Nuisance Call Guidelines

Simon Woodhead

Simon Woodhead

21st March 2013

In common with all UK CPs/Network Operators we recently received a joint letter from OFCOM and the ICO on the behaviour they expect in respect of:

  • Unwanted sales calls and spam texts
  • Abandoned and silent calls

The letter is attached for you to read and to ensure compliance with. However, as the Network Operator potentially carrying your traffic we wanted to highlight a few areas of observation and discuss some of the measures we take before they impact you.

The spectrum of “sales calls” is very wide, ranging from manually dialling qualified prospects through to aggressive automated diallers calling sequential unqualified (and often non-existent) numbers. Whilst qualitatively determining what is “good” or “bad” is impossible for us we can do so quantitively.

We spend a disproportionate amount of resources pre-qualifying new customers; an unhealthy proportion of sales enquiries are seeking to find homes for traffic which fits squarely at the “bad” end. We loathe such traffic and do not knowingly allow it on network. This is our contribution to consumer protection but it also makes commercial sense when the costs of such traffic are fully analysed. Hopefully legitimate new customers do not notice this process but our team are very familiar with the indicators of this type of traffic, the answers they will have rehearsed to our questions and some revealing further questions they will not have been asked before.

Unfortunately some occasionally slips through, usually by way of a historically excellent CP beginning to pass it. We therefore analyse traffic statistics and are hopefully quick to assess what is “bad” quantitively. To assess our own performance we monitor in real-time the timings for every call in and out of the network; what was the delay between the call being initiated and us returning ringing progress (PDD), how long did a call take to answer, how long was a call whether it answered (ACD) or not. We also monitor the success metrics for calls, not just the percentage which answer (ASR) but the percentage of each type of failure. You’d expect us to do this for our onward carriers to manage the quality of our routes but the statistical quality of a call is affected by both the onward carrier and the originator. For example, if a carrier is only sent calls to non-existent numbers then their ASR will be poor yet that is the fault of the traffic not theirs – they cannot magic a number into existence if someone has dialled one which does not exist. Put another way, the information we already record to monitor our own and carrier performance is exactly relevant for quantitively analysing “bad” traffic.

However, we have recently refined all of this as whilst low ASR is an indicator of potentially high calls to invalid numbers, and low ACD indicates the presence of recorded messages or unsolicited calls, we have identified other metrics that indicate this type of traffic. We now calculate the usage of our network resources down to the micro-second, so even failed calls are measured – quite rightly as they’re using valuable resources. We’re measuring the ‘efficiency’ of customer traffic across our network by comparing actual resource utilisation with billable minutes and we’re finding a direct correlation with “badness”. It goes far beyond simple ASR and ACD and gives us a single unified score per account which in some cases can be surprising, but rightly so on closer investigation. Of course, what we need to refine is what score we consider to be the boundary and that will evolve, as will the other metrics we monitor.

If you’re reading this the chances are you’re sufficiently diligent for your traffic to be the right side of the line. If you’re not, you will have heard from us and if you cease to be, you will hear from us. However, you need to ensure your own and your customers’ compliance with the guidelines. You are responsible for your own regulatory affairs after all and what we consider acceptable may be far from so to OFCOM and the ICO when they look more closely at it.

We think there is specific action you need to take in the light of the guidelines irrespective of your business:

  • The guidelines give specific quantitive and qualitative requirements. You need to be measuring these as you cannot manage them otherwise. We suggest measuring them per customer, if not per CLI, or if you have such visibility per campaign.
  • You need to ensure you are passing valid CLI for all calls. We expect and require this but do not yet enforce it. We’ll repeat that one: You need to ensure you are passing valid CLI for all calls.

Moving forwards you can expect us to introduce a few measures which you should note as they may impact your traffic:

  • We will begin blocking call attempts with invalid CLI.
  • We will use our efficiency statistics to regulate network capacity available in place of manual channel limits – good customers get the resources they need to grow, bad traffic will be automatically constrained to the point of eventually being blocked. This is automatic with no opportunity to give the answers we want to hear – the stats are there, we just need to let it loose after further testing.

Please take steps now to protect consumers and ensure our measures are not going to impact you. Aside from being good industry participants this kind of traffic is economically damaging and if the problem continues to grow we’ll all be paying more in the way of levies to fund more regulation. Nobody gains from that.

As always, if you have any questions or concerns about this, please get in touch.

Related posts