Define the start and end boundaries and the customer requirement.
3. Walk the flow from the customer backward or from the beginning forward.
4. Record actual observations, including queues, rework and information delays.
5. Add supplier, customer, scheduling and replenishment information.
6. Calculate total processing time and total lead time.
7. Identify bottlenecks, unstable steps, large queues and disconnected schedules.

A product family should not be selected merely because it is convenient. Product or service families should share sufficiently similar routes so that one map represents a meaningful flow. If routes differ materially, separate maps or a carefully designed family matrix may be necessary.

Constructing a Future-State Map

A future-state design generally asks:

• What is the customer demand and takt time?
• Which process should control production or service release?
• Where can continuous flow be created?
• Where is a supermarket or controlled FIFO lane needed?
• How will work be scheduled and replenished?
• How can workload be levelled?
• What quality-at-the-source controls prevent defects moving downstream?
• What information is necessary, and how often should it be communicated?

The future state must be feasible. It should not simply remove every queue or assume perfect reliability. Some buffers are justified by variability, supplier risk, safety or the need to decouple processes. The design should specify ownership, standard work, measurement and an implementation sequence.

9. Integrated Worked Example: A Service-Request Process

Consider an illustrative internal IT service desk. A request enters through a portal, is classified, assigned to a technician, resolved and closed. Observation over one week produces the following averages:

The average lead time is therefore 180 minutes, assuming the stages and waits form the defined end-to-end boundary. Process efficiency is:

The service desk completes 320 requests during a 40-hour week. Average throughput is:

Applying Little’s Law, average WIP is:

where three hours is the average lead time. The largest waiting period is before diagnosis and resolution. This does not prove that the diagnostic team is the only cause: assignment rules, skill matching, priority policies and incomplete information may create the queue. A current-state map would therefore connect the queue to the relevant information flows rather than merely demanding that technicians work faster.

A plausible future state could include a standard request template, automatic categorisation for common requests, a visible FIFO lane for standard work, explicit escalation rules and a daily workload review. The design should be tested against security requirements, urgent incidents and the risk that automation misclassifies complex requests.

10. Prioritising Improvement Opportunities

A useful prioritisation method combines customer impact, frequency, delay, risk and feasibility. High-volume activities with long queues are attractive targets, but a low-frequency failure may deserve priority if it creates safety, legal or severe customer consequences.

Root-cause analysis should distinguish symptoms from causes. For example, “large queue at approval” is a symptom. Possible causes include incomplete inputs, unclear authority limits, excessive batching, uneven arrival patterns or unreliable software. Countermeasures should be linked to a verified cause and evaluated with before-and-after data.

Useful supporting tools include Pareto analysis for defect or delay categories, cause-and-effect analysis, standard-work combination sheets, workload balancing, mistake-proofing, control plans and visual management. Simulation or discrete-event modelling may be justified when variability, multiple priorities and complex interactions make static maps inadequate.

11. Technology, Data and Contemporary Practice

Digital process mining can reconstruct process paths from event logs, revealing actual sequence variation, rework and waiting. It is particularly useful in administrative and service processes, provided event definitions and timestamps are reliable. A process-mining model may show that a large proportion of cases bypass the documented route, but interpretation still requires interviews and contextual knowledge.

Artificial intelligence can assist classification, demand forecasting, document extraction and anomaly detection. However, automation is not automatically lean. An inaccurate classifier can create hidden rework, biased prioritisation or privacy risks. Governance should address data quality, explainability, access control, cybersecurity, human override and monitoring for unequal service outcomes.

Performance data should be collected for improvement rather than surveillance. Employees should understand how measures are used, and managers should avoid rewarding local utilisation or speed when those incentives damage quality, safety or the total flow. Regulatory obligations concerning records, accessibility, occupational safety and data protection remain part of the process design.

12. Limitations and Critical Evaluation

Maps are representations, not reality. They simplify variation, informal knowledge and social interaction. Average cycle and lead times can conceal wide distributions; therefore, percentiles, control charts or stratification by product, priority and shift may be required. Averages are especially dangerous in services where a small number of complex cases dominate customer harm.

