Are your demand forecasts good enough?
Against most forecasts, post-lockdown has seen shoppers returning to stores and leaving online. How well did you forecast this? What does your forecasting model say about how long this is going to last and what the future shifts in channel might be?
The retail market was already tight. With a recession on the way, it will tighten further. The industry needs all the help it can get to manage more tightly, forecast better, and generally optimize offerings. There are lots of tech tools today that can enhance decision-making in all these areas. But retailers are struggling to fully appreciate the potential benefits, embrace these tech tools and integrate them into the heart of their businesses.
Data science, AI, machine learning - countless articles have been written over the last couple of years about these incredible new tools and their benefits. However, these aren’t silver bullets that will provide powerful insights at a click of a button; leveraging these benefits depends on the skills of the team deploying the tools.
Before companies can leverage the unparalleled accuracy and efficiency of data science, AI, and machine learning, they must first understand how to power them.
Data Science and SMARTS are here to enhance your team
SMARTS (the collective term for AI, ML engines) are something of a black box; the mathematical calculations they produce are complex and only the answer is provided. Knowledge of how the calculation works, which data it inputs (and the quality of this data), and the end impact of changing something is vital and a skill that has never previously been required within a retail organization.
Data science is commonly used to create across many target operating model pillars – elasticities for pricing, seasonality profiles in forecasting & replenishment, and ideal range structure for planning. It improves with every data point of history it receives but without an AI/machine learning (self-teaching) element to it, the coefficients created require manual updating as accuracy reduces over time. Coefficients, themselves, are the values associated with variables that use a multiplicative effect (uplift/dampener) on demand. For example, there are price coefficients and as price changes, the demand element gets multiplied by the coefficient associated with the price change amount.
It’s not just about the tools, it’s about the people using them
Data science, AI, and machine learning tools achieve very little on their own. To unlock their full benefits, you must have a team skilled in deploying them. Having a data science team in-house might be a wise investment to manage the above. SMART development could then be made without any reliance on external contractors who may not understand the nuance challenges you face as a business, or be able to work to the timelines you require. However, building an internal team of highly skilled data scientists is an expensive undertaking, and employment without value-adding work can cripple a company’s cash flow. External contractors or consultancies on the other hand allow companies to access specialist people during specified times. Whilst the initial investment is higher, costs are limited to the duration of the work, and benefit calculations are provided to ensure value for investment.
As data science, AI and machine learning solutions continue to grow and power retail decisions, it’s critical that businesses assess what capabilities they have in place to maximize these tools and technologies.
To gain the optimal benefits, the quality of the data they rely on and the skills of the staff involved must both be high. TPC continues to work with major retailers around the world to define strategies and both select and assist with the implementation of innovative tools that enhance their trading performance. To learn more, please visit our website, find us on LinkedIn or call for an informal chat.