The Weather Channel’s Jim Cantore has never killed anyone on national television. But we’re pretty sure we witnessed a moment when he was tempted to back in 2012.
As Hurricane Sandy’s dangerous storm surge started to rise around him during a live broadcast from New York’s Battery Park, he mentioned that he was standing in water that was several feet deep — only to be immediately corrected by an authoritative voice from the studio.
“Jim,” the voice insisted. “Our computer models show that the water level is actually much lower than that.”
Cantore’s experience is a great example of what can happen when people place blind trust in their technologies and algorithms. It’s also a cautionary tale for many supply chain professionals, because even the most advanced IT and cloud solutions have the potential to steer us wrong if the data is inaccurate or incomplete.
But you probably knew that already, didn’t you?
What’s less obvious is how you can do a better job of correcting this issue, which is why we’d like to share the following helpful acronym to help get your data accuracy efforts on the right track.
Invest time up-front to create standard terms and definitions for all of your supply chain data points. (For example, how should an exception be defined? What does on-time mean?) This will help prevent many of the small variations in nomenclature that can lead to major data discrepancies.
Nip obvious data errors in the bud. If an operator knows that some key shipping parameter has changed since a pallet, shipment or parcel was entered into the system, he or she should be trained and incentivized to stop and make the update immediately rather than assuming that someone will correct it later down the line. The longer such a mistake — or any other one such as a typo — is allowed to remain, the more off-base your data points are likely to become.
Tie data to its source. Clear attribution will help you do a better job of assessing everything from each source’s validity to the age and relative cleanness of each data point provided. Just as important, it will inspire the parties that are responsible for populating your data to bring their “A” game.
Ensure your telemetry devices are in good working order. Always double-check the health of any ELD or other pieces of hardware that fuel your data collection efforts, because even the most rugged of these devices will occasionally break or experience a glitch — and compromise your data integrity accordingly.
Give your operations people a seat at the planning table. By letting your end users define the business problem they want you to solve, you’ll prevent your IT department from getting too granular about minor data points that don’t add value — and allow them to spend more time collecting, analyzing and scrubbing the points that do.
Recognize and correct obvious red flags. As you analyze your data, be on the lookout for unusual sudden spikes or drops that don’t appear to be tied to any major changes in your business (or other anomalies like a really substandard score for a carrier that has historically set the bar for excellence). These could indicate a serious inputting error that needs to be addressed or a new standard of measurement that needs to be revisited.
Involve all of your company’s personnel, not just your senior executives or facility managers. At the end of the day, good data collection is a team sport that’s reliant on every player doing his or her part — including the all-important people who drive your trucks, operate your forklifts and work in your DCs. If they buy into your vision, they’ll be more willing to do much of the hard work that’s required to capture your grassroots data correctly every time. If not, you’ll find some of your best-laid plans stymied by partial compliance or incorrect usage.
Take the time to collect, vet and organize your data correctly. Although no one truly wants their data collection efforts to move at a glacial pace, “definitive” numbers that are produced too quickly (and without the benefit of any checks and balances to ensure their accuracy) could be even more detrimental, because all they’ll do is help you make bad supply chain decisions faster. Don’t produce data quickly just to get the check in the block.
Your team’s intuition and experience are powerful tools. Use and respect them. Much like Jim Cantore was actually the better judge of how deep Hurricane Sandy’s storm surge was because — wait for it — he was actually standing in it, your IT experts and engineers often have a depth of firsthand, non-technical experience that can be just as valuable as any sophisticated algorithm. Empower them to bring this practical knowledge (such as knowing that certain carriers’ data needs to be fact checked more frequently than others) to bear as they prepare your data for the next step. Your ultimate results will be better and stronger for it.