An ERP program has 8 divisions, 180 high-level requirements, a $12 million budget, and 12 weeks to approve scope. Would you run requirements centrally or federate the analysis, and what would you put in place first?

Discovery should be federated by division, with one central requirements architecture and approval gate. - I would assign a domain BA to each division and give all eight the same capability map, requirement hierarchy, and definition of ready. - A central team would own cross-division rules, shared data definitions, and the 25 or so interfaces where local decisions can create rework. - By week 4 I would baseline the top-level scope, then use weekly dependency reviews and a decision log to reach approval by week 12.

Six divisions disagree on who may approve customer-credit rules for a $4 million platform due to start build in 5 weeks. Sales wants revenue ownership, Risk wants veto power, and Operations owns execution; what decision-rights model do you choose?

Credit Risk should own policy and declines, Operations should own execution, and Sales should have input without override authority. - Any exception to a decline would require formally delegated Credit Risk authority; a salesperson or revenue owner could request an exception but never approve it. - The rights matrix would name one approver for each policy, limit, evidence, and exception decision, with segregation between revenue ownership and credit authority plus a 48-hour escalation path. - I would facilitate a workshop using a limit increase, decline exception, and evidence change, then document the sponsor's ratified model within 10 business days.

A regulated lending change spans 5 departments, 220 requirements, 4 regulations, and an audit in 10 weeks. What traceability depth do you mandate, and would you delay development until the matrix is complete?

No regulated story should enter delivery without end-to-end traceability; unrelated low-risk work can continue. - Each obligation would link to its legal source, policy rule, business requirement, acceptance test, control owner, and evidence location; ordinary usability requirements would stop at requirement-to-test links. - The first 2 weeks would complete the 60 requirements tied to customer harm or statutory reporting, with a weekly report of orphan obligations and unsupported tests. - A gap would remain open until Legal or Compliance confirms it is legally permissible, either a remediation owner and date or an exception approved by formally delegated authority is recorded, and the accountable control owner approves the treatment; documentation alone is not acceptance.

A $9 million insurance platform must serve 7 countries in 9 months, but a new regulation adds 35 mandatory requirements while each country requests 20 local enhancements. What enters the committed scope, and how do you defend that decision?

The committed scope should contain all 35 regulatory requirements and only the local features required for lawful day-one operation. - I would map every request to a legal obligation, core business outcome, or discretionary benefit and require a named source for any claim of mandatory scope. - Country enhancements would be ranked by quantified value and dependency, with the highest-value items filling capacity left after compliance and core processing. - I would show the steering group a scope ledger with deadline impact and reserve the remaining enhancements for country waves after the 9-month launch.

A benefits portal has 140 candidate requirements from 6 departments, but the $3 million release has only 14 weeks and capacity for about 65. How do you define the minimum viable requirements and choose what is deferred?

The minimum is the smallest compliant, supportable end-to-end service that delivers the target benefit. - I would start with the employee journey, mandatory policy rules, data inputs, exception paths, and operational controls rather than selecting isolated high-priority features. - Every requirement would score against benefit, legal risk, dependency, and effort, but an end-to-end process slice would outrank a collection of disconnected high scores. - I would baseline roughly 55 requirements and hold 10 as contingency, with explicit deferred outcomes and owners rather than a hidden overflow backlog.

An $8 million order-management program has 9 divisions, 160 requirements, and 3 releases over 11 months. Finance wants billing first, Sales wants quoting first, and Operations says fulfillment dependencies make both unsafe; what sequence do you choose?

The release sequence should follow the end-to-end dependency map, starting with shared product, customer, and order rules. - Release 1 would establish canonical master data and the minimum quote-to-order path for one representative division. - Release 2 would add fulfillment and billing together where financial events depend on shipment status, instead of creating temporary reconciliation work. - Release 3 would onboard the other eight divisions in waves, using the first division's acceptance evidence to remove local variants that lack business value.

Seven departments have approved 200 requirements for a $6 million program, then submit 24 changes 6 weeks before build completion; estimates show $900,000 and 10 weeks of impact. Which changes do you accept, and what governance do you invoke?

The change authority should accept only changes that prevent a breach, restore usability, or create more value than their full delay cost. - Each request would include the changed outcome, source, affected requirements, test impact, estimate, and cost of waiting; preference changes would not reach the board. - I would group dependent requests so the change authority sees three or four coherent options rather than voting on 24 fragments. - The sponsor would choose scope, budget, or date explicitly, and I would issue a new baseline within 2 business days of that decision.

Five divisions describe the same $15 million customer transformation through 320 conflicting process, data, policy, and system requirements, with an investment decision due in 6 weeks. How do you structure the analysis, and what do you refuse to approve?

Before investment approval, I would organize the 320 requirements by outcomes, capabilities, value streams, rules, information, and solution needs. - A capability map would expose duplicate requests, while BPMN flows and a business glossary would reveal where divisions use different words or processes for the same outcome. - I would trace solution features upward to outcomes and downward to affected rules, data, and controls, keeping technology design outside the BA-owned decision. - I would recommend that the investment owner withhold approval from capabilities with no measurable outcome, no decision owner, or unresolved cross-division dependency by week 6.

A shared-services team claims a $2.5 million workflow will save $4 million annually across 6 divisions, but no one agrees on the current cost of handling 1.2 million cases. You have 4 weeks before the funding vote; what baseline do you approve?

The funding vote should be deferred unless Finance and the process owners approve a reconciled baseline covering every material cost category. - The baseline would combine volume, touch time, labor cost, error rework, and service levels from representative samples across all six divisions, reconciled to payroll, vendor, and operational totals. - The model would separate avoidable cost from fixed overhead, because releasing 40 FTE of effort creates cash savings only when capacity is removed or credibly redeployed. - I would disclose evidence coverage and confidence by cost category; if material costs remain weak, the decision is a deferral or an explicitly conditional vote with no committed savings, not approval of an unsupported case.

