Failure demand in the contact centre: find it, measure it, remove it

Failure demand is demand caused by a failure to do something, or to do something right, for the customer. It is usually set against value demand: the contacts customers make because they want something from you, such as buying, changing an order, asking a question for the first time or reporting a new problem. Value demand is the reason a contact centre exists. Failure demand is the work an organisation makes for itself, and its customers, by getting something wrong.

The distinction is about cause, not about the call type on a menu. Two customers asking about the same bill can be making very different contacts: one wants to understand a new tariff, the other is calling for the third time because a refund has not arrived. The first is value demand. The second is failure demand, and it will keep coming until the refund process is fixed. That is why failure demand cannot be found from queue names or reports alone; someone has to look at what customers are saying.

Failure demand takes capacity that should go to customers who want something new. It lengthens queues, so every caller waits longer, including those with value demand. The customers making failure demand contacts arrive already frustrated, so the conversations are harder and longer, and agents spend their day apologising for problems they did not cause and cannot fix. Over time that wears good people down.

The usual response to rising volume is to forecast it and staff for it. That keeps service levels up, but if a large share of the volume is failure demand, the organisation is paying to absorb its own mistakes, and the people who could fix the causes never see them. Measuring failure demand turns a staffing problem into an improvement opportunity, and gives contact centre leaders evidence to take to the parts of the business where the failures start.

Start with real contacts, not reports. Take a sample from each main queue and channel, listen to the calls or read the messages, and classify each one: value demand or failure demand, and, for failure demand, what failed. Ask agents and team leaders to help; they already know the common causes and will spot patterns faster than anyone. A few hundred contacts across a couple of weeks is usually enough to see the shape of the problem, and the conversations it starts are worth as much as the numbers.

Once you know the main categories, make the measure routine so you can see whether things improve. Wrap-up codes in Genesys Cloud let agents record the nature or outcome of each interaction, and those codes feed reports, dashboards and Architect flows. A short list of failure demand codes, kept simple enough that agents use them consistently, gives a running view. Speech and text analytics can add evidence at scale: topic spotting detects subjects using predefined phrases, such as "I have already called" or "still waiting", and content search finds interactions by words and phrases.

• Sample real contacts from each queue and channel
• Classify each as value or failure demand, and name what failed
• Involve agents and team leaders in the classification
• Add a short, consistent set of wrap-up codes for the main failures
• Track the share of failure demand over time, not only volumes

Most failure demand starts outside the contact centre: in letters and emails customers cannot follow, a website that hides the answer, a billing process that makes errors, an order system that promises dates it cannot meet, or a hand-off between teams where cases go missing. Some starts inside: callbacks not made, cases closed before they are resolved, or advice that differs from one agent to the next. Group your failure demand by cause and owner, and the biggest few usually stand out.

Handle time can point the way. When the time taken on one call type suddenly rises, the cause is rarely the agents; something has changed elsewhere, and customers are calling about it. Treat that spike, and any sudden rise in repeat contacts, as a prompt to listen to recent calls. Then take what you find to the people who own the cause, with examples in customers' own words. A recording of a confused customer reading out your letter is more persuasive than a chart.

It is tempting to point automation at the biggest queues. If those queues are full of failure demand, automation makes it easier for customers to chase you, but the failure is still there and the customer is still unhappy. Fix the cause first: rewrite the letter, keep the promise, make the callback, close the case properly. Every failure removed takes contacts out of every channel at once, and it takes nothing away from customer experience.

Automation earns its place on value demand that is simple and frequent. Dave Tidwell, Managing Director of QVCCS, puts it plainly: simple transactions should always be capable of self-service, online or in the IVR, visual or voice. Which transactions those are varies by sector, so start from your own demand analysis. Automating caller identification and verification is another sound step, because it shortens every call that needs it. Agents are then free to spend their time on the complex and sensitive contacts that need a person.

Failure demand falls when it is owned by leaders across the organisation, not only by the contact centre. Report it next to service level and volumes, so it is discussed in the same meetings. Give each main cause an owner outside the contact centre where that is where the cause lives, agree what will change and by when, and check the codes and samples afterwards to see whether the contacts have gone.

Keep the analysis going. New products, price changes, system changes and new regulations all create new failure demand, often quickly. A small monthly sample, reviewed with team leaders and agents, keeps the picture current and shows agents that what they hear every day is being acted on. Over time, the contact centre becomes the organisation's best source of evidence about what is not working for customers, which is a better role than absorbing the consequences.