Don't be left in the dark
Updated: Sep 11, 2019
With the resurgence of rolling blackouts sweeping across South Africa, load-shedding has once again become the default conversation starter, a hot news topic and the subject of much amusement across social media.
While we can all commiserate with what it feels like to be on the wrong end of load-shedding with its disruption to work and life schedules, it was only recently that I considered the flip-side of what load-shedding means for Eskom, the power utility. Instead of trying to increase revenue in order to grow, Eskom is forced to reduce its income generation by restricting the sale of its primary product across the country. This self-induced debacle must have significant financial impact on an already struggling enterprise.
While this example of lost opportunity is hard to miss, it is useful to consider that many organisations that do not fully leverage their reach, efficiency or productivity are in essence load-shedding themselves. Using advanced data analysis to gain greater understanding of the insights that lie hidden in the streams of data oft collected but not always utilised, could shed light on opportunities within.
Demand forecasting, is used to help ensure electricity supply meets demand. This tool has multiple uses in almost every industry whether it is to forecast revenue, product inventory, call centre volumes or website traffic (think airline sales or Black Friday mania). Automatic forecasting solutions can indeed predict generalised trends sufficiently. However, off-the-shelf solutions can be highly inflexible making it difficult to incorporate additional information (e.g. holidays, weather, irregular events) and result in limited prediction accuracy. While increasing prediction accuracy by 1-2% or even 10% might sound trivial, in certain instances this can result in optimizations that save or generate hundreds of thousands of dollars.
Photo by Devin Avery on Unsplash