Get these right before tapping point-of-sale for supply chain changes.
DON'T Overanalyze factors you can't control
If the root cause of a problem is controlled by others, such as distributors or retailers, determine whether you can change the business process or behavior. Demand-signal analysis can easily spot poor promotional execution, for instance, but it doesn't help a manufacturer if its field staff can't work with retail managers to straighten out stocking and merchandising errors.
DO Pick the big opportunities
Supply chain optimization, logistics planning, retail category management, promotional execution, shrinkage control, stock-out forecasting--these all are good prospects to improve through demand-signal analysis. But start with the problems that can deliver the biggest payoffs. That typically means improving promotion performance and reducing stock-outs.
DON'T Reinvent the wheel
Collecting, normalizing, and cleansing high-volume demand-signal data is costly and time consuming, particularly when a manufacturer is dealing with multiple retailers, all of which likely send data in different formats. Look for data networks and aggregation specialists that work with key retailers and leading manufacturers. Wal-Mart's Retail Link, Retail Solutions' Demand Signal Repository, and Vision Chain's Demand Driven Supply Network are leading sources of demand data.
DO Use demand data to spot "phantom inventory"
Consider this scenario: Shipment data suggests an item should be in stock, but demand-signal data shows it's not selling. Is the product a dud, or is it actually out of stock? Stocking errors, breakage, theft, and bad scans can lead to "phantom inventory" that's really not in the store. Some analytics software can compare sales histories with demand data to spot suspect conditions.
DON'T Set off false alarms
Demand-signal analysis can be used to predict and prevent problems, but watch out for false alerts. Say a manufacturer's order history shows suntan lotion should be selling at a store in Florida, but point-of-sale data shows no scans. That may not be a stock-out situation. Retail Solutions does demographic or geospecific store cluster analysis to spot factors such as bad weather. Make sure your approach includes reality checks before triggering alerts or, worse, automatic replenishment.