Innovation may be the one word used in consumer goods industry that is met with both equal excitement and disdain. On one hand it is an inspiring banner by which marketing, sales and leadership can get behind and begin believing in increased market share, infringing on white spaces not pioneered by the competition, and blasting through business boundaries to forge new revenue streams. Operations, however, view innovation in a much less favorable light. Supply Chain Managers and Site Directors all across the world view Innovation the company’s one stop shop for a unique bonus killing piece of inventory known as the Slow Moving / Obsolete Stocks.
Ask anyone who has worked in a beverage company and they will have excel sheet upon excel sheet of bygone warehouses filled with cherry-flavored-this and zero-sugar-that. The cherished battle cry of ‘move fast and break things’ often doesn’t resonate with an operations team, simply because most Fortune 500 conglomerates don’t move very fast and are ironically very bad and breaking anything. Instead they move slowly and sit on un used inventory until a marketing intern can find a way to swindle distributors into picking up SKUs that will move even slower than the companies that produced them.
However burdensome innovation may be to operations, it is still core factor in company growth. No company can compete without consistently improving it’s portfolio and taking risks. As such, the real question is not whether or not innovate, but how to do so efficiently. How does one differentiate between a slow start and a slob? when does that new flavor go from being the next big thing to the next big warehouse glut?
The answer may very well lie classification. Innovation is not a black box, in fact, the features that could help define a successful innovation are often heavily monitored metrics by the very companies that have launched them.
Seasonality, weather, current penetration of market share, current price, price of competition, rate of sale of competition and rate of sale of innovation could all reveal hidden patterns that innovation managers could use to pre-empt failure or identify success. Millions of dollars could be saved in marketing expenses, warehousing, and clogged distribution networks. More importantly, the same funds saved on failed innovations could be used to cast a wider net, allowing to move on from enough bad projects in order to find potentially profitable ones.
A properly fed XG Boosted Random Forest could hypothetically identify a bad project from a good one, by identifying key markers of success such as sales/USD invested, ROS at a specific point in time and market share gained a ta specific point in time.
The future of efficient innovations in consumer goods will undoubtably rest on better data, but the most profitable innovators will definitely be those capable of capitalizing on data science, machine learning and forests of failure.
Oracle Operations 