Hot Topics
- Home
- /
- Contact center data quality
Figure 1 – Comparison of how a call flows as recorded by the Switch and CTI of a contact center is shown above. The call traverses through IVR, waiting to get to an agent. Except the hold time, other durations are consistently recorded by both Switch and CTI. End of hold time is different in these systems leading to discrepancy in calculating total talk time for the call.
Contact center data quality
Introduction
Good data quality is a primary concern for contact center executives because accuracy and completeness of data is essential for making decisions. However, executives face data quality issues even after investing considerable time and effort in implementing performance management strategies for their contact centers. One of the reasons is that the contact center infrastructure is often built through an integration of heterogeneous sub-systems.
Flawed integration design as a cause for poor data quality
Contact centers deploy different systems like Intelligent Voice Recognition (IVR), Automated Call Distributor (ACD), Computer Telephony Interface (CTI), and Switch etc. While these systems are integrated to support contact center’s operations, reporting data from these systems for performance management is a challenge because measures from these systems are not standardized leading to inconsistencies in the reporting layer.
Consider a Contact Center, which uses Switch and CTI to record the call flow. CTI is used for performance reporting as it stores more call specific details when compared with switch. Figure 1 shows the events as recorded by these systems. Both the systems start recording events from the time a call arrives. From the figure, it is clear that till the call was placed on hold, events were recorded correctly. There is a time lag in CTI recording the event of call being removed from hold. This introduced a difference of 32 seconds in the reported talk time.
Gain insight from customer interactions
Conventional reports cannot uncover complex patterns and relationships present in interaction data. Analytics improves performance management of contact centers by providing such insights.
Since the calculation of one metric in a contact center is dependent on another, an error in calculating a base measure will have a cascading effect on other metrics as shown in figure 2. Let us analyze the previous example, which had a time lag of 32 seconds in talk time. Average Handle Time (AHT) and Agent occupancy are metrics directly calculated from talk time. Any error in talk time will affect these metrics and also Agent Pay and Cost Per Call, which are calculated from AHT and Agent Occupancy.
Data quality issues can surface even in a mature Contact Center with established processes during system upgrades or due to changes in the upstream systems.
Benchmarking contact center performance
This case study is centered on how EXILANT helped the client to establish & improve performance management solution for their contact center despite having subsystems in diverse technologies.
Figure 2 - This illustrates how data quality issues in a Contact Center can have cascading effect. As shown in this figure, the metrics are derived from the base measures. Since the calculation of one metric in a contact center is dependent on another, an error in calculating a base measure will have a cascading effect on other metrics.
Solution
It is very important to realize that a functioning contact center is not automatically ready for implementing performance management strategies. This is because the rigor applied to validate data quality during implementation may not rise to the standard expected for performance management.
Addressing data quality issues is an important step in ensuring accuracy of reports and metrics in the Contact Center. Proper strategy has to be in place to analyze the data at the start of the project to find out these issues. Once the issues are unearthed, priority has to be given to fix them. If quality issues that propagate from a source system cannot be fixed at source, a plan has to be devised to substitute dirty data with clean data. Careful attention to these steps will prevent data quality issues from cropping up during the User Acceptance Phase (UAT).
There should be an extensive data analysis phase at the beginning of the project. During this phase, source of data, relationship between data elements and usability of data have to be analyzed. Data from different systems should be compared for consistency, accuracy and completeness through total count and missing count calculations. If inconsistencies are identified between systems, they should be drilled down from aggregate level, down to an individual call to find the root cause.
Derive the right Contact Center metrics
Metrics help spot issues, identify root causes and control factors affecting customers. However, complexity and proliferation of systems in a contact center makes it difficult to derive the right metrics.
In order to compare data between systems, the most reliable system should be chosen as the System of Record (SOR). System of Record can even be log files from transactional systems. Typically in Contact Centers with multiple systems, switch is considered more robust than other systems and is often chosen as the System of Record.
During system upgrades, data has to be compared between the old and new systems to check their consistency and accuracy. This has to be performed at every layer including extraction, transformation, loading and presentation. Scripts can be written to automate this entire process.
Maintaining clean data is an ongoing activity in the Contact Center. Queries have to be written to check for missing feeds, referential integrity violations in the data periodically. The result of these queries has to be analyzed for discrepancies. Fixing these discrepancies on a continuous basis will lead to increased accuracy, completeness and timeliness of reports and metrics.
Summary
Data quality is one of the underestimated challenges in performance management of Contact Centers, which can render the reports and metrics useless. It is important to have these challenges addressed early in the project to ensure reliability of performance data. With our experience in implementing performance management for Contact Center, we can help clients handle such challenges.
Recent Links :
Performance management in Contact Centers
Performance management in Contact Centers has been a challenge due to varied technologies and enormous data involved. Through our multi disciplinary approach, we can better support our clients in their performance management initiatives.
Benchmarking contact center performance
This case study is centered on how EXILANT helped the client to establish & improve performance management solution for their contact center despite having subsystems in diverse technologies.
Gain insight from customer interactions
Conventional reports cannot uncover complex patterns and relationships present in interaction data. Analytics improves performance management of contact centers by providing such insights.