A claims automation option costs $5 million over 3 years and promises $8 million of benefit across 4 departments, while a $1.4 million process change promises $3.5 million. The board meets in 12 days; which cost and benefit categories do you include before recommending one?

Both options need a like-for-like comparison of incremental cash flows and operational outcomes over three years. - Costs would include implementation, licenses, integration, data cleanup, training, transition productivity loss, and ongoing support, not just vendor price. - Benefits would separate cashable headcount or vendor savings from capacity, error reduction, leakage prevented, and faster settlement. - I would recommend the process option if its lower risk and faster realization remain stronger after dependencies and benefit confidence are priced, even if its headline benefit is smaller.

A $7 million compliance platform returns $2.4 million per year for 5 years, uses an 11% discount rate, and loses half its benefit if adoption stays below 70%. With approval due Friday, do you recommend full funding, a pilot, or rejection?

My recommendation is staged funding after the case prices the pilot and benefit delay; without those inputs by Friday, defer. - The base-case NPV is about $1.87 million and discounted payback is about 3.7 years: the 5-year annuity is worth about $8.87 million at 11% against the $7 million cost. - At half the benefit, present value falls to about $4.44 million and NPV to about -$2.56 million, so adoption can reverse the decision. - The gate would model a 12-week pilot capped at $700,000 as incremental unless Finance confirms it is included in the $7 million, shift benefit timing accordingly, and require at least 75% active use plus evidenced control-time reduction before the sponsor releases full funding.

A 5-division procurement process can be improved by building for $6 million in 14 months, buying a package for $3.8 million in 8 months with a 25% fit gap, or changing policy for $900,000 in 4 months. What option do you recommend at the 3-week options review?

I would change the policy first unless the remaining gap still blocks the measured outcome. - I would compare all three against the same capabilities, control requirements, 5-year total cost, implementation risk, and time to first benefit. - The package's 25% gap would be split into differentiating needs, legal needs, and local preferences; only the first two justify configuration or extensions. - I would preserve build as a later option only if the policy change and package cannot meet a quantified requirement worth more than their extra cost and delay.

A $10 million service program forecasts $3.2 million annual savings across 4 regions, but Finance, Operations, and HR each say another function owns realization. Funding closes in 2 weeks; who do you make the benefits owner, and what must they sign?

The benefit owner must be the business executive who controls the affected process, budget, or capacity, appointed with the program sponsor rather than by the BA. - The relevant Operations executive would own handling-time and rework realization, while HR advises on workforce actions and Finance independently validates the baseline, formula, and cash treatment. - Each nominated owner would sign the target, date, source, dependencies, and action that converts released capacity into value, with one executive sponsor accountable for the total case. - I would recommend removing any benefit not accepted by a qualified business owner within 2 weeks rather than carrying it as assumed savings.

A 7-country onboarding case assumes 18% annual volume growth and 85% digital adoption, creating a claimed $6 million benefit, but current growth is 6% and two countries prohibit one digital step. With 10 days left, do you rework or reject the case?

I would rebaseline the case on current evidence and keep 18% growth as upside rather than reject the opportunity. - Country-level volumes and legal constraints would replace one blended model, with adoption capped where the prohibited step remains manual. - I would show which benefits survive at 6% growth, 60% adoption, and a 6-month delay, then identify the assumption that changes the decision. - If the base case no longer clears the investment threshold, I would recommend a two-country experiment rather than approve the $6 million claim.

A portfolio has $8 million available: a pricing initiative needs $5 million for a likely $9 million NPV, a service redesign needs $4 million for $6 million NPV, and a mandatory control change needs $3 million with no revenue. Six divisions need an answer in 1 week; what do you fund?

Fund the mandatory control change and pricing initiative, and record the service redesign as the opportunity cost. - Mandatory does not mean unexamined, so I would confirm the minimum compliant scope and reserve only the evidenced $3 million. - Between discretionary options, pricing has the stronger NPV, but I would test confidence, resource overlap, and downside before committing the remaining $5 million. - The decision paper would show the foregone $6 million NPV and the trigger for reconsidering it if pricing confidence falls before contracting.

A 3-division automation pilot has spent $600,000 of a $4 million budget after 12 weeks; it reduced cycle time by 8% against a 30% target and exception rates rose from 4% to 11%. Do you stop, pivot, or scale at Monday's gate?

Stop scaling and allow only a short diagnostic pivot with a fixed cap and exit criteria. - The gate failed both the outcome and control measures, so sunk cost is not evidence for continuing the original design. - The bounded option would allow 4 weeks and no more than $150,000 to test whether two identified exception causes can restore the rate below 5% and lift cycle-time reduction above 20%. - If either threshold fails, I would recommend closing the option and documenting reusable process findings instead of consuming the remaining $3.4 million.

Three workstreams in a $20 million program each claim the same $2.5 million reduction in customer contacts, and the board expects a corrected case in 5 days. How do you allocate the benefit and decide whether the program still proceeds?

The $2.5 million contact reduction belongs once at outcome level; each workstream gets contribution measures, not duplicate cash. - A benefit dependency map would show how self-service, simpler policy, and better data jointly remove the same contacts. - The program sponsor would name one accountable owner for the $2.5 million, while I would use leading measures for each workstream, such as digital completion, avoidable-contact rate, and data accuracy. - I would recommend proceeding only if the de-duplicated case still clears the investment hurdle; otherwise the sponsor should remove or resequence the weakest workstream.

