Dear,
Currently, the average SLA calculation is based on a simple average of the percentages from each call interval. However, we have identified that this method can skew the results, as it does not account for the actual call volume in each group, giving equal weight to intervals with very few calls as it does to those with high volumes.
Issue with the Current Method (Simple Average):
- Example:
- 1,092 calls with 66.46% SLA (725 answered within ≤5 seconds)
- 74 calls with 49.4% SLA (37 answered within ≤5 seconds)
- Current calculation: (66.46% + 49.4%) / 2 = 57.93%
This result does not reflect reality, as the 1,092 calls (representing 93.6% of total volume) carry the same weight as the 74 calls (just 6.4% of total volume).
This is clearly biased, as one call would have the same impact as thousands.
Proposed Solution: Weighted Average
We recommend adjusting the SLA calculation to use a weighted average, ensuring each interval’s contribution is proportional to its call volume. The formula would be:
Benefits of the Change:
More accurate results, reflecting the true impact of each interval.
Eliminates distortions caused by small sample sizes.
Aligns with performance analysis best practices, where larger volumes should carry greater weight.
Formal Request:
We kindly request the implementation of the weighted average method to ensure the average SLA is calculated based on the actual number of calls answered within each interval, rather than a simple average of percentages.
We appreciate your attention to this matter and are available to discuss technical details or assist with implementation.
Best regards,
Luciano Moreira