In my last post I introduced a vision for the Smart Trash that would automatically identify the items you are throwing away. What would you do with the data collected? The waste management company may not have much use for the data, but manufacturers and retailers who are trying to predict what consumers are going to buy next would find it very valuable.
How would the smart trash can work? There are a couple of different options. Version one (circa 2017) would probably rely on the use of RFID tags and readers. If manufacturers put RFID tags on each of the items you purchased then the smart trashcan would be able to identify them automatically. Version two (circa 2018) might add a camera to the lid. As items are being disposed of the camera would automatically recognize the item using “visual search” technology found in Google Goggles. Perhaps the smart trash can might have multiple cameras at different depths that could see through trash bag liners to identify items at rest. Version three (circa 2020) might be more advanced with capabilities to identify items based upon smell. Perhaps, the trash can would be fitted with sensors that can detect odors and identify items based upon their chemical composition.
You might be wondering what data retailers and manufacturers use to forecast demand today and whether smart trashcans would provide an improvement. Today, the primary data used for forecasting demand is the information about what shoppers are buying at individual stores or what is called “Point-of-Sale” data. Every night retailers and manufacturers run reports to understand how many of each item was sold in each store. They then try to guesstimate how much inventory they have on hand and whether or not they are going to run out of stock in the coming days (weeks or months). If they are running low on inventory then will need to issue a replenishment order.
Would trash can data provide better insights than Point of Sale data? This begs a good question. What provides better insights into future sales – what people are buying or what they are throwing away? Is monitoring Point-of-Sale data a better approach than monitoring waste?
Let’s first think about items that are regularly purchased – batteries, diapers, detergent, shampoo, soda, milk, bread and salty snacks. I would argue that monitoring consumption (via trashcans) of these repeat purchases is a better indicator of near-term demand. If someone throws out a milk container they are very likely going to buy a new one in the next 24 hours. In many cases, the disposal of an item after it is consumed is the event that triggers the need to buy another one.
But what about items which are not consistent, repeat purchases? Examples might include toys, electronics, clothing, shoes, etc. For these inconsistent purchases you might question the validity of the correlation between waste patterns and future purchases. Just because you throw something away doesn’t mean that you are going to purchase it again – immediately or ever.
The value of using the trash data is clear for groceries and regular purchases. Will Nestle, Procter & Gamble and Tesco begin giving away free kitchen trash cans to consumers just to collect data and be optimally positioned for replenishment orders? Or even better what if your smart trash can was linked to your online grocery account? Items detected in your trash can (or recycling bin) could be automatically identified then transmitted to the garbage truck upon pickup at your house. A replenishment algorithm could review your list of “always in stock” items to determine if the item should be replaced immediately. If yes, then a home delivery provider might visit a few hours later to drop off new supplies on your doorstep. Amazon Fresh be extended to include Amazon Trash. Walmart might buy a waste management company. The Smart Trash Can could create a myriad of new opportunities in the supply chain.