In retail banking, the queue is rarely caused by one slow teller. It is usually the visible result of a connected value stream in which counter transactions, document verification, referrals, approvals, and adviser work compete for the same limited people and systems.
Value Stream Mapping (VSM) makes that whole journey visible. Instead of measuring only the time a teller spends with a customer, it captures the complete flow of material and information, from arrival and triage to transaction completion or adviser referral.
This distinction matters. A branch can have a 4.6-minute average counter transaction and still create a 2.1-day lead time for new account opening because work waits between stages. The fundamental purpose of VSM is to expose that gap and redesign the process around customer value, demand, capacity, and quality.
This guide uses a worked example of a branch serving 240 customers per day across four teller counters and two service pods.
1. Define the value stream before drawing the map
A useful bank-branch map begins with a clear scope. Mapping “all branch operations” produces a diagram that is difficult to measure and even harder to improve.
For this example, the scope is:
- Start: Customer arrives at the branch.
- End: Transaction is completed, or the customer is successfully referred and the related advice or application process is completed.
- Demand: 240 customers per day.
- Resources: Four teller counters and two service pods.
- Service families: Routine transactions, complex counter requests, and referrals for new accounts or products.
From the customer’s perspective, value means receiving an accurate outcome with minimal waiting, repeated explanations, or additional visits. From the bank’s perspective, the process must also satisfy risk, compliance, security, and documentation requirements.
That balance connects VSM to the Voice of the Customer, the Voice of the Business, and the Voice of the Process. A strong future state does not remove necessary controls; it removes avoidable delay and duplication around them.
For guidance on establishing boundaries, see this Lean Six Sigma project-scoping guide.
2. Current-state map: nine stages from arrival to completion
The current-state map should be built through direct observation, transaction data, staff interviews, and customer feedback. Do not rely only on documented procedures. The documented process describes what should happen; the map must show what actually happens.
Worked current-state flow
| Stage | Typical activity | Active time | Queue or delay | Observed issue |
|---|---|---|---|---|
| 1. Arrival and triage | Customer explains need and is directed | 1.0 min | 0–3 min | Complex requests enter the teller queue |
| 2. Ticket issue | Ticket or queue position is created | 0.5 min | 0–1 min | Ticket categories do not always reflect service time |
| 3. Counter transaction | Teller processes the request | 4.6 min | 6.8 min peak wait | Demand exceeds available counter capacity at peaks |
| 4. Document verification | Identity, forms, and supporting documents checked | 4.0 min | 3–20 min | Missing documents generate rework |
| 5. Referral handoff | Teller records and transfers referral information | 2.0 min | 8–25 min | Handoff depends on verbal messages or paper |
| 6. Adviser appointment | Customer waits for or attends adviser session | 25 min | Up to 1.5 days | Two service pods absorb variable demand |
| 7. Product application | Application information is entered and reviewed | 20 min | 2–8 hours | Data is sometimes re-entered |
| 8. Compliance checks | Verification, approval, and risk controls | 15 min | 4–24 hours | Approval batching creates delays |
| 9. Completion | Confirmation, document issue, and close-out | 3 min | 0–4 hours | Completion can wait for a final authorisation |
The branch’s first-contact resolution (FCR) is 63%, meaning 37% of customers require another interaction, referral, return visit, or follow-up. Counter rework is 22%, primarily associated with incomplete or incorrect documentation.
The process also uses approximately 34% of staff time on non-value-adding activity, including searching for documents, duplicating entries, waiting for dual authorisation, walking between desks, and correcting incomplete applications.
Current-state metrics
For the representative referral and new-account pathway, observed value-adding work totals approximately 75 minutes. The total lead time is approximately 2.1 days, or 3,024 minutes.
[
\text{PCE}=\frac{\text{Value-Added Time}}{\text{Total Lead Time}}\times100
]
[
\text{PCE}=\frac{75}{3,024}\times100=2.5%
]
A 2.5% Process Cycle Efficiency means the customer’s elapsed time is dominated by waiting, queues, rework, handoffs, and approval delays rather than active service.
Use the Process Cycle Efficiency Calculator to separate value-added time from non-value-added time and rank the largest sources of delay.
3. Identify all eight DOWNTIME wastes
The branch’s queue spiral is not one waste. It is a system of connected wastes.
- Defects: Missing identification documents, incorrect account details, incomplete forms, and data-entry errors create callbacks and repeat visits.
- Overproduction: Preparing product packs, printed forms, or referral work before customer demand and eligibility are confirmed creates unnecessary work-in-process.
- Waiting: Customers wait for a teller, tellers wait for dual authorisation, and applications wait in compliance batches.
- Non-utilised talent: Experienced tellers spend time searching for forms or correcting avoidable errors instead of resolving complex customer needs.
- Transportation: Paper files, signed forms, and supporting documents move between teller desks, service pods, printers, and approval points.
- Inventory: Open applications, unprocessed referrals, paper forms, and queued tickets become service inventory.
- Motion: Staff walk to shared scanners, printers, cash areas, or supervisors; customers move between counters and adviser pods.
- Extra-processing: Information is written on paper, entered into a queue system, retyped into a customer relationship platform, and entered again into an application.
These wastes interact. Missing documents create defects; defects create rework; rework consumes scarce staff capacity; reduced capacity increases waiting; increased waiting encourages more customers to join the queue at peak times.