Order-to-cash crosses 6 departments, handles 40,000 orders a month, takes 21 days, and locks up $3 million in working capital; leadership wants a redesign proposal in 6 weeks. Where do you set the process boundaries and what do you change first?

The process boundary should run from accepted customer order through cleared cash. - I would map queue time, touch time, rework, decisions, systems, and controls across all six departments using a representative sample of order types. - The first change would target the constraint with the largest delay contribution, not the loudest team, and would include upstream causes of its rework. - The target process would have an owner for the full 21-day flow and measures for cycle time, first-pass yield, overdue value, and exceptions.

A 5-department service processes 60,000 requests a month in weekly batches, creating 4.5 days of waiting despite only 18 minutes of touch time. Which Lean change do you recommend for a 10-week pilot, and what would you measure?

Pilot smaller daily pull batches with a clear WIP limit, while retaining every current approval. - Daily intake would pull only the volume each downstream team can finish, with queue-age triggers exposing blocked requests instead of hiding them until the next weekly batch. - The 10-week pilot would compare median and 95th-percentile queue age, weekly throughput, touch time, and defect or rework rate against the current weekly-batch baseline. - If high queue age persists, I would use bottleneck data to adjust WIP or capacity; approval removal would remain a separate control decision for the control owner and Compliance.

A billing process across 4 teams has 92% first-pass yield, producing 9,600 defects and $1.8 million rework each year. You have 8 weeks for a Six Sigma recommendation; do you automate inspection or eliminate the main defect cause?

Remove the verified defect cause before automating inspection; faster detection does not improve first-pass yield. - I would stratify defects by product, channel, team, and rule, then validate the dominant cause with samples rather than relying on workshop opinion. - If one ambiguous pricing rule creates most defects, the change would standardize that rule and prevent invalid combinations at entry. - A control chart would track yield and defect mix for 8 weeks, with automation reconsidered only for residual variation that cannot be designed out.

Business Analyst interview questions

100 real questions with model answers and explanations for Senior Business Analyst candidates.

See a Business Analyst resume example →

Practice with flashcards

Spaced repetition · Hunter Pass

Questions

requirements

Discovery should be federated by division, with one central requirements architecture and approval gate.

I would assign a domain BA to each division and give all eight the same capability map, requirement hierarchy, and definition of ready.
A central team would own cross-division rules, shared data definitions, and the 25 or so interfaces where local decisions can create rework.
By week 4 I would baseline the top-level scope, then use weekly dependency reviews and a decision log to reach approval by week 12.

Why interviewers ask this: The interviewer is testing whether the candidate can scale elicitation without losing enterprise consistency or decision speed.

conflictownership

Credit Risk should own policy and declines, Operations should own execution, and Sales should have input without override authority.

Any exception to a decline would require formally delegated Credit Risk authority; a salesperson or revenue owner could request an exception but never approve it.
The rights matrix would name one approver for each policy, limit, evidence, and exception decision, with segregation between revenue ownership and credit authority plus a 48-hour escalation path.
I would facilitate a workshop using a limit increase, decline exception, and evidence change, then document the sponsor's ratified model within 10 business days.

Why interviewers ask this: The interviewer is evaluating whether the candidate creates explicit decision rights while preserving independent credit authority and segregation from revenue ownership.

requirementsrequirements-traceability

No regulated story should enter delivery without end-to-end traceability; unrelated low-risk work can continue.

Each obligation would link to its legal source, policy rule, business requirement, acceptance test, control owner, and evidence location; ordinary usability requirements would stop at requirement-to-test links.
The first 2 weeks would complete the 60 requirements tied to customer harm or statutory reporting, with a weekly report of orphan obligations and unsupported tests.
A gap would remain open until Legal or Compliance confirms it is legally permissible, either a remediation owner and date or an exception approved by formally delegated authority is recorded, and the accountable control owner approves the treatment; documentation alone is not acceptance.

Why interviewers ask this: The interviewer wants evidence that the candidate applies risk-based traceability without treating a documented regulatory gap as automatically acceptable.

requirements

The committed scope should contain all 35 regulatory requirements and only the local features required for lawful day-one operation.

I would map every request to a legal obligation, core business outcome, or discretionary benefit and require a named source for any claim of mandatory scope.
Country enhancements would be ranked by quantified value and dependency, with the highest-value items filling capacity left after compliance and core processing.
I would show the steering group a scope ledger with deadline impact and reserve the remaining enhancements for country waves after the 9-month launch.

Why interviewers ask this: The interviewer is checking whether the candidate can separate mandatory scope from stakeholder preference and make the trade-off auditable.

mvprequirementscapacity

The minimum is the smallest compliant, supportable end-to-end service that delivers the target benefit.

I would start with the employee journey, mandatory policy rules, data inputs, exception paths, and operational controls rather than selecting isolated high-priority features.
Every requirement would score against benefit, legal risk, dependency, and effort, but an end-to-end process slice would outrank a collection of disconnected high scores.
I would baseline roughly 55 requirements and hold 10 as contingency, with explicit deferred outcomes and owners rather than a hidden overflow backlog.

Why interviewers ask this: The interviewer is testing whether the candidate defines minimum viable scope as an operable business outcome rather than a numerical cut of the backlog.

requirementsdependencies

The release sequence should follow the end-to-end dependency map, starting with shared product, customer, and order rules.

Release 1 would establish canonical master data and the minimum quote-to-order path for one representative division.
Release 2 would add fulfillment and billing together where financial events depend on shipment status, instead of creating temporary reconciliation work.
Release 3 would onboard the other eight divisions in waves, using the first division's acceptance evidence to remove local variants that lack business value.

Why interviewers ask this: The interviewer is evaluating whether the candidate sequences requirements by business and data dependencies instead of stakeholder rank.

requirementsestimation

The change authority should accept only changes that prevent a breach, restore usability, or create more value than their full delay cost.

