Managing the Hold Queue
We know a lot about the hold queue from a technology standpoint. It's an easy matter to quantify the time people spend on hold. The whole notion of service level comes from the idea of answering a certain number of calls within a certain amount of time. Behind that is the idea that the people not being served are sitting in a virtual waiting room until their number is called.
What do we do with those people? The options include:
• Playing them messages and/or music to enhance (or ameliorate) the experience.
• Allowing them to leave the queue to perform some other action, like querying an IVR system for basic information.
• Doing nothing at all.
Few people will argue that you should do nothing, but that's often what happens, particularly when the time spent in the queue is too short to allow for any meaningful activity. We should count this as a success.
Playing messages on hold is an old idea that is still popular, and with good reason. One executive at a music-on-hold production company once told me that if you look at messages presented to consumers on hold as a form of advertising, it has the lowest cost per impression of any comparable media.
Under these circumstances, you want to use the hold queue as a value opportunity. You want to tell people what you have to offer and bring them information about your company to which they might otherwise not be exposed.
The hold queue exists because more customers are trying to reach you than resources are allocated for handling those customers. If you equalize those two factors, the hold queue disappears but your costs go too high.
Therefore, you want to build in a certain tolerance level. You need to realistically assess what type of waiting time your customers will absorb, and under what circumstances. If they want to buy something from you, you want to answer their call immediately. The more valuable the customer, or the potential transaction, the lower the tolerance the customer will have for waiting in queue.
So if you are selling $5,000 cruises to people who have booked in the past, answer immediately. Likewise, you want to take the call from the Platinum cardholder, the frequent flier and the guy who has been to your Web site to gather all the information he needs to decide between buying your widget or your competitor's widget.
In the same spirit, you probably also want to minimize the hold times of people who owe you money. The collections environment has a call-center dynamic all its own. But it's pretty clear that when someone responds to a collections mailing, the worst possible thing you can do is keep that person waiting on hold. If they call to resolve, it's up to the center to allocate the resources to make the call shorter and deal with the customer quickly, before the opportunity is lost.
That's what the hold queue is all about. It is a delicate balancing act between the call center's cost of doing business and the opportunity presented by the customer.
With too few resources allocated, the opportunity will go away. With too many resources, the opportunity may not be worth pursuing. Like everything else about running a call center, how you handle callers on hold depends on the relationship between the urge to reduce costs and the need to drive revenue.
Back to using messages. There are two ways to use hold media. One is what I will call the opportunistic method. That involves telling people more about your company and products you offer in short ten-second bursts. There is little else you can call these but commercials. They are most prevalent on inbound telesales queues, and they can help interaction somewhat by providing cross-sell and upsell information at no cost.
You see this a lot less on the service side. Frankly, people do not enjoy that medium when they call with a problem. In fact, it's a bit of a vicious circle. Because when you allocate resources, you are almost always going to tolerate a longer hold time on non-revenue-generating service lines than on sales lines. As a result, advertising messages cycle through more often during a typical hold session, and the annoyance factor increases.
That's why most of the service applications we've seen tend to use hold media somewhat differently. They accept the premise that you're going to have to wait longer for service than you are for sales and that service from an agent is a more precious resource than service from the Web or an IVR system. So it makes sense to give people as much information as possible to shorten or complete that call before it hits an agent.
In a true technical support center, for example, there would likely be a series of selection criteria applied before the caller ever went into the hold queue. These can be identifiers, telling the routing mechanism who the caller is and whether there should be any special priority assigned. It may also include special information gathered, such as an account number or the number of a service contract, that will give the automated router some classification for the kind of problem that might be coming.
The simplest implementation of this type of application is "press one for Windows, press two for Mac," but it can be used more artfully to get the call to precisely the right person, therefore reducing hold time and the amount of transferrng among agents the call requires.
You can also play messages during the queue that alert the caller to the true length of that queue. If you tell a person that there are 20 calls ahead of him, or that the average call takes 4 minutes, or that there's an estimated 15-minute wait before reaching an agent, that's power in the hands of the customer. That means you won't have to spend 30 seconds mollifying someone who's waited too long before solving his problem.
You can also invite the caller to use other options. Giving out a Web address is a good idea. Offering a faxback system is also good, but less likely to be used during the wait. Offering to let callers pass the time in an interactive diagnostic voice response system without losing their place in the queue can be valuable on your end as well as theirs.
If you are CTI-capable, you can pump information about what the caller has queried right to the agent's desktop, saving time and giving the rep more to work with in diagnosing a problem.
In all, how you handle your hold queue will say a lot to your customers about how you feel about them. It presents a lot of opportunities, but can often be a neglected component of call center operations.