Remediation Cannot Outrun Duplicate Creation
Meridian Retail Bank has 480,000 customer records in its CRM. Twelve percent are duplicates, 8% lack a postal code, and the CRM generates about 220 complaint tickets per month tied to mailings and service history. Data quality is remediated manually by two FTEs who correct roughly 1,800 records per month. New records arrive through five channels, none of which validate against existing customers; sales staff are measured on speed of record creation. The data quality dashboard shows 14 dimensions, all rated amber, and has not changed in three quarters. The COO asks why the dashboard is not improving and whether to hire four more remediation staff.
1. Where should the bank intervene first? Use the numbers to support your answer.
Read the worked answer
Fix the inflow before adding remediation capacity. Current remediation is 1,800 records per month, or 21,600 per year, against a stock of about 57,600 duplicates (12% of 480,000). If new duplicates are created faster than 1,800 per month, the backlog grows regardless of hiring. The highest-leverage points are the five intake channels and the sales incentive. Add real-time matching and validation at intake, make record quality part of the sales measure, and route only exceptions to remediation. Then prioritize the 8% missing postal codes, about 38,400 records, because they break downstream mailings and service matching and are cheap to fix at source.
Hiring four more remediators treats the symptom and locks in a manual process. Approve it only as a temporary bridge with a defined end date and a measured reduction in inflow.
2. The COO asks why the 14-dimension dashboard is all amber and unchanged. What is wrong with the measurement approach and what would you replace it with?
Read the worked answer
The dashboard measures broad dimensions without process ownership, decision thresholds, or a baseline anyone can act on. Amber on 14 dimensions gives no one a next action, so the score persists. Replace it with a small set of operational measures tied to process steps: duplicate creation rate per channel per week, postal-code completeness at point of entry, first-pass validation rate, remediation backlog and age, and complaint tickets per 1,000 mailings. Assign each measure an owner and a threshold that triggers a specific action.
Keep two or three outcome measures for executives, such as complaint rate, mailing waste cost, and regulatory exceptions, and report trend rather than color.
3. Build the business case for the intake fix versus hiring four remediators. What assumptions matter most?
Read the worked answer
Compare 12-month cost and risk. Four remediators cost salary plus overhead and add roughly 3,600 records per month of capacity, but if inflow exceeds that the backlog still grows, and they add no prevention. The intake fix has upfront cost for channel integration, matching tooling, training, and incentive change, with lower ongoing cost.
The decisive assumptions are the monthly new duplicate creation rate, cost per complaint ticket, mailing waste per bad address, remediation cost per record, and the speed of channel changes. Get the creation rate from a 30-day sample by channel; if it is below 1,800 per month, remediation can keep up and the fix can be sequenced later. If it is above, prevention wins on both cost and service. Present a range, not a point estimate.