By Simon Woodhead
There’s a massive change coming July 1st in that most UK mobile operators will be surcharging calls to their subscribers based on where calls appear to come from, largely but not entirely based on the underlying network number. Several fixed-line operators are following next month. I’ll save my fury at the absolute avoidability of this if we had a regulator who listened to the warnings of consequences of their own proposed schemes, so let’s instead look at what I think is going to happen.
The surcharges are not trivial. NDAs prohibit me from saying how much any one provider is charging but you’ll see, looking round the market, that those passing these charges on are doing so at up to £2 per minute. That is 528x the regulated wholesale price of a mobile call – the so-called MTR – and I suggest that there is no operator in the market-place at any level, least of all wholesale, charging anywhere near that to their customers. Even BT is apparently only charging its mobile customers 96x MTR out-of-bundle.
Retail prices are probably going to rise
Contrary to the presumed intent in Ofcom driving down wholesale prices, and the very legal Act underlying their formation – their principal duty being to further the interests of citizens in relation to communications matters – operators are going to have to pass these surcharges on to consumers. What we don’t know yet is what proportion of calls are going to see a 528x increase in cost and thus how much that increase will be.
Companies are likely to go under
No company at any level of the supply chain can sustain a 528x increase in costs that is not passed on. There’s going to be bill-shock at every level, not just small players. There’s a former FTSE-100 operator who, as far as I can tell, has completely forgotten to handle surcharges and is going to attract quite a lot of traffic! This being a small industry, I believe their main transit supplier is surcharging everything at the maximum rate to reduce their own risk, ergo, they’re going to see more traffic than normal at 528x the cost.
Calls might not complete
Telephony networks obviously have redundancy built in, some more than others, of course! We also all have multiple routes to any one destination so even though we directly interconnect with many terminating networks, we’ll have several options to transit other networks to get there should we need to. Ofcom paints a rosy picture of perfect competition and fluidity of supply but in the real-world there’s alarmingly few choices for transit and many use them heavily. The mobile operators in particular seemingly do everything they can to avoid direct interconnections, in our experience – we have more than most, but not the full deck – so as a destination they depend on (and consume) transit heavily.
The largest transit provider in volume who is not the incumbent – in my opinion, there’s few competent stats on this – has essentially withdrawn from the mobile market as a result of these surcharges. They can’t handle the billing system changes needed so have changed their rates to be the maximum they possibly could be while they do. They fully acknowledge the traffic will move away. That is a huge huge step for a business to take: sorry customers, we’re closing the door because the risk of leaving it open is too huge!
Whilst that illustrates the gravity of this situation, there’s practical consequences too. Where is that traffic going to go? I can speculate and would just say I hope Ofcom will be watching it unfold before they next paint the rosy competitive picture! Practically, if a majority of operators are using their second and third routes (where they have second and third routes!), is there adequate capacity on them? If there is adequate capacity what does it mean for redundancy if those routes fail?
Beyond actual routing, anyone who has implemented this properly (like we have) is going to see a multiplication of call routing processing. We already perform hundreds of thousands of lookups per second to route calls with everything handled in real-time, with consistently the lowest PDD amongst our peer group when tested. Much of that is our unique fraud controls but even those who don’t offer them are going to see at least a doubling, more like 4-5X of call routing look-ups. God help those who are still doing so in SQL Server 2000, and their customers! In all likelihood though, they’re just passing through the surcharge and winging it on the buy-side as they can sort it out next month in a few spreadsheets – the total opposite to our approach which I find telling!
And of course there’s already profiteering
Of those who have passed surcharges on, some are doing so in different ways. Most are simply passing through the full complexity of different surcharge ladders and triggers for every network, whilst others are simplifying it. There’s merit in that and they’re even charging a lower than maximum surcharge, perhaps to look attractive, but they’re doing so on calls that otherwise wouldn’t be surcharged to make up the difference, or more likely more.
We also notice that one major transit provider is charging up to a £2 surcharge on calls to Simwood mobile numbers despite us not levying any surcharge on them.
That’s all a bit dirty if you ask me.
I hope, rather like the millennium bug, that the above doesn’t reap havoc – it is the last thing the economy and society needs – but am absolutely confident that they’ll happen at some level even if word of it doesn’t break the surface.
However, there is some good news for Simwood customers.