4. Analyse the bottleneck using demand, capacity, and variation
The Analyse Phase of DMAIC should combine visual and statistical evidence rather than relying on opinions.
Start with a time observation sheet that records arrival time, queue entry, service start, service completion, referral time, adviser start, and final completion. Stratify the data by:
- Day of week and time of day.
- Routine versus complex transaction.
- Teller counter versus service pod.
- New customer versus existing customer.
- Complete versus incomplete documentation.
- Referral type and approval requirement.
The branch receives 240 customers per day. If 420 minutes of customer-facing capacity are available per counter, total teller capacity is:
[
4 \times 420 = 1,680 \text{ minutes per day}
]
Average demand requires:
[
240 \times 4.6 = 1,104 \text{ transaction minutes per day}
]
That average appears manageable. However, the 6.8-minute peak wait indicates that average capacity is masking variation. Demand arrives in waves, complex requests take longer, breaks reduce available capacity, and referral work interrupts routine flow.
The critical question is therefore not simply, “How fast is the teller?” It is:
Where does demand exceed effective capacity, and what variation causes the queue to grow?
Use run charts, Pareto analysis, box plots, and control charts to separate common-cause variation from special-cause events. A bottleneck may be the teller counter during lunchtime, the adviser pods after referral peaks, or compliance approval at the end of the day.
5. Build the future state around flow
The future-state map should not simply instruct staff to work faster. It should redesign the flow so that routine demand is separated from variable, complex work.
Future-state design principles
1. Introduce demand levelling
Use historical arrival data to create staffing coverage for predictable peaks, such as paydays, lunch periods, and month-end. Allocate one flexible team member to triage and referral support during high-demand windows.
2. Strengthen front-end triage
At arrival, classify customers into:
- Self-service or digital support.
- Express teller transaction.
- Standard teller transaction.
- Adviser referral.
- Document-preparation support.
This prevents a 25-minute advice request from entering the same queue as a simple deposit.
3. Create standard work
Standard work should define the best-known sequence for:
- Greeting and request clarification.
- Identity and document pre-check.
- Transaction processing.
- Referral criteria.
- Digital handoff.
- Completion confirmation.
A document checklist at triage can reduce the current 22% rework rate before the customer reaches a teller.
4. Deflect suitable demand to self-service
Customers needing balance information, routine statements, card support, or simple transfers may be directed to secure digital or self-service channels. Deflection must be supported, not forced: staff should remain available for customers who require assistance or have accessibility needs.
5. Use a pull-based referral system
Instead of pushing paper referrals to an adviser, place validated referrals in a visible digital queue. The adviser pulls the next customer based on priority, readiness, and available capacity.
6. Reduce approval batching
Use risk-based approval rules and defined service-level triggers. Routine, low-risk applications should move continuously, while exceptions receive specialist review.
Current versus future performance targets
| Measure | Current state | 90-day future-state target |
|---|---|---|
| Peak queue wait | 6.8 min | 3.0 min |
| Average counter transaction | 4.6 min | 4.1 min |
| First-contact resolution | 63% | 82% |
| New-account lead time | 2.1 days | 0.75 day |
| Documentation rework | 22% | 8% |
| Non-value-adding staff time | 34% | 20% |
| Representative PCE | 2.5% | 4.6% |
The future PCE assumes approximately 50 minutes of value-adding work within a reduced 1,080-minute lead time:
[
\frac{50}{1,080}\times100=4.6%
]
The target is not achieved by eliminating necessary compliance activity. It is achieved by reducing queues, repeat entry, batching, and avoidable handoffs.
6. Sequence the improvement across 90 days
Days 1–30: Measure and stabilise
- Confirm the project charter, scope, customer requirements, and operational definitions.
- Observe at least 200 customer journeys across peak and non-peak periods.
- Validate the current-state map and baseline metrics.
- Introduce a document pre-check and clear triage categories.
- Display daily wait, FCR, rework, and referral backlog measures.
Days 31–60: Pilot the future state
- Pilot demand-levelling rosters during the two highest-volume periods.
- Test an express pathway for routine transactions.
- Introduce standard work for tellers and advisers.
- Replace paper referral movement with a controlled digital handoff.
- Test self-service deflection for suitable low-complexity requests.
- Review results weekly using customer and staff feedback.
Days 61–90: Control and scale
- Establish control limits for queue wait, rework, FCR, and referral lead time.
- Create reaction plans when performance exceeds agreed thresholds.
- Audit adherence to standard work without turning the process into a compliance exercise.
- Confirm ownership between branch leadership, teller operations, advisers, and compliance.
- Scale successful changes to other branches only after the pilot demonstrates stable results.
Turn a branch map into measurable capability
Value Stream Mapping gives bank leaders a shared view of customer demand, staff capacity, information flow, and operational risk. It also gives improvement teams a disciplined way to connect Lean principles with DMAIC, statistical analysis, standard work, and control planning.
If you want to lead this type of improvement project, explore Lean Six Sigma online training or pursue the CSSC-accredited Green Belt course. The Green Belt pathway develops practical capability in process mapping, data collection, root-cause analysis, hypothesis testing, piloting, SPC, and sustaining gains.
Enrol in Lean Six Sigma certification training and learn how to transform queues, handoffs, and hidden delays into measurable customer value.
Kaizen. Kai-Care. Kai-Done. Lean Six Sigma.