Each request would include the changed outcome, source, affected requirements, test impact, estimate, and cost of waiting; preference changes would not reach the board.
I would group dependent requests so the change authority sees three or four coherent options rather than voting on 24 fragments.
The sponsor would choose scope, budget, or date explicitly, and I would issue a new baseline within 2 business days of that decision.

Why interviewers ask this: The interviewer is checking whether the candidate makes late change control evidence-based and forces sponsors to own the resulting constraint trade-off.

requirementssystem-designconcurrency

Before investment approval, I would organize the 320 requirements by outcomes, capabilities, value streams, rules, information, and solution needs.

A capability map would expose duplicate requests, while BPMN flows and a business glossary would reveal where divisions use different words or processes for the same outcome.
I would trace solution features upward to outcomes and downward to affected rules, data, and controls, keeping technology design outside the BA-owned decision.
I would recommend that the investment owner withhold approval from capabilities with no measurable outcome, no decision owner, or unresolved cross-division dependency by week 6.

Why interviewers ask this: The interviewer wants to see an analysis architecture that reduces complexity without drifting into ownership of enterprise technology architecture.

workflow-analysis

The funding vote should be deferred unless Finance and the process owners approve a reconciled baseline covering every material cost category.

The baseline would combine volume, touch time, labor cost, error rework, and service levels from representative samples across all six divisions, reconciled to payroll, vendor, and operational totals.
The model would separate avoidable cost from fixed overhead, because releasing 40 FTE of effort creates cash savings only when capacity is removed or credibly redeployed.
I would disclose evidence coverage and confidence by cost category; if material costs remain weak, the decision is a deferral or an explicitly conditional vote with no committed savings, not approval of an unsupported case.

Why interviewers ask this: The interviewer is testing whether the candidate protects the funding decision when the baseline lacks sufficient evidence for material costs.

promisesconcurrencycost-benefit-analysis

Both options need a like-for-like comparison of incremental cash flows and operational outcomes over three years.

Costs would include implementation, licenses, integration, data cleanup, training, transition productivity loss, and ongoing support, not just vendor price.
Benefits would separate cashable headcount or vendor savings from capacity, error reduction, leakage prevented, and faster settlement.
I would recommend the process option if its lower risk and faster realization remain stronger after dependencies and benefit confidence are priced, even if its headline benefit is smaller.

Why interviewers ask this: The interviewer is evaluating whether the candidate builds comparable options and distinguishes real cash from non-cash capacity.

decision-making

My recommendation is staged funding after the case prices the pilot and benefit delay; without those inputs by Friday, defer.

The base-case NPV is about $1.87 million and discounted payback is about 3.7 years: the 5-year annuity is worth about $8.87 million at 11% against the $7 million cost.
At half the benefit, present value falls to about $4.44 million and NPV to about -$2.56 million, so adoption can reverse the decision.
The gate would model a 12-week pilot capped at $700,000 as incremental unless Finance confirms it is included in the $7 million, shift benefit timing accordingly, and require at least 75% active use plus evidenced control-time reduction before the sponsor releases full funding.

Why interviewers ask this: The interviewer is checking whether the candidate calculates the economics correctly and makes pilot cost, delay, and adoption an explicit funding gate.

procurementconcurrency

I would change the policy first unless the remaining gap still blocks the measured outcome.

I would compare all three against the same capabilities, control requirements, 5-year total cost, implementation risk, and time to first benefit.
The package's 25% gap would be split into differentiating needs, legal needs, and local preferences; only the first two justify configuration or extensions.
I would preserve build as a later option only if the policy change and package cannot meet a quantified requirement worth more than their extra cost and delay.

Why interviewers ask this: The interviewer wants to see that the candidate includes process change as a real option and avoids defaulting to technology.

benefits-ownership

The benefit owner must be the business executive who controls the affected process, budget, or capacity, appointed with the program sponsor rather than by the BA.

The relevant Operations executive would own handling-time and rework realization, while HR advises on workforce actions and Finance independently validates the baseline, formula, and cash treatment.
Each nominated owner would sign the target, date, source, dependencies, and action that converts released capacity into value, with one executive sponsor accountable for the total case.
I would recommend removing any benefit not accepted by a qualified business owner within 2 weeks rather than carrying it as assumed savings.

Why interviewers ask this: The interviewer is testing whether the candidate facilitates accountable benefit ownership without assigning it beyond the BA's authority.

onboardingdecision-making

I would rebaseline the case on current evidence and keep 18% growth as upside rather than reject the opportunity.

Country-level volumes and legal constraints would replace one blended model, with adoption capped where the prohibited step remains manual.
I would show which benefits survive at 6% growth, 60% adoption, and a 6-month delay, then identify the assumption that changes the decision.
If the base case no longer clears the investment threshold, I would recommend a two-country experiment rather than approve the $6 million claim.

Why interviewers ask this: The interviewer is evaluating whether the candidate challenges optimistic assumptions and produces a decision-ready alternative under deadline.

pricing

Fund the mandatory control change and pricing initiative, and record the service redesign as the opportunity cost.

Mandatory does not mean unexamined, so I would confirm the minimum compliant scope and reserve only the evidenced $3 million.
Between discretionary options, pricing has the stronger NPV, but I would test confidence, resource overlap, and downside before committing the remaining $5 million.
The decision paper would show the foregone $6 million NPV and the trigger for reconsidering it if pricing confidence falls before contracting.

Why interviewers ask this: The interviewer is checking whether the candidate makes constrained portfolio choices explicit, including what value is deliberately forgone.

error-handling

Stop scaling and allow only a short diagnostic pivot with a fixed cap and exit criteria.

