If you live here in the States, you may never have heard of the telecommunications company EE. But they are a major player in the United Kingdom with brands like Orange and T-Mobile. According to their Wikipedia page, they have around 28 million customers. EE has a new service offering that I must admit is kind of intriguing. Here is how it is described on their web page.
Priority answer service
From 6 August 2014 we’re also introducing a priority answer service. It’s available to all customers on pay monthly and SIM only plans.
Our priority answer service gives you the choice to get support even faster for just 50p per call when you call 150 and want to speak to customer services. It’s always available so if there’s a queue, you can be moved towards the front – ideal if you’re in a hurry.
How much it costs
The charge for this is 50p. If you’re on a plan that includes standard charging for customer services at 25p, you’ll only be charged an extra 25p for priority answer – so the total for the call with priority is 50p.
The 50p charge applies regardless of how long the call lasts.
To save the Americans the trouble of Googling this, 50p works out to about 84¢. So what do you think happens when customers are given the chance to jump the queue for less than a buck?
Before getting to that question, let me just note that EE has some interesting ways of managing their customer service lines. For example, if you call outside of prime business hours (i.e., you call between 8pm and 10pm on weekdays and 6pm and 8pm on weekends), you also get hit with a 50p charge. That is actually in line with costs. Assuming they are staffing to provide the same service level (e.g., average wait is less than some target), then the system will operate at a higher utilization during busier periods. That, in turn, implies lower costs. If the threat of an extra charge induces some calls to slide from 7pm to 4pm, the difference in utilization (and hence costs) gets more skewed. That’s a pretty cool service operations example.
OK, so much for the tangent, what do you think they happened with customers? You’re right: They were annoyed and whined on social media. See PCPro for a sample. The same article has the following tweet from a reporter.
Let me note that there is really nothing theoretically uncertain here. As long as EE is using the same agents to answer both priority and regular calls (and according to The Telegraph that is what they are going to do), the wait time for regular calls has to go up much as apples “theoretically” fall from trees to the ground. Note that with this priority scheme, the average customer may not be worse off (average here means averaging over high and low priority customers). If the average service times of those jumping the queue is the same as those that tough it out in the regular queue, then the time the average customer waits won’t change — if everyone takes the same amount of work, it doesn’t matter in which order the work is done. Of course, if those opting for priority service tend to have more complicated and time-consuming needs, then those at the back of the line are particularly screwed.
In addition, there is a question of how you manage the agents. I have been implicitly assuming that agent time is never wasted. If an agent finishes a call and another call is waiting, the agent moves immediately to handle that call. But that’s not how I would manage this call center. Assuming that there is higher service level that needs to be achieved for high priority calls, it may make sense to follow a threshold policy in answering lower priority calls. That is, when an agent is free to take a call, she immediately takes a high priority call if one is waiting. If there are no high priority calls, she takes a low priority call only if there are at least some given number of other agents already idle. Note that this saves more capacity for high priority customers and hence allows for hitting their targeted service level with fewer agents. Of course, from the perspective of a non-queue jumper, this implies a double whammy: there are fewer agents and they are harder to get to.
This uncertainty of what the impact is for both kinds of call seems to be the real problem with this plan. Here is how someone put it in the Daily Mail (50p to jump mobile queue: T-Mobile and Orange customers can pay fee to have their call answered first when they ring helpline).
Rob Kerr, mobiles expert at telecoms price comparison website uSwitch.com, said:
‘Priority queuing will create a two-tier system.
‘It’s all very well slashing waiting times for those who pay the 50p premium – but EE’s not made it clear if that means non-priority queuers will be left hanging on for longer than usual as a consequence.
EE could use all the cash it raises 50p at a time and expand total capacity enough that all customers are better off. Indeed, EE’s explanation for its new charges is (essentially) that it has brought call center jobs back to the UK and needs to fund that added expense. One interpretation of this plan then is that the added cost of improved service is going to be borne by those that use the service as opposed to mobile subscribers who rarely call into the call center.
It is also worth noting that EE hasn’t (as far as I can tell) publicly promised a waiting target for priority customers. That is, they are asking you to pony up 50p without explicitly saying how much shorter your wait will be. All you know is that your wait won’t be as long as those in the regular line. That might seem unreasonable, but you have to realize that they cannot always guarantee that the priority wait will be less than say, 30 seconds. The inherent uncertainty of queuing systems makes it too costly to always staff to handle the worst possible load on the system. Even if they did, the wait for regular calls would generally be so low that no one but the unaware would pay to jump the queue.
Finally, let me close with a shameless plug. If you find this stuff interesting and you have to teach an Ops class, check out this case I wrote several years ago.