Short Introduction to Analytics & Big Data
Regardless of the growing disparity of the traffic and revenues, operators don’t want to limit their customers’ data usage. Supplying the demand is the very fundament of any business. Hence the importance of proper capacity management!
During the mobile data surge – which most operators are witnessing – capacity management becomes a truly vital daily operation. Most operators I have worked with do have some type of systematic capacity management processes in place. In the simplest form it means:
From what I have seen in our projects, the most imminent discrepancies are typically hidden in the first and third parts of the workflow: It’s easy to forget the requested traffic that was never delivered due to capacity constraints. That is, to forget the undelivered traffic.
When working with operators’ network engineers, I’ve often noticed a perception about the network being close to its limits, but still carrying the requested traffic. When we then audit the true capacity utilisation and uncover also the undelivered data, the results are typically somewhat shocking. In some cells, the proportion of undelivered data can be even 50% of total requested traffic.
The challenge is that traditional performance counters often don’t directly reveal the traffic that never exists. Excluding the undelivered data however results in too optimistic view of the capacity and faulty traffic forecasts.
Uncovering the undelivered data gives also means to quantify the associated lost revenue. Often it is only after such analysis when operators truly understand how business-critical the capacity management process is. From what we have seen in our projects, I can conclude that the revenue loss due to capacity limitations is usually measured in million €/month scale.
And not to forget, this is lost business from the already acquired customers! The way I see it:
The needed capacity expansions can be made very cost effective when the precise need is predicted and optimised on per cell level. This is why analytics-based capacity prediction and capacity management should be the first line of action for any operator with data traffic surge.
And in my view, analysing the undelivered data is the number one thing in capacity management.
The article belongs to a mini-blog series about CoDriver™ Predictive Analytics. In the series we will uncover hidden trends and causalities in mobile data business dynamics. Read the first article here.