The gate failed both the outcome and control measures, so sunk cost is not evidence for continuing the original design.
The bounded option would allow 4 weeks and no more than $150,000 to test whether two identified exception causes can restore the rate below 5% and lift cycle-time reduction above 20%.
If either threshold fails, I would recommend closing the option and documenting reusable process findings instead of consuming the remaining $3.4 million.

Why interviewers ask this: The interviewer is testing whether the candidate can enforce stop gates despite sunk-cost pressure and still preserve a bounded learning option.

program-management

The $2.5 million contact reduction belongs once at outcome level; each workstream gets contribution measures, not duplicate cash.

A benefit dependency map would show how self-service, simpler policy, and better data jointly remove the same contacts.
The program sponsor would name one accountable owner for the $2.5 million, while I would use leading measures for each workstream, such as digital completion, avoidable-contact rate, and data accuracy.
I would recommend proceeding only if the de-duplicated case still clears the investment hurdle; otherwise the sponsor should remove or resequence the weakest workstream.

Why interviewers ask this: The interviewer is evaluating whether the candidate detects benefit overlap and is willing to change scope when the corrected economics weaken.

concurrency

The process boundary should run from accepted customer order through cleared cash.

I would map queue time, touch time, rework, decisions, systems, and controls across all six departments using a representative sample of order types.
The first change would target the constraint with the largest delay contribution, not the loudest team, and would include upstream causes of its rework.
The target process would have an owner for the full 21-day flow and measures for cycle time, first-pass yield, overdue value, and exceptions.

Why interviewers ask this: The interviewer wants to see end-to-end value-stream thinking tied to a financial outcome rather than local optimization.

batchconcurrency

Pilot smaller daily pull batches with a clear WIP limit, while retaining every current approval.

Daily intake would pull only the volume each downstream team can finish, with queue-age triggers exposing blocked requests instead of hiding them until the next weekly batch.
The 10-week pilot would compare median and 95th-percentile queue age, weekly throughput, touch time, and defect or rework rate against the current weekly-batch baseline.
If high queue age persists, I would use bottleneck data to adjust WIP or capacity; approval removal would remain a separate control decision for the control owner and Compliance.

Why interviewers ask this: The interviewer is checking whether the candidate improves flow with smaller batches and WIP limits without casually removing approvals.

defectsconcurrency

Remove the verified defect cause before automating inspection; faster detection does not improve first-pass yield.

I would stratify defects by product, channel, team, and rule, then validate the dominant cause with samples rather than relying on workshop opinion.
If one ambiguous pricing rule creates most defects, the change would standardize that rule and prevent invalid combinations at entry.
A control chart would track yield and defect mix for 8 weeks, with automation reconsidered only for residual variation that cannot be designed out.

Why interviewers ask this: The interviewer is evaluating whether the candidate applies defect evidence to root-cause removal instead of automating waste.

