There's a lot of talk about the adoption of data science technologies over the years and how it's the next game-changer. However, a 2019 article published by the Harvard Business Review highlights that the uptake may not be as quick as we first thought.
The article notes that of the 65 Fortune 1000 companies surveyed, around 72% have yet to "forge a data culture." These are companies with significant budgets set aside — 55% of the companies surveyed have $50 million or more to focus on data science projects.
So, if the big end of town, with their substantial budgets, is slow to adopt data science, where does it leave the rest of the business world?
Having worked with a variety of companies, it has become apparent there is a large capability gap preventing companies from advancing. The typical business is standing on the edge of a chasm. Where they currently stand, they are comfortable and getting okay results. But on the other side of the chasm is the promised land. The problem is that the ticket to get across is prohibitively expensive, and there is uncertainty about whether the grass will be greener on the other side.
What's needed is a tool that democratises data science — something that makes it realisable for the general population.
What's needed in the market is a solution with four key characteristics: it must be robust, repeatable, adaptable, and understandable.
While every company is unique, every company sells something, and there are mechanisms that drive that demand (such as pricing, promotion, advertising), as well as external factors that affect demand (such as weather and economic factors).
By having a robust, repeatable, understandable, and adaptable algorithm, we can begin to explore and model how these various factors affect demand.
The current raft of older-style demand planning tools have understandable and repeatable algorithms, but they lack the adaptability and robustness to leverage the variety of data available. They use one data set: the demand history. In contrast, CauSelf looks for patterns in the demand history and assesses other factors that drive and affect demand (like price, advertising, weather, etc.).
Using a greater data set not only helps to improve accuracy and the quality of the demand planning process, but it also means users can gather insights that were previously not easily found — such as understanding what pricing works best and what promotional mechanisms deliver better value.
The newer analytical tools have very sophisticated and adaptable algorithms and can analyse complex and large data sets. However, they are often heavily data-dependent, which makes repeatability difficult or time-consuming. The complexity of their calculations also means that users don't understand the logic, which can impede user acceptance.
With CauSelf, we have designed a robust, repeatable, adaptable, and understandable algorithm that answers the common question of what drives demand. Having a standardised and robust algorithm doesn't mean that we need to sacrifice performance.
Over the last few months, CauSelf has worked with several different businesses and confirmed that the solution delivers strong results. In one exercise, we compared the results from CauSelf, achieved with 4 hours of effort, to bespoke models generated by a data scientist as part of a 3-month engagement. CauSelf had a more accurate forecast for 66% of the 85 products analysed.
We believe that the CauSelf solution can help you bridge that chasm and make it to the promised land.
Book a free demo →