Wallboards nobody watches: designing real-time contact center reporting

Most contact center wallboards show about fourteen numbers, get glanced at twice a day, and change nobody’s behavior. That is not a display problem, it is a purpose problem. A real-time view exists to change a decision in the next few minutes. Anything on it that cannot do that belongs in a report instead.

The test for every tile

Ask what someone would do differently if the number moved. If the honest answer is “nothing until the weekly review”, it is not a real-time metric.

Applied strictly, this removes most of what is usually displayed. Calls answered today, average handle time month to date, abandonment this week — all useful, none actionable inside five minutes.

What survives the test

• Contacts waiting, now.The single most actionable number in a contact center.
• Longest current wait.More useful than the average, because the average is comfortable while one person has been holding for nine minutes.
• Agents available versus agents on contacts.The supply side of the same equation.
• Agents in an unexpected state.Not a surveillance tool — a count of people who are unavailable when the queue is deep is a prompt to go and find out why.
• Service level so far today, against target.The one cumulative figure that earns its place, because it tells the floor whether the day is still recoverable.

That is five numbers. Five is close to the right answer for a wall-mounted display.

Design so it can be read at eight meters

A wallboard is read from across a room, in peripheral vision, by someone doing something else. That constrains it hard:

• Large figures, few of them. If it needs to be read closely it will not be read.
• Change state by color and by something else — a position, a size, a label. Roughly one in twelve men has some degree of color vision deficiency, and a red-versus-green wallboard is exactly the case that fails.
• No animation except when something changes. Constant movement becomes invisible within a day.
• One target line per figure, drawn, so “good” is spatial rather than remembered.

Different views for different people

The floor wallboard and the supervisor’s screen are not the same artefact and should not be the same layout.

• The floor.The queue and the day, in five tiles, visible at a glance.
• The supervisor.Per-agent state, per-queue depth, and the ability to act: move someone, open a skill, call a break back.
• The agent.Almost nothing — their own state and how deep their queue is. More than that competes with the customer for attention.

Alerts instead of watching

The honest conclusion of “nobody watches the wallboard” is that watching is the wrong mechanism. A threshold that raises an alert — longest wait over four minutes, available agents under three — reaches someone whether or not they happened to look up.

Keep the thresholds few and set them where they genuinely require action, or they become the same background noise the display was.

Then check it is telling the truth

Real-time views are usually assembled from a different path than the historical reports, and the two drift. It is worth reconciling them occasionally — the number on the wall and the number in the report should agree at the end of the day, and where they do not, the wall is the one people trust.

A short specification

1. Five tiles, chosen by the “what would change” test.
2. Target drawn on each, not remembered.
3. State signaled by more than color.
4. Two or three alert thresholds that reach a person.
5. Separate, denser views for supervisors.
6. Reconciled against the historical report once a month.

Our call center solutions drive real-time views and threshold alerts from the same event stream as the historical reporting, which is what keeps the wall and the report from disagreeing. If your current display is a fourteen-tile screen nobody looks at, it is worth seeing what five would look like.