Locked questions

21
A regulated refund process spans 5 teams and must fall from 10 days to 4 days before a 6-month contract renewal, but removing dual approval may expose $700,000 of annual fraud. What process decision do you make?
concurrency
22
A contact-center process across 3 sites receives 18,000 cases weekly, has capacity for 16,000, and faces $500,000 in penalties if the backlog is not cleared in 8 weeks. Do you add staff, automate triage, or remove demand?
backlogcapacityconcurrency
23
A $6 million target operating process must support 7 regions in 9 months; four regions want local processing, while three support a shared center that is 22% cheaper but adds a one-day handoff. Which model do you choose?
concurrency
24
A redesigned claims process affects 1,200 employees in 5 departments, must launch in 10 weeks, and needs 80% adoption to realize $2 million of benefit. Do you choose a big-bang rollout or waves, and what adoption evidence controls the decision?
decision-makingconcurrency
25
A bank wants to remove 4 of 7 handoffs from a lending process handling $200 million annually, but Compliance requires segregation of duties and the redesign is due in 5 weeks. Which handoffs do you remove without weakening control?
concurrency
26
Process-mining logs contain 2 million purchase-to-pay events across 6 departments and reveal 47 variants; the standard path takes 6 days while the largest exception path takes 19 days and costs $900,000 yearly. Which variant do you standardize first?
concurrencyerror-handlingprocess-mining
27
A 4-department fulfillment process must handle a 35% seasonal peak in 12 weeks while keeping 95% of orders within 48 hours; cross-training costs $300,000 and automation costs $1.2 million. Which capacity option do you choose?
capacityconcurrency
28
Six divisions use four definitions of active customer across 11 systems, changing a $5 million retention case by up to 30%; the board needs one figure in 4 weeks. Which canonical term do you choose and who owns it?
business-glossarysystem-designretention
29
Five business units report 18% more ticket closures, but reopen rates rose from 9% to 16% and CSAT fell 7 points before a retention review in 3 weeks. What KPI tree do you recommend, and who governs it?
retentionkpi-definitionclosures
30
A 9-month customer migration spans 12 systems, 7 departments, 25 million records, and a $14 million budget. What conceptual and logical models do you require before approving build, and where do you stop short of physical design?
system-designdesignmigrations
31
An MDM initiative finds 14% duplicates in 8 million supplier records across 6 systems; false merges could block payments, but unresolved duplicates cost $1.8 million yearly. With an 8-week design deadline, what golden-record and match decision do you make?
system-designdesignestimation
32
Four finance systems disagree by $6 million at monthly close, 3 divisions use different adjustment rules, and the close must fall from 8 days to 5 within one quarter. Which source and reconciliation tolerance do you approve?
conflictsystem-designreact
33
A customer-data program covers 7 countries and 11 systems; laws require retention between 2 and 10 years, while Marketing wants all history for a $4 million analytics case due in 6 weeks. What lineage and retention requirements do you approve?
system-designlineageretention
34
Operations claims a rule change will recover $4 million of annual revenue across 5 product lines, but the evidence sits in 30 million order rows and scope freezes in 9 days. What SQL analysis do you run, and what decision threshold do you set?
sql
35
A migration has 12 million policy records from 5 systems, 22% missing a standardized status, and only 7 weeks before dress rehearsal; manual cleansing costs $900,000. What Python evidence do you produce, and do you cleanse or transform?
migrationssystem-designpython
36
A $16 million migration will move 25 million customer records from 9 systems in 8 weeks; the vendor proposes 99% row-count accuracy, but 1% could mean 250,000 bad records. What acceptance criteria determine go or no-go?
specsmigrationssystem-design
37
Six delivery teams have produced 240 requirements, but none shows coverage of the program's 14 customer journeys. How do you build the coverage view and decide when analysis is complete?
requirementscoverage
38
Nine teams must deliver 14 business interfaces for a $10 million launch in 7 months, but 5 interfaces still have conflicting ownership and data rules. Do you freeze contracts now or allow continued discovery, and how do you sequence the dependencies?
ownershiptypesdependencies
39
Five product teams see 30% of requirements change every month, causing an estimated $700,000 of quarterly rework, while a regulatory release is fixed in 4 months. What do you freeze, and what remains open?
requirementsestimation
40
A vendor package covers 750 requirements for 8 departments, costs $4 million, and must go live in 9 months; the vendor reports 88% fit, while departments call 180 gaps mandatory. Which gaps justify configuration, extension, or process change?
requirementsconfigconcurrency
41
A $11 million customer product must launch in 8 countries within 6 months; 70% of 210 requirements are common, while tax, language, consent, and document rules differ. Do you define one global baseline or eight local backlogs?
backlogrequirementsbacklog-management
42
A service used by 6 divisions handles $80 million daily; 99.9% availability costs $2 million, 99.99% costs $4.5 million, and each hour of peak outage costs about $300,000. Which nonfunctional target do you recommend before contracts close in 2 weeks?
43
A 5-department platform supports $25 million of weekly transactions; a 1-hour RTO and near-zero RPO cost $3 million more than a 12-hour RTO and 4-hour RPO, and requirements must close in 3 weeks. What continuity targets do you choose?
requirementstransactions
44
Eighteen analysts across 12 projects use incompatible requirement styles, and audits find 23% lack acceptance evidence; you have 6 weeks and $80,000 to set a standard. What do you mandate, and what do you leave flexible?
45
Eighteen analysts produce about 900 requirements a month across 10 teams, but a central reviewer can inspect only 120 and escaped requirement defects cost $600,000 per quarter. Do you choose central review, peer review, or risk-based sampling before next month's release?
samplingdefectsrequirements
46
Ten delivery teams spend a median 5 days producing requirement documents, while leadership wants 2 days without raising the current 12% rework rate; you have 4 weeks to choose reusable templates. What survives and what do you remove?
47
Six analysts on a $7 million program have a 18% requirement-defect rate, mainly ambiguous rules and missed exceptions; you have 90 days and no authority to replace them. What mentoring intervention do you choose, and what result must it deliver?
mentoringinfluenceerror-handling
48
Seven division heads disagree on a $10 million scope, 45 requirements remain contested, and the executive steering meeting is tomorrow morning. What decision do you ask for, and how do you communicate it in 10 minutes?
conflictcommunicationrequirements
49
A portfolio contains 14 initiatives worth $60 million, 5 begin procurement in 3 weeks, and requirement quality ranges from 45% to 90% against the current checklist. Which initiatives pass the gate, and how do you avoid rewarding document volume?
procurement
50
Eighteen analysts across 9 programs split requirements among Jira, Azure DevOps, spreadsheets, and documents; an audit starts in 8 weeks, and only $120,000 is available. Do you mandate one tool now or standardize the repository rules first?
requirementsspreadspreadsheets
51
The CFO wants to cut $3 million from a warehouse transformation, while the COO says all 46 sites must adopt it before peak season in 12 weeks. The current plan costs $5 million. What do you do in the next 48 hours?
decision-makingwarehouse
52
Legal says a new lending journey must show a specific adverse-action reason, but Product wants an instant one-screen decision and the model vendor returns only a score. Launch is in 10 weeks and $8 million of revenue is forecast. How do you resolve it?
procurement
53
Eleven senior stakeholders disagree on the priority of 24 pricing rules, three delivery teams are blocked, and an $18 million renewal needs an auditable quote in nine weeks. How do you break the deadlock?
stakeholder-managementconflictcommunication
54
Procurement for a $6 million automation program starts in 10 days, but the accountable benefit owner withdraws the claimed $4 million annual saving after discovering only $1.3 million is cashable. What do you do?
procurement
55
An ERP program has spent $7.6 million of a $9 million budget, delivered 62% of critical scope, and has 11 weeks left; the CEO insists on the original scope and date. What recommendation do you make?
56
A refund policy affects 14 markets and 12,000 cases per month, but Legal, Finance, and Operations each deny ownership while developers have already built three different interpretations. What do you do this week?