Caller ID clean-up
It is over a year since Simwood started rejecting calls where the CLI was mal-formed – the kind of traffic that will attract the highest surcharge from terminating networks. Whilst Ofcom have seemingly done nothing to enforce it, their General Condition C6 requires this but we’re unique in doing it, perhaps because they don’t seem to enforce it. The pain our customers endured over a year ago now needs to be felt across the industry – it is just that they’ll see a 528x price increase from our peers rather than a warning email from us as our customers did!
Everyone else cleaning up their CLI is good for us and the consumer though. We’ve never been able to turn on inbound CLI filtering because, frankly, there is so much crap spewed out by other networks. Now that crap is expensive and certain mobile operators finally have economic incentive to comply with 20-year-old guidance that long predates GCC6, turning it on might become possible.
That is good news. It is just a shame Ofcom couldn’t enforce their rules with the powers that they have rather than relying on unintended consequences to do it for them!
Simwood is not passing on surcharges
We’ve said from the start that because your CLI is clean and because we’ve long had dynamic real-time loss-protection on the network we want to avoid passing on the surcharges. I just don’t see how the average wholesale customer will implement them so this feels the right thing to do. What this means though is that you may see calls rejected where you trip arbitrage limits, i.e. if an unnatural proportion of your traffic attracts surcharges. That is necessary and we’re unapologetic – send us clean traffic and calls will go through, with us taking the pain on the occasional surcharge. We’ve also warned this may cause us to raise prices, depending on the blend.
Every one of our competitors that has acted, is passing the surcharges on. So they’re going to become a thing our customers have to adapt to unless they send 100% of their traffic to Simwood. Whilst we’d welcome that, there’s a dilemma here worthy of a game theory study!
Some bright-spark is going to send us as much of their surcharged traffic as they can get away with and send the non-surcharged somewhere cheaper – cheaper only because they’re not absorbing the surcharge or worse still are making on it. That’ll cause us to put our prices up, and the more we do the more likely it is we’ll just attract the rubbish. There comes a point of equilibrium where we’re only carrying surchargeable traffic at a ridiculous rate. That obviously is not desirable or sustainable.
So, whilst we’d hoped the regulator might show their teeth and halt this nonsense, or competitors might do the right thing by customers too – such that surcharges were just a new irritant in the carrier space that doesn’t get passed down the chain – that hasn’t happened. We will therefore almost definitely need to introduce surcharges.
We might be surprised of course and indications are that we’re picking up a lot of new transit business on July 1st because of this approach. If surcharges are just an occasional irritant and we’re not abused then we’ll maintain this position.
But don’t try it on
For the record, our dynamic loss protection is applied as a limit per customer per destination – a profit and loss tally essentially with a permitted level of loss. Those who have taken the mick previously or we just know are untrustworthy, will not be allowed to pass any loss-making traffic. Those who send enough to trip their loss-limit will also not be allowed to pass any loss-making traffic until they pass breakeven. Those who sail on the loss limit by just sending us as much rubbish as they can will be quickly marked as untrustworthy and we’ll only accept non-surcharged traffic from them.
The same is true of the obvious bodges to avoid surcharges. We know what they are and will mitigate them because they expose us to retrospective risk.
Given the supply side of things, it is possible we will need to turn off channel bursting and strictly enforce channel limits. If you rely on bursting above your committed channel limit you may wish to consider increasing ASAP. We are substantially increasing capacity on our side to account for expected hot-spots as traffic flows shift and we win new business but do not know how things will look downstream. There are going to be massive swings of traffic as a result of the market withdrawal mentioned above.
Whilst the lowest risk position at the moment for you is to send your traffic to Simwood and be assured of no surcharge risk, if we need to add surcharges, that still holds true. For years and years you’ve been able to pass headers with your traffic that cap the maximum rate applied on calls. This was intended to combat fraud but is perfectly applicable here. So, for example, if you don’t want to pay surcharges above say 10p, you just set 10p in the header, we’ll work out what we’re going to charge you and if the rate is above 10p then we’ll reject the call. You set that rate in the header per call so you can be as dynamic as you like with it. Also don’t forget our real-time calls in progress views that’ll show you any expensive running calls if you haven’t added the header to prevent them.
Of course, we’ve learned to our dismay over the years that all the best fraud controls are pointless if customers simply then send the rejected call elsewhere and get screwed anyway! That is sad for us – we incur the cost to protect customers for free because it is the right thing to do, meanwhile competitors are rewarded with windfall margin for choosing to profiteer.
Stay safe folks and I hope the next month or two doesn’t harm any of our customers, or indeed us.