While automated forecasting is hardly in its infancy – its existence was celebrated well over two decades ago – impressive leaps and bounds have been made, even in just the last couple of years. Linked to how customer service has evolved and adapted with more and more expectations placed on the contact center and agent, forecasting tools have also had develop and expand in their capabilities, ‘good enough’ is no longer good enough.
A wealth of of detailed data – Advances made in forecasting software now make it possible to take into account not just the past few weeks of data, but rather years of detailed data. This obviously results in much greater forecasting accuracy. Correlated forecasting for specific events, frequently the cause of workload fluctuations, can now also be performed. Planners can easily note what is a yearly, repeatable event or a one-off anomaly in the data and forecasting algorithms will take this into account when forecasting for the future.
Intraday forecasting – Forecasting, recognized as the first and most crucial step in workforce management (WFM) – not to mention the single, most effective cost-cutting tool used in contact centers, back offices and retail outlets today – can now be performed not only for days or weeks ahead of time, but also right up to the minute. That’s just what the WFM doctors call for – an instant diagnosis for essential and effective intra-day remedies.
Multi-skilled forecasting – Going one step further, today’s sophisticated algorithms are equally able to support an integrated method, directly calculating requirements in a multi-skilled environment. This eliminates the repetition of analytical simulations for each different channel. Now, single, forecasted requirement sets for all inter-woven skilled activities, irrespective of the channel can be produced.
Multi-skilled overflow – Bringing secondary skills into the equation as such and accounting for overflow to available, multi-skilled agents helps reduce the tendency to over-staff and better meet service-level targets.
Out with averaging, in with time-series model – Taking over from the inflexibility of weighted-history averaging, time-series models are offering the possibility for forecasters to properly compare apples to apples, and months to months. The latest time-series models are serving as the basis for most automated forecasting solutions today. It’s no wonder when you consider their capabilities of producing forecasts with 95 percent accuracy or greater – and at lightening speed.
Crystall-ball-gazing pundits predict that soon, you’ll say goodbye to Erlang C and weighted-history forecasts.
Others predict that, the word “forecasting” will no longer be used in the years to come: “Forecasting” as currently envisaged, will disappear into business-as-usual. How? Get answers to these questions and more by downloading our latest whitepaper on Forecasting.
Rest assured though, the role of resource planners will remain essential in the years to come. The continual need for experts with deep knowledge and awareness of changing customer demands, and the organization’s ability to respond, will never go out of fashion; especially once you find that perfect balance between the workload and staffing.