ownership
57
Forty-eight hours before a board meeting, the CFO reports 31% product margin and the COO reports 24% for the same quarter; a $20 million investment decision uses that metric. How do you handle it?
soft-skillsmonitoringcss
58
Sales has promised a $6 million customer a custom order workflow in six weeks, but Product says it would fork the platform and add $800,000 annual support cost. What decision process do you lead?
promisesconcurrencyworkflow-analysis
59
The board orders $4 million of annual savings, but five functional heads each label all of their processes critical and the budget locks in three weeks. How do you get to a defensible decision?
concurrency
60
UAT has 327 cases, but only 214 have run with four days left; 18 failed and 5 are severity 1. The sponsor wants sign-off to protect a $2 million launch campaign. What do you do?
severity-priorityuatsponsor
61
Seventy-two hours before launch, 2 of 12 critical refund journeys double-charge customers in the release candidate; a $1.2 million campaign is already booked. What is your recommendation?
62
A migration rehearsal for 1.8 million accounts shows target balances differ from the source on 2.7% of records, while the accepted limit is 0.05% and cutover is six days away. Do you proceed?
migrations
63
The approved performance requirement is a 2-second p95 response for 25,000 concurrent users, but the final test shows 4.8 seconds at 18,000 users. Launch is in eight days. What do you do?
concurrency
64
After 410 UAT cases pass, you discover the environment used stale data and bypassed the tax service that handles 28% of production orders. Sign-off is due tomorrow. What do you do?
uat
65
Operations signs a release, but Compliance rejects 7 of 80 controls and Sales says a delay risks $9 million in contracts. The launch authority is unclear. How do you reach a go/no-go decision?
governance
66
A package vendor proposes a three-month CSV workaround for a failed tax integration handling 12,000 transactions a day; Operations estimates it needs 14 people and creates a one-day posting delay. Launch is in two weeks. What do you decide?
procurementtransactionsestimation
67
Five days before a holiday launch backed by $3.5 million in media, testing finds 0.7% of orders calculate tax incorrectly; forecast volume is 350,000 orders and refunds must be manual. Do you launch?
testing
68
The final cutover rehearsal takes 17 hours against an 8-hour outage window, rollback has never been tested, and a regulatory deadline is in 10 days with a $900,000 penalty. What do you recommend?
estimationrollback
69
A claims redesign promised a 40% cycle-time reduction, but 12 weeks after rollout average time moved from 5.2 to 5.0 days across 60,000 claims. The sponsor wants expansion to three more units. What do you do?
sponsorpromises
70
Invoice automation raises straight-through processing from 54% to 72%, but exception rates also rise from 6% to 19% across 80,000 monthly invoices and month-end close slips two days. How do you recover?
concurrencyerror-handling
71
Automation reduced cycle time but skipped the mandatory accessibility and manual-assistance route for 9% of vulnerable customers. What do you do?
a11yvulnerabilities
72
A new intake process cuts front-office handling from three days to one, but quality review backlog grows by 11 days and 2,000 cases are waiting. Leadership still reports the initiative as green. What do you do?
backlogconcurrencybacklog-management
73
Eight weeks after a $2 million case-management rollout, only 18% of 1,200 agents use it weekly and average handle time has not changed. The sponsor requests another $700,000 for features. What do you recommend?
sponsor
74
A $4.5 million automation case assumed 35 FTE could be removed, but a time study after approval shows only 8 FTE of addressable work. Procurement wants to sign the vendor this week. What do you do?
procurement
75
Translated consent wording in 6 of 9 countries omits the mandatory purpose, and 38% of pilot consents in those countries are invalid. What do you do?
76
A seven-workstream transformation is nine months in, has spent $7 million, completed 46% of milestones, and realized 12% of a 30% benefit target. The sponsor asks for a recovery plan in 72 hours. What do you produce?
milestonesprogram-managementrecovery
77
An order-to-cash redesign reports 15% lower operating cost, but disputes double, days sales outstanding rises by 9 days, and working capital worsens by $11 million. Should rollout continue?
78
The night before a board meeting, you find a many-to-many join has overstated quarterly revenue by $18 million in the Power BI dashboard. The board pack has already been circulated. What do you do?
joinsbi
79
Two days before bonuses close, Growth changes the active-customer definition from 30 to 90 days, raising the KPI by 18% and moving three teams above target. How do you respond?
kpi-definition
80
Forty-eight hours before a campaign to 2.1 million customers, profiling finds 7.4% duplicate master records and conflicting consent on 31,000 pairs. What do you do?
profiling
81
Three days before cutover, a 14-million-row migration has matching row and control totals, but status-date distributions shift 18% of records to default values. How do you proceed?
distributionsmigrations
82
An SQL analysis claims churn fell 12% after an app launch, but its sample includes only app users, who are 41% of customers; the board is considering another $6 million investment. What do you do?
churnsql
83
Consent withdrawals recorded in CRM fail to reach campaign tools in 4 systems for 31,000 customers. What do you do?
system-design
84
A program forecast $10 million in annual savings, but first-quarter actuals are $800,000 against a $2.5 million plan; Budget wants to book the full-year benefit next week. What do you recommend?
85
Service teams report 96% SLA compliance, but 27% of tickets are closed and reopened within 48 hours and CSAT has fallen 14 points; bonuses use the SLA. What do you do?
86
A regulatory filing is due in 72 hours, but two source systems disagree on the delinquent-account population by 4,300 records out of 18,000 because they use different snapshot times. What do you do?
conflictsystem-designsnapshot
87
A regulator's FAQ conflicts with the statute interpretation 60 days before a deadline across four products. How do you lead the response?
estimation
88
An external audit starts in 12 days and finds 31 of 220 key controls have no traceable owner or evidence; a qualified opinion could breach lending covenants. What do you do?
89
After an acquisition, one company defines customer as a legal entity and the other as a household; three systems and 6 million records must support day-one billing in eight weeks. How do you handle the conflict?
soft-skillssystem-design
90
A vendor upgrade in 5 months deprecates 12 configured ERP workflows used by 4 business units. How do you assess the options and decide the rollout wave?
workflow-analysisconfigprocurement
91
A vendor contract promises to resolve business-critical defects promptly, but a production outage is costing $120,000 per day and the vendor says this incident is excluded. What do you do in the first 24 hours?
procurementpromisesincidents
92
An outsourced analysis team delivers 480 requirements ten days before fixed-price build starts, but 37% are not testable and 22 pairs conflict. The supplier says clarification is out of scope. What do you do?
requirements
93
A regulator requests end-to-end complaint-resolution evidence for 500 sampled cases in 7 days, but event timestamps and owners are missing in 19% of them. How do you organize the response?
e2e
94
A BA on your team wrote that approvals must be fast; the feature went live six hours ago with 9-second p95 latency and abandonment has risen from 6% to 14%. What do you do?
latency
95
Two senior BAs disagree on BPMN versus use cases, and reviews for 260 stories have stalled. How do you resolve the method dispute?
use-casesconflictprocess-modeling
96
A developing BA must refine 90 regulatory requirements in five days while two other workstreams need your analysis for release decisions. How do you mentor without missing the deadline?
mentoringestimationprogram-management
97
Six product teams reject a new requirements standard because it adds an estimated 30 minutes per story, while escaped requirement defects are running at 11% and caused $400,000 of rework last quarter. What do you do?
requirementsestimationdefects
98
Four weeks before a launch tied to a $15 million contract, an executive asks you to remove 28 unresolved critical requirements from the steering deck because the customer is nervous. What do you do?
requirements
99
An $8 million program finished three months late after 41 scope changes and 23 UAT escapes; leaders blame delivery teams and want a two-hour retrospective next week. How do you lead it?
uat
100
The lead BA leaves unexpectedly three weeks before go-live, leaving 640 requirements, 27 open decisions, four analysts, and no current handover. What do you do in the first 48 hours?
requirements

