A couple of years ago the folks at McKinsey published a report called “Big data: The next frontier for innovation, competition, and productivity”. You can find it online pretty easily. Among other things, they wrote:
“Inventory management. With the additional detail offered by advanced analytics mining multiple datasets, big data can continue to improve retailers’ inventory management, Best-in-class inventory management provides full transparency at the SKU level, while bar code systems linked to automated replenishment processes reduce the incidents of running out of stock. Leading retailers are improving stock forecasting by combining multiple datasets such as sales histories, weather predictions, and seasonal sales cycles. Together, improved inventory management allows retailers to hold a lower level of stock because supplies are coupled much more tightly with demand signals, while reducing the number of sales lost because of merchandise stock-outs.”
I would be interested – intrigued, for that matter – to know just who these ”leading retailers” are - the ones that have improved their forecasting and thereby have simultaneously reduced inventory and lost sales due to stock-outs. It isn’t Walmart, not Walgreens or Target either. In fact, I have seen absolutely no evidence to support the theory of big data, complex algorithms and massive computers having any great impact whatsoever on the performance of big manufacturers or big retailers.
The fact is that there are no facts – but there is a rather vociferous mob of Kool Aid drinkers who love the big theory. The justification – the stats – supporting the nonsense the likes of McKinsey is that computers are better forecasters than humans. What they fail to see is that proving a computer is a better forecaster than a person is like proving that Chanel Number 5 makes pigs smell better than they do when you put Calvin Klein’s Obsession on them. Why put perfume on pigs in the first place?
No one drinks longer, deeper and more delusionarily at the computers-are-so-much-better-than-people-at-everything Kool Aid trough than Amazon and their latest is downright loony: Anticipatory Shipping. Anticipatory shipping is basically shipping something to you that you haven’t ordered based on a forecast that you will, in fact, order the item before it gets to you. When you do – voila! – the lead time is near nothing, and all thanks to the mind-blowing algorithms embedded in Amazon’s mighty computers.
What are the odds of that happening?
You can read all about this in Amazon’s patent application. Yep, you read that right. They are so in love with the idea they applied for a patent. And the government is so clueless, they granted the patent. All it is, however, is forecasting in brain-numbing detail and then using trucks and UPS facilities as extremely high cost and extremely inefficient warehouses, then using more and bigger computers to keep track of and sort out all of the inventory they have strewn across the countryside. How forecasting demand and then putting the stuff projected to sell in inventory deserves a patent is beyond me … but they did it. They didn’t just put perfume on the forecasting pig – they marinated that pig to the point is skin is permanently wrinkled.
Forecasting is a useful tool for planning capacity, for projecting gross demand as a gross volume indicator for negotiating and managing blanket purchase orders, and for selective inventory building for companies with a particularly wicked seasonal curve. Making, buying and stocking specific items – a la 1961 vintage MRP thinking – push production – is a strategy for losers and the evidence of that is overwhelming.
There is no substitute for relentlessly shortening cycle times and lead times. Delivery excellence comes only from flexibility – being able to make and ship whatever is ordered. Forecasting, however, assumes inventory is necessary; and inventory exists solely to protect poorly conceived and poorly executed processes – pigs. And that is why any discussion of whether people or computers are better forecasters is such a meaningless debate.