A contact centre crystal ball

Andre le Roux

Mathematical models provide the next best thing to certainty. They alert, evaluate operational risk, and propose business and resourcing solutions. And now the best of these models also determine expected experience delivery.

The nature of customer experience metrics implies that they can be captured in a contact centre plan as a time-series metric. Such metrics exhibit seasonality. This further implies that they can be forecasted using the same sorts of methodologies that the best planners use to forecast call volumes and sick time. For instance, analysts forecasting metrics that exhibit a time-series trend might utilise exponential smoothing or regression models. If the metric displays seasonality, it might make sense for the analyst to use methods like Holt-Winters.

The better news is that these custom metrics for experience delivery can be modelled, predicted, and used in the normal course of strategic planning. For each resource plan, these metrics can determine the expected week-over-week service level, abandon rate, occupancy, cost, revenue, and average speed of answer. They can also determine customer service metrics like Net Promoter Score or agent quality score. Therefore, there's no reason that expected customer experience isn't part of the regular capacity plan, forecasted along with all other service metrics.

Strategic planning systems

Simulation and mathematical modelling systems - aka strategic planning systems - automatically develop forecasts and resource plans for multichannel and multi-skill contact centre operations. Strategic planning models such as those used in Interaction Decisions from Interactive Intelligence have some terrific advantages over home-grown spreadsheets.

First, strategic planning models work in two directions. They evaluate for any week-over-week scenario the service, revenues, costs, and customer experience scores expected under any planning scenario. In the other direction, they determine the least-cost staff plan required to hit the service goals associated with any scenario.

Second, they are proven accurate. The best planning systems include a validation step to prove that, for each of the contact types and contact centres in a network, the model is accurate when compared to real contact centre data. This is not easy, as every contact centre and contact type is truly different. The models must be smart enough to consider these differences and be recalibrated as the operation changes.

Third, systems for strategic planning must be fast. It helps no one if the models are too slow for decision makers. The best systems can be run interactively, requiring only minutes to evaluate any scenario. In all, speed, accuracy, and breadth of analytics enable a different sort of decision-making process.

In the past with a static, spreadsheet-based planning process, decision-making and analytics were only passing acquaintances. Analysts had little time available to answer executive what-ifs. Using advanced modelling, however, an analyst can now answer the executive's query in real time, interactively.

It's a process that enables a different relationship with decision making, in which all major decisions are vetted and all repercussions of resource decisions are known, including the expected customer experience. For the contact centre executive, strategic planning systems are the closest thing yet to a crystal ball.