When the TeleManagement Forum defined revenue assurance in its TR131 technical document, they made it clear it was about things that telcos control – their systems, processes, data and the like. It was not about things they do not control, including the behaviour of their customers. There is good reason for that distinction. It is hard enough to distinguish good from bad in the field of revenue assurance without making it a speculative enterprise too. Unfortunately, not everyone agrees.
I worked at one telco which used to regularly report how much value they added. Every month usage records would leak somewhere or other, and by the magic of revenue assurance they would get recycled or saved or recovered or whatever. By virtue of this magic, the revenue assurance department would claim they had added value to the business, which was reasonable. And they calculated that the value added = the rated value of the records they saved. But that was wrong. Because postpay customers of that telco all benefited from allowances, the actual total value of postpay usage billed was significantly less than the total value of postpay rated records. So the true economic value of saving those records, however you calculated it, was strictly less than their rated values. So in this telco, the revenue assurance department were not even very good at predicting the incremental revenues they were responsible for generating. They were instead guilty of systematically exaggerating them, which was ironic given that was the common criticism that they levelled at Marketing.
There is a fortune to be made out of predicting customer behaviour. But let us be clear – predicting customer behaviour is a gamble, and gambling is not a science, no matter how much science or mathematics you use. Studying the behaviour of people, like studying the form of a racehorse, is prone to error. Even in the most heavily-scrutinised and simplest cases, political elections, human behaviour often confounds expectations. Election pollsters have got it wrong surprisingly often. This is despite their prediction being binary in most cases: either candidate A wins OR candidate B wins. There is a famous picture of President Harry Truman on the night of his victory holding aloft a newspaper that ran the headline “Dewey defeats Truman“. But polling has not got that much better since 1948. The pollsters were wrong about the outcome in the 1992 UK general election and much has been written about how the US television networks “called” the outcome of the key Florida state in the 2000 US presidential election first for Gore and later for Bush – see this CNN report for a relatively unbiased take on what went wrong that night.
So opinion pollsters, who try very hard to get it right and build upon a long legacy, sometimes get it wrong. Now consider the average revenue assurance department: a small department in a telco, with very limited statistical skills, unreliable and partial data, no training in customer behaviour and every reason to exaggerate their successes and hide their failures. Would you trust them to correctly predict how customers will act? If they sometimes lack the skills or honesty to measure revenue assurance accurately, they should not be trusted to reliably gauge customer behaviour. So telecoms companies should beware when trusting revenue assurance departments to second-guess how to make more money from customers. RA managers may be tempted to play the role of fortune tellers, but my prediction is that they are better encouraged to be honest and objective reporters of business performance. Because if they report on the success of their own guesses, you can be sure the predictions will be more impressive than the reality.