Takt time is a demand-based planning rate, not a measure of actual capability. Cycle time below takt does not guarantee delivery if downtime or defects are high. Process efficiency can encourage harmful reduction of necessary controls if used as a stand-alone target. Little’s Law describes a stable aggregate relationship but does not identify the cause of delay. VSM can also become a one-time workshop artifact unless future-state actions have owners, dates and review measures.

Lean analysis should therefore be combined with statistical thinking, human-factors analysis, financial evaluation and risk management. The aim is not minimum activity at any cost; it is a capable, responsive and sustainable process that consistently delivers what customers and stakeholders legitimately require.

Applied Case: Reducing Delays in a University Admissions Process

Background

The following is an illustrative case. A university receives postgraduate applications through an online portal. Applicants complain that decisions are slow and that they receive repeated requests for information.

Problem

A SIPOC exercise identifies applicants, the portal, academic departments, admissions officers and the central records team as major participants. A current-state map shows that applications are downloaded in batches, manually re-entered, sent to departments, returned for clarification and then checked again by central staff.

Over a representative period, active work totals 52 minutes per application, while average lead time is 8 working days. The process-efficiency ratio, if one working day is defined as 480 minutes, is:

The low ratio indicates extensive waiting and batching, not that staff are unproductive.

Analysis

The largest delays occur before departmental review and during clarification. Interviews reveal that the portal permits applications to be submitted without mandatory evidence, and departments use different criteria for requesting additional documents. A local target based only on the number of files processed encourages batching at the department level.

Approach and Solution

The future-state design introduces mandatory input validation, a standard evidence checklist, daily electronic release of complete applications, a controlled FIFO queue and explicit escalation rules. A small protected capacity is retained for urgent scholarship deadlines. Staff are trained in the new standard work, and an exception path is maintained for unusual qualifications.

Outcome

After a pilot, average lead time falls to 3.5 working days, while first-time completeness improves. The university continues to monitor decision quality, applicant equity and staff workload rather than relying only on speed. The key lesson is that lead-time reduction came from improving information quality and flow control, not from demanding faster individual work.

Chapter Summary

Process analysis makes customer requirements, transformation activities, delays, handoffs and constraints visible. SIPOC establishes the high-level architecture; process maps describe sequence and responsibility; service blueprints add customer actions, frontstage and backstage work; and VSM connects material or case flow with information flow across the entire value stream.

Takt time expresses the required rhythm from available time and demand. Cycle time describes output spacing or processing capability under defined conditions. Capacity is constrained by available time and effective cycle time, while lead time includes processing and all forms of waiting. Throughput is the completion rate, WIP is unfinished work, and Little’s Law relates WIP, throughput and lead time in a stable system. Process efficiency compares value-adding time with total lead time but must be balanced with quality, risk, equity, resilience and employee well-being.

A current-state map should be based on observation and reliable data. A future-state map should specify feasible flow, pull or release controls, quality at the source, workload levelling, ownership and implementation measures. Lean metrics are powerful diagnostic tools, but none should become an isolated performance target.

Key Terms

SIPOC: A high-level representation of Suppliers, Inputs, Process, Outputs and Customers.

Process map: A visual representation of activities, decisions, sequence and handoffs.

Service blueprint: A service map separating customer, visible employee, backstage and support activities.

Value stream: The end-to-end activities required to deliver a product or service.

Value stream mapping: A visual method for analysing material, case and information flow.

Current-state map: A representation of how a process operates at present.

Future-state map: A proposed representation of a desired, feasible process condition.

Takt time: Available operating time divided by customer demand for the same period.

Cycle time: Time between completed units or time required to complete a defined operation, depending on the stated measurement convention.

Processing time: Time during which work is actively performed on a unit.

Lead time: Elapsed time from process entry to exit, including processing and delay.

Throughput: Completed output per unit of time.

Capacity: Maximum achievable output under specified resources and operating conditions.

