Buyers and merchandisers need to make hundreds of individual decisions a week, the majority of which have a significant impact to the business, supply chain, and bottom line. Whilst businesses are rich in data, storing years of sales, stock, and price information down to granular levels of detail (e.g. SKU by location by week), this level is far too vast for humans to analyse alone. It is impossible for us to discern patterns from large quantities of data without comparable support. Data science and algorithms are incredibly adept at making sense of this data both efficiently and effectively and are best positioned to assist retailers in making business-critical decisions.
Human nature: the bias problem
It has been repeatedly proven that humans are naturally risk averse, and tend to be very poor at forecasting risk and reward over time (hence why there aren’t many more Warren Buffetts). By nature, people are also highly inconsistent, with even the day of the week proving to affect how much risk people at work are willing to take. By not supporting teams with the right tools and analytics, there’s a significant amount of benefit to retailers being left on the table.
The Bell Shape: Humans tend to congregate in the middle when making pricing decisions, opting for “safer” discounts away from the minimum (e.g. 20% off) and maximum (e.g. 60% off)
Data science: laying the foundations
Choices that are repeatable and data-led, such as pricing, allocation, and replenishment, are perfect candidates for algorithms to solve, with increased consistency and far less bias than humans. The accuracy of these decisions is then vastly improved, with data science being able to identify demand signals much better.
This optimised approach is consistently more accurate in forecasting risk and reward than humans are and is, therefore, more assertive in making markedly different choices. Data science-led forecasting and optimisation results in accuracy up to 7 times better than people, unlocking significant benefits to business, from improved margin in pricing decisions to improved inventory management and sell-through via optimised forecasting, replenishment, and allocation.
The U Shape: Data science-led decisions are much more assertive, knowing which products can be trusted to trade well at a lower discount and which products need aggressive discounting at the headline message
The best of both worlds
The computational power of data analytics and data science will always have a quantitative advantage over the amount of data humans are able to process. However, unless you have access to some of the most advanced AI, humans will always have an advantage on more nuanced decisions in creative and strategic spaces.
Data science can recommend the best options to humans who can then add their additional business, customer, and market knowledge to those recommendations to determine the best choice. At TPC, we know the best of both worlds is where data science empowers humans to make improved decisions. As the competitive market intensifies, harnessing data-driven decision-making should be seen as a necessity, not a luxury.