Companies pursue supply chain speed as a competitive advantage but often wreak havoc. Instant communication and voracious data transfer can easily overwhelm today’s managers. Trying to manage with speed is like texting while driving—you know it’s dangerous but there’s a strong urge to respond while on the go. There are many examples of supply chain speed gone wrong. Starbucks’ latest problems with mobile ordering are just one.1 Texting while driving is never recommended. Managing for speed, on the other hand, can provide competitive advantage if the pitfalls are avoided.
One major pitfall is the thinking that more data and more data faster leads to better performance. This a major theme of many software vendors—“Buy our software and you’ll be able to see data in real time, respond faster and make more money.” Don’t get me wrong, there are many instances where real time access to data is extremely helpful. The ability to get instant feedback on tornadoes and hurricanes has saved countless lives. That’s not what I mean.
The problem comes when companies feed data to their operations faster than their operations can respond. This is fairly common with supply chain ERP systems. I have been at companies where it was a point of pride that they regenerated their ERP system once or twice a day. This is great for getting new sales orders into the system and if the operation's cycle time2 is less than half a day. Otherwise, it’s madness to expect any supply chain's operations to successfully respond to all real time changes. Feeding data to a system faster than the system can respond leads to chaos and expediting—costly wastes of resources. For a classic illustration, see what happens when demand signals exceed Lucy’s and Ethel’s capacity to respond. As Lucy astutely observes, "I think we're fighting a losing battle." Starbucks barristas know how she feels.
Supply chain speed is subject to the laws of nature and operations science. Nature won’t write you a ticket if you exceed your operations system’s speed limit but the outcome will be costly. “Starbucks reported that transactions, a measure of customer traffic, dropped 2% in the most recent quarter, in large part due to problems caused by mobile ordering.3”
When trying to synchronize demand and supply in presence of variability, buffers will develop. There are only three buffers: capacity, inventory and time. Here are Starbucks’ options for providing a speedy response:
- Capacity – The Google graph shows frequency of visits on Friday to my local Starbucks—not my visits, all Google-tracked visits. Starbucks can add more baristas, and maybe espresso machines, to handle the 7AM crush and deliver within the prescribed three minute wait time. The problem is that having enough capacity to meet peak demand between 6AM and 10AM means that capacity is sorely underutilized for the rest of the 5AM to 1PM early shift.
- Inventory – probably not an option, who wants a cup of coffee that has been made an hour or so before the 7AM rush? It violates the business proposition of a fresh cup of coffee.
- Time – With all that high-powered technology, why not let customers know how long they will wait? Customers place an order, get a confirmation and a projected wait time. They may not be thrilled but at least they won’t be unpleasantly surprised to wait 10 minutes when they thought they were going to wait for three. Customers may be willing to wait longer if they can trust that 10 minutes means 10 minutes. Also, use of the time buffer usually reduces the need to add expensive capacity buffers.
- Reducing variability – when variability is reduced, less buffering is required—this is a law of nature. This is not a great option for Starbucks since customers won’t order just one type of coffee. One possible option is to shape demand by making complex orders more expensive. Anyone for a Caramel Macchiato, Venti, Skim, Extra Shot, Extra-Hot, Extra-Whip, Sugar-Free coffee?—that’s $10 between 7AM and 9AM on Fridays. Less complex orders during peak hours should make the baristas more efficient and increase output. The tradeoffs between revenue per cup, revenue per hour and customer goodwill would have to be assessed.
Starbucks will probably work its way out of this problem. It focused too much on speed and mobile orders without fully understanding the consequences. The unfortunate aspect is that the problem is entirely predictable. Focusing on supply chain speed and ignoring the laws of operations science is like always texting while driving. It might work for a while but eventually bad things happen. Don’t react after things get bad and, maybe, irreparable. Apply the laws of operations science to manage supply chain speed predictably and profitably. See Factory Physics for Managers for operations science concepts you can use now.—ESP
1 Kate Taylor. “One of Starbucks' biggest strengths is becoming a huge problem for the chain.” Business Insider. January 27, 2017. http://www.businessinsider.com/starbucks-mobile-ordering-problems-2017-1
2 “cycle time” here means the time to complete production from start to finish, not the time at a single machine or at a single operation.
3 Kate Taylor. “One of Starbucks' biggest strengths is becoming a huge problem for the chain.” Business Insider. January 27, 2017. http://www.businessinsider.com/starbucks-mobile-ordering-problems-2017-1