TL;DR we’re blocking more calls, both inbound and outbound, to combat nuisance calls.
Increasingly often, I find this industry like a parallel universe. The disconnect between what is said and what is really done is incredible. If our politics wasn’t [Simon, no!] in such a perilous state, it’d warrant the kind of media attention bestowed in the past on other fundamentally dishonest entitled industries. As Pete has clarified, the current furore around new CLI requirements is utter BS, and the sensationalist tabloid-esque newsletters from our regulator heaping praise on the monopolist for what some of us did in 2018, causing friction with both the monopolist and the regulator, is mind-boggling. Then there’s those who claim coverage in 300 of 195 countries, to be CPaaS simply to inflate their valuation, or to be profitable only on the basis of non-GAAP accounts – you can read “non-GAAP” and “Enron” in the same sentence here, it wasn’t me who said it Your Honour. Meanwhile, cretins adopt what was brilliantly framed the other day as “end of life pricing” in dirty origin surcharges, driving their customers away from phone use to OTT services like WhatsApp whilst simultaneously virtue signalling. Down here in the real world, little old ladies are having their life-savings stolen by an industrial level of fraud against UK consumers. Some of us care about that.
As a mostly wholesale operator, we’re generally several steps removed from the consumer but we’ve always tried to protect those who, however indirectly, depend on us. They may not be our customer, and it might cost us money to do so, but if we can stand between them and harm, we should and we will. Sometimes that puts us at odds with our own customer, who has a different set of values, and those few tend not to remain our customer. In fact our sales function for the last few decades has been primarily scrote-filtration, with the usual things you’d associate with sales being way down the list – we could be 100x the size in a matter of weeks if we enabled the mass of shite traffic looking for a home. And that brings me to the point of this post…
There are those who will provide this traffic with a home, often contrary to their public position, and as a percentage of traffic across networks it is on a relentless march upwards. Yet something else to persuade modern users to not bother with the phone if it increasingly just conveys scams, fraud and aggressive sales. We really need to deal with it, more aggressively than might have been acceptable or even technically possible in the past.
There’re two routes we need to tackle:
Outbound traffic from our carrier customers
Now the rest of the industry is at least talking about doing what we did last decade (we know they’re not actually all doing even the minimum), and we’ve already greatly elevated our KYC, we’re pretty confident we do not have customers with all-bad traffic. But as we have said before, when carriers work on aggregates and ratios, what becomes important is the overall blend of traffic and a long-standing good customer has both a massive asset in their account, and huge moral hazard as bad traffic is blended in. Whether this comes through lax KYC-controls or from a briefcase of cash in a car park, good customers do turn bad, or more often good customers simply look less good as more bad traffic is blended into their mix. Imagine several layers in a supply chain, each tuning their ‘blend’, and there’s an awful lot of shite sailing under the radar. We need to deal with that on a per-call basis to weed it out.
Accordingly, effective today, some customers will see certain calls rejected with a 603 Decline where they have previously completed. This is based on behaviour over the last 24 hours, and the last several months. We’ve set the bar pretty high for now but we will be tuning this further. It is not subject to a set of metrics we’ll publish so you can tune your dialler to comply with, nor something shouting at our support colleagues will change – clean traffic will pass. We’re using a combination of our own data and developing the decision making with further AI. Ultimately we don’t want to block calls unnecessarily but where we have a high degree of confidence that they will result in consumer harm, we will.
Inbound calls to our carrier customers
Inbound is far harder, not least because we have regulatory obligations to provide end-to-end connectivity, and they trump the ‘guidance’ everyone else is getting excited about (that we implemented last decade, in case I haven’t mentioned it). It is also more nuanced. As we train our models with live data there are some edge cases. For example, the highest volume contender for blocking sees over 90% of calls actively rejected by recipients and only 7% of calls connect, so that’s a clear cut one to block, right? But then 7% of calls do connect and have decent call durations, so that swings back to less clear-cut. We don’t want to block legit calls just because many don’t want them; we don’t want to permit scam calls just because a few people are conned into engaging – they look the same statistically. This is therefore a journey and we’re starting with a relatively high bar but will definitely be turning the screw.
We already block calls on Ofcom’s DNO list. Every operator should but we know with certainty that they don’t (despite virtue signalling), so we filter them inbound as well. We are also now blocking the worst of the worst looking calls by virtue of our own on-net data. There’s a handful of these with current data but they pass a huge volume of calls. There’s more we could block this way (and would on outbound) but we learned in 2018 that by virtue of relative market share, whichever way a call fails and however right that is, it’ll be deemed our fault (and thus our support deluge) over a major brand; there’s way too many calls of this nature from certain suspects that the level of false positives would be huge until they get their act together. We refuse to rely on consumer forums and the kind of dangerous data that muppets at other firms live by (and the kind that lead to me having death threats in the early noughties), but we do really need to get to the intent of the call. That’s where AI comes in.
We have a secret project in flight that I’m itching to tell you (and many of my colleagues!) about, and that wasn’t on our quarterly road-map. It is nothing to do with nuisance calls specifically but can be applied to be really really helpful in this use case. We’re also keeping a watchful eye on what some of our overseas colleagues are up to – sampling the media on a small percentage of calls and looking for triggers via AI. This makes us slightly edgy from a privacy point of view but is eminently possible with the Potato. So lots more to come here as this journey continues!
Simwood Potato®
Speaking of the Potato though, it almost goes without saying that all this functionality is available on any network globally by virtue of the Simwood Potato and specifically BYoC. Point your numbers with a legacy provider to us over BYoC and this blocking is automatically applied, protecting your customers where your other carrier would sooner let them be ripped off while creaming off some origin surcharges. They’ll say all the right things of course, in one case with absolutely no substance and consistently in direct response to posts on this blog.
***
Watch this space because we’ve got the bit between our teeth on this one and we’re only just getting started.
And finally, Pete would like me to remind you all about your obligations with respect to sub-allocating numbers (that includes porting-in numbers for your clients) and the Ofcom’s pronouncements with respect to the same here.
