Managing Call Drops

The phenomenon of "call drops" is known to all, as we have been experiencing it daily. To tackle this issue TRAI came out with a punitive measure wherein the consumer was needed to be compensated by the operator for every dropped call (with some caps). This regulation was later struck down by the supreme court. Thereafter, TRAI decided to overhaul the metric for measuring the intensity of dropped calls. The purpose of this note is to discuss this changed metric and the experience emanating out of it in the last two quarters.

Past   

Earlier, TRAI aggregated the "call drop" rate (% of total dropped call experienced in a BTS in the busy hour) of all the BTSs in the LSA to come out with an average number. This gave a very skewed picture, as the worst performing BTSs got overshadowed by the better-performing ones, thereby presenting the operator with no obligation to improve their performance, as the aggregated number was always less than the regulatory benchmark.

Now

The "call drop" rate for each BTS in the LSA is collected over a 90 day period and is placed in a matrix with data of each BTS (in a row) stacked one below the other. After doing so, each row is then sorted horizontally in an ascending manner. Then data of the 81st column is selected after omitting the remaining 9 days (10% of the total) to account for uncertainties (beyond the control of the operator). This selected column of data is again sorted in an ascending manner, and thereafter, the one at the 90th percentile is picked to represent the benchmark, which is called QSD(90,90), i.e Quality Spatial Distribution (90,90). The acceptable level of this benchmark (as per TRAI) is less than or equal to 2%.

Instead of sorting the rows first, one can sort the columns, and pick the data in the row representing the 97th percentile. Thereafter, this row can again be sorted and the 90th percentile data can become the 2nd benchmark called QTD(97,90), i.e Quality Temporal Distribution (97,90). The acceptable level of this benchmark (as per TRAI) is less than or equal to 3%.  

Analysis

After the metric for call drop was changed by TRAI only data for last two quarters are available. This I have captured pictorially. See figure below. 

One can see that many operators in various circles are above the benchmark for both the metric (QSD & QTD). Though there has been some improvement in the March quarter (compared to the previous), but some (Tata, Idea) have deteriorated as well. 

Impact

Since the punitive impact on non-compliance is low, one can question the efficacy of the new metric. Does this serve the intended purpose? At least the perceived impact on the ground of improvement is absent. Hence, the only way to revamp the situation is to enable the consumer with a choice of a better network. How would the consumer know the quality of the network in advance without experiencing it? One approach would be to package the BTS level data graphically (tagged to a location) for the consumer see for himself which network it performing better. This will also trigger competitive sentiment among the operators, thereby motivating them to improve. 

Conclusion

Though the new metric is a signification improvement from the earlier one, however, in absence of punitive measures, the objective might not get fully achieved. By punitive measures, I do not mean cash penalties, but laying out a system which makes the network quality transparently visible to the consumer. Globally this is done by drive test data published by independent agencies. In India, in absence of such an option, one can publish the coverage map of BTSs and its performance (clearly defined by TRAI) transparently to enable the consumer to make an informed choice thereby motivating other operators to catch up in order to prevent churn. 

(Views expressed are of my own and do not reflect that of the employer)

PS: Find the list of other relevant articles in the embedded link.