MELLORCATalk to Mellorca
Impact

Data Quality & Customer Experience

When Customer Details Differ Between Applications

Inconsistent customer data across applications creates operational friction, undermines service quality, and leads to unreliable reporting.

Mellorca Impact2026-10-02

Overview

In many organisations, customer details live in multiple systems that were built at different times, for different purposes, or with different data models. When a single customer appears under slightly different names, addresses, identifiers or contact channels across these apps, the result is not merely a data anomaly. It is a visible, painful symptom: duplicated outreach, misdirected service requests, conflicting order histories, and reports that tell divergent stories about a customer's activity.

The business tolerates this situation because the underlying data landscape has grown organically, yet the cost of this tolerance compounds over time as the same customer interacts with marketing, sales and support in different contexts.

What it looks like in a real business

Take a Cape Town insurance brokerage with around sixty staff. Clients are captured in the CRM by the sales team, in the policy administration system by the operations team, and in the accounting package for debit orders. A client moves house and phones to update her address. The call centre agent changes it in the CRM. The renewal pack is then posted to the old address from the policy system, and the debit order confirmation goes to an email address she stopped using two years ago. Nobody made a careless mistake. Each team updated the record it owns. The business simply has three versions of the same person and no rule for which one wins.

This is the pattern described in why one customer exists differently in three systems, and it is common in firms that grew by adding a tool per department.

Operational consequences

Operational teams face a cascade of friction. Customer service agents may need to reconcile records manually, delaying response times and increasing the chance of human error. Marketing and sales risk sending inconsistent communications or duplicating outreach, eroding customer trust. Analysts struggle to stitch a single view of the customer, making it harder to diagnose issues, satisfy service-level agreements or measure the effectiveness of campaigns. In short, data fragmentation translates into friction in execution and in reporting.

Warning signs

  • Staff keep a personal list of "the correct" contact details for important clients.
  • Customers say they have already told you something, often more than once.
  • Customer counts differ between the CRM, the billing system and the marketing platform.
  • Returned post, bounced emails and failed debit orders are handled one at a time with no root-cause review.
  • Merging duplicates is a task someone does "when there is time".

The customer-facing side of this is covered in the pain of customers repeating information.

Financial and strategic implications

The financial footprint includes wasted labour, rework and missed opportunities. Time spent reconciling data is time not spent on value-add activities such as personalised customer experiences or strategic analytics. Reporting confidence erodes when stakeholders rely on conflicting subsets of data, which can slow decision-making and erode governance credibility. Over time, the cost of misalignment compounds, affecting margins, customer satisfaction and the organisation's ability to scale.

There is also a compliance dimension. Data protection law generally expects personal information to be accurate and kept up to date where the purpose requires it; section 16 of South Africa's POPIA requires reasonably practicable steps to keep personal information complete, accurate, not misleading and updated where necessary, and the UK regulator's guidance on the equivalent accuracy principle asks organisations to take reasonable steps to correct inaccurate data promptly and to record where information came from. Three conflicting addresses for one person make that harder to demonstrate.

Estimating the cost in your own business

Avoid borrowed industry averages. Instead, measure a sample month: how many records were corrected, how long each correction took, how many communications were sent to wrong details, and how many service tickets were caused by mismatched data. Multiply staff time by loaded hourly cost and add direct costs such as reprinting, courier fees or failed collection charges. Even a rough figure will show whether the problem is worth a project. How to calculate the real cost of a repetitive task sets out a simple method.

How to diagnose the problem

IBM's data quality guidance lists dimensions that make diagnosis concrete. For customer records, three matter most:

  • Uniqueness: how many duplicate customers exist within each system and across systems?
  • Consistency: do the same fields agree between systems for the same customer?
  • Accuracy: are the values correct against an agreed source of truth?

A practical exercise is to take fifty active customers and compare their name, registration or ID number, email, phone and address across every system that holds them. Record each mismatch and which system was right. The pattern of errors usually reveals where data enters badly and which system people actually trust. It also gives you a baseline, so that later you can show whether the fix is working rather than relying on anecdote.

Remediation actions

Addressing this pain begins with establishing a trusted, single source of truth for customer identity. Actions include implementing identity resolution, creating a golden record for core customer attributes, and standardising data definitions across systems. IBM describes master data management as consolidating customer details through deduplication, reconciliation, standardisation and integration, so that one authoritative version exists.

Technical steps involve data validation at entry points, regular deduplication, and parent-child relationship mapping across apps, such as linking branches to a head office account. Process-wise, align ownership for customer data, enforce governance policies, and embed data quality checks into both onboarding and ongoing operations. Practical diagnostics include mapping data lineage, identifying high-friction touchpoints, and validating critical attributes against a trusted reference.

Trade-offs

A full master data management platform suits larger organisations with many systems. For a smaller business, the more proportionate route is usually to choose one system of record for customer identity, push changes outward through integration, and make other systems read-only for those fields. The decision is discussed in how to decide which system owns which data. The risk of the lighter approach is that it depends on discipline: if staff can still edit addresses anywhere, drift returns.

What to do first

  1. Pick the customer attributes that cause the most damage when wrong, typically legal name, ID or registration number, billing email, and address.
  2. Agree which system owns each of those attributes and name a person accountable for it.
  3. Run a deduplication pass on that system before connecting anything else to it.
  4. Add validation at the point of capture: required fields, format checks and a duplicate warning.
  5. Integrate outward so that other systems receive updates instead of capturing them separately.
  6. Repeat the fifty-customer comparison quarterly to see whether consistency is improving.

Takeaways

The pain is real and measurable in everyday operations and in finance. A disciplined data quality programme, grounded in a single customer identity, consistent definitions and automated validation, turns this friction into a controllable risk and a source of competitive advantage. For organisations that want reliable reporting and good customer experiences, moving toward a unified customer data model is not optional. It is a foundational capability, and it also makes later automation and AI work far more dependable.

Sources and further reading

Related Mellorca reading

Article summaryInconsistent customer data across applications creates operational friction, undermines service quality, and leads to unreliable reporting.
Explore our solutionsStart a conversation