Work-in-process: Units that have entered but not completed the process.

Bottleneck: The resource or activity constraining system throughput.

Process efficiency: Value-adding processing time divided by total lead time.

Little’s Law: The stable-system relationship WIP = throughput × lead time.

FIFO: First-in, first-out sequencing of work.

Fail point: A location where an error or delay can become visible to a customer or downstream process.

Review Questions

Short Answer Questions

1. What is the purpose of a SIPOC diagram?
2. Distinguish takt time from cycle time.
3. What activities are included in lead time but excluded from processing time?
4. State Little’s Law and its principal assumptions.
5. Why is a service blueprint useful in addition to a conventional process map?

Long Answer Questions

1. Explain how a team should construct and validate a current-state value stream map.
2. Discuss the relationship among takt time, bottleneck cycle time and capacity.
3. Compare process maps, service blueprints and value stream maps in terms of purpose and level of detail.
4. Critically evaluate process efficiency as a lean performance measure.
5. Explain how information quality, batching and queue design influence lead time in service operations.

Analytical / Application Questions

1. A process has 360 available minutes per day and demand of 90 units. Three sequential activities have cycle times of 3, 5 and 4 minutes. Calculate takt time, identify the bottleneck and estimate bottleneck capacity.
2. A service process completes 12 cases per hour and has average WIP of 36 cases. Estimate average lead time using Little’s Law. State one reason why the estimate may not describe every individual case.
3. Design a SIPOC and future-state improvement approach for a university, hospital or financial-service process of your choice. Identify one flow problem, one information problem, one risk and three measures for evaluating the change.

Multiple Choice Questions with Answers

Question 1

A process has 420 available minutes per shift and demand is 105 units. What is the takt time?

A. 2 minutes per unit

B. 4 minutes per unit

C. 4.5 minutes per unit

D. 5 minutes per unit

Correct Answer: B. 4 minutes per unit

Question 2

Which measure includes processing, waiting, movement and inspection time?

A. Processing time

B. Cycle time only

C. Lead time

D. Takt time

Correct Answer: C. Lead time

Question 3

Which SIPOC element identifies the parties receiving process outputs?

A. Suppliers

B. Inputs

C. Outputs

D. Customers

Correct Answer: D. Customers

Question 4

A process completes 8 cases per hour and has average lead time of 3 hours. According to Little’s Law, average WIP is:

A. 2.67 cases

B. 11 cases

C. 24 cases

D. 32 cases

Correct Answer: C. 24 cases

Question 5

If value-adding time is 20 minutes and total lead time is 200 minutes, process efficiency is:

A. 5%

B. 10%

C. 20%

D. 90%

Correct Answer: B. 10%

Question 6

In a sequential line, the bottleneck is generally the operation with:

A. The shortest cycle time

B. The longest effective cycle time

C. The largest physical area

D. The highest material cost

Correct Answer: B. The longest effective cycle time

Question 7

Which feature is especially characteristic of a service blueprint?

A. Only machine utilisation data

B. Separation of customer, frontstage and backstage actions

C. Exclusive focus on raw-material movement

D. Removal of all service exceptions

Correct Answer: B. Separation of customer, frontstage and backstage actions

Question 8

A future-state value stream map should primarily describe:

A. An ideal process with no constraints or variation

B. A feasible improved flow and its control mechanisms

C. Only the organisation’s financial budget

D. The historical process before measurement

Correct Answer: B. A feasible improved flow and its control mechanisms

Question 9

If available time is 405 minutes and effective cycle time is 4.5 minutes per unit, theoretical capacity is:

A. 90 units

B. 95 units

C. 102 units

D. 182 units

Correct Answer: A. 90 units

Question 10

Which statement about Little’s Law is correct?

A. It identifies the root cause of every delay

B. It applies only to manufacturing

C. It relates average WIP, throughput and lead time in a stable system

D. It states that cycle time must equal takt time

Correct Answer: C. It relates average WIP, throughput and lead time in a stable system