Questions

1An ERP program has 8 divisions, 180 high-level requirements, a $12 million budget, and 12 weeks to approve scope. Would you run requirements centrally or federate the analysis, and what would you put in place first?

2Six divisions disagree on who may approve customer-credit rules for a $4 million platform due to start build in 5 weeks. Sales wants revenue ownership, Risk wants veto power, and Operations owns execution; what decision-rights model do you choose?

3A regulated lending change spans 5 departments, 220 requirements, 4 regulations, and an audit in 10 weeks. What traceability depth do you mandate, and would you delay development until the matrix is complete?

4A $9 million insurance platform must serve 7 countries in 9 months, but a new regulation adds 35 mandatory requirements while each country requests 20 local enhancements. What enters the committed scope, and how do you defend that decision?

5A benefits portal has 140 candidate requirements from 6 departments, but the $3 million release has only 14 weeks and capacity for about 65. How do you define the minimum viable requirements and choose what is deferred?

6An $8 million order-management program has 9 divisions, 160 requirements, and 3 releases over 11 months. Finance wants billing first, Sales wants quoting first, and Operations says fulfillment dependencies make both unsafe; what sequence do you choose?

7Seven departments have approved 200 requirements for a $6 million program, then submit 24 changes 6 weeks before build completion; estimates show $900,000 and 10 weeks of impact. Which changes do you accept, and what governance do you invoke?

8Five divisions describe the same $15 million customer transformation through 320 conflicting process, data, policy, and system requirements, with an investment decision due in 6 weeks. How do you structure the analysis, and what do you refuse to approve?

9A shared-services team claims a $2.5 million workflow will save $4 million annually across 6 divisions, but no one agrees on the current cost of handling 1.2 million cases. You have 4 weeks before the funding vote; what baseline do you approve?

10A claims automation option costs $5 million over 3 years and promises $8 million of benefit across 4 departments, while a $1.4 million process change promises $3.5 million. The board meets in 12 days; which cost and benefit categories do you include before recommending one?

11A $7 million compliance platform returns $2.4 million per year for 5 years, uses an 11% discount rate, and loses half its benefit if adoption stays below 70%. With approval due Friday, do you recommend full funding, a pilot, or rejection?

12A 5-division procurement process can be improved by building for $6 million in 14 months, buying a package for $3.8 million in 8 months with a 25% fit gap, or changing policy for $900,000 in 4 months. What option do you recommend at the 3-week options review?

13A $10 million service program forecasts $3.2 million annual savings across 4 regions, but Finance, Operations, and HR each say another function owns realization. Funding closes in 2 weeks; who do you make the benefits owner, and what must they sign?

14A 7-country onboarding case assumes 18% annual volume growth and 85% digital adoption, creating a claimed $6 million benefit, but current growth is 6% and two countries prohibit one digital step. With 10 days left, do you rework or reject the case?

15A portfolio has $8 million available: a pricing initiative needs $5 million for a likely $9 million NPV, a service redesign needs $4 million for $6 million NPV, and a mandatory control change needs $3 million with no revenue. Six divisions need an answer in 1 week; what do you fund?

16A 3-division automation pilot has spent $600,000 of a $4 million budget after 12 weeks; it reduced cycle time by 8% against a 30% target and exception rates rose from 4% to 11%. Do you stop, pivot, or scale at Monday's gate?

17Three workstreams in a $20 million program each claim the same $2.5 million reduction in customer contacts, and the board expects a corrected case in 5 days. How do you allocate the benefit and decide whether the program still proceeds?

18Order-to-cash crosses 6 departments, handles 40,000 orders a month, takes 21 days, and locks up $3 million in working capital; leadership wants a redesign proposal in 6 weeks. Where do you set the process boundaries and what do you change first?

19A 5-department service processes 60,000 requests a month in weekly batches, creating 4.5 days of waiting despite only 18 minutes of touch time. Which Lean change do you recommend for a 10-week pilot, and what would you measure?

20A billing process across 4 teams has 92% first-pass yield, producing 9,600 defects and $1.8 million rework each year. You have 8 weeks for a Six Sigma recommendation; do you automate inspection or eliminate the main defect cause?

Locked questions