Cycle Counting Best Practices While You Still Ship

Bin A-12 said 40 lamps in the system. The shelf had 36. Somebody posted a minus 4 before lunch. A picker was already walking over with an order for 4 that had not been confirmed. By 2 p.m. the shelf and the screen were both wrong, in opposite directions. Cycle counting best practices are how you check a slice of the building without closing it, and how you refuse to "fix" a number that a late scan already explains. What are cycle counting best practices? Count part of the stock on a schedule, while the rest of the building keeps shipping, until everything has been counted over a period of time. The GAO's inventory count guide describes that approach, and it contrasts it with a wall-to-wall count of everything at one moment. The practices that make the slice honest are plain. Hide the system quantity until the shelf number is written down. Freeze picks and receipts for that bin, not the whole site. Look for an open transaction before you adjust. A second person recounts a gap, and the reason gets written down. A full shutdown still has a place. It is the wrong weekly habit once orders are due the same day. How often should you cycle count? The GAO guide says leading operations counted a portion daily, weekly, or monthly until the whole inventory was covered, and they aimed to get every item inside the year. Some pulled high-dollar items more often than the rest. Use that as the shape, not as a quota copied from a poster. An example cadence, which you should change if your misses cluster somewhere else: Items Example cadence Why Fast or high-value SKUs Every month A miss here is cash and a canceled order The middle of the catalog Every quarter Often enough to catch a process slip Slow SKUs Twice a year Still inside a year, without burning the count team A bin that missed last count Next week, then back to its class Repeating gaps are a process, not bad luck Count the problem locations more than the quiet ones. A perfect bin counted every week teaches you nothing. A bin that was short twice is the one that should come back early. Why should the counter not see the system quantity? If the screen says 40, people find 40. They count two cartons, remember the number, and write 40 without opening the second carton. That is a confirmation, not a count. A blind count hides the book quantity until the physical number is in. The counter writes 36. Only then do you show 40. The 4-unit gap is now a fact to explain, not a number someone talked themselves into. Have someone other than the person who just picked that bin do the count when you can. The picker who set the carton aside "for a minute" should not be the one certifying the shelf. What do you check before you adjust? Freeze bin A-12 for new picks and new putaway. Do not freeze the building. Then look for quantity that is real but not on the shelf yet. On this example, the system says 40 and the blind count says 36. An open pick for 4, not confirmed, explains the gap. The lamps are on a cart. Confirm the pick. The book becomes 36, the shelf is 36, and you adjust nothing. If there is no open pick, no unposted receipt, and no unit-of-measure mix (a case of 4 keyed as 4 eaches, or the reverse), send a second person to recount blind. Adjust only after that recount agrees, and write the reason: miss-pick, missed receipt, damage, or unknown. What you find What you post Open pick for the same gap Confirm the pick. No inventory adjustment Receipt scanned to the wrong bin Move the units. No write-off Case versus each Fix the unit, then recount Second blind count matches the first, and nothing explains it Adjust, with a reason code Second count matches the system Keep the book quantity. Coach the first count Posting minus 4 in the first ten minutes deletes units that were on a cart. The next pick shorts a customer, and the accuracy chart gets worse because you "fixed" it. How do you measure count accuracy? Federal Prison Industries defines accuracy as the number of correct counts divided by the total number of item counts, and that agency's goal is 95 percent or greater. It is their bar, not a law for every stockroom. The definition is still the right one to copy: exact matches over counts, not "close enough" averaged into a percent. Example from a single week: 50 counts, 47 exact. Accuracy is 47 divided by 50, which is 94 percent. That week misses the agency's 95 percent goal. Do not hide the 3 misses inside a dollar variance that looks small. A $2 gap on a fastener and a missing $200 lamp are both misses. Track them, then sort the reasons. Damaged units are not a quiet pass either. If 2 of the 36 are cracked, the shelf count for sellable stock is 34, and the 2 belong in a damage code, not in the number you will pick tomorrow. Leaving them in A-12 as good stock is how the next order fails at pack-out. A photo on the adjustment beats arguing from memory a week later. Zero bins belong in the program. When a pick empties A-12, the picker confirms it is actually empty. That check is cheap, and it catches the unit that was supposed to be the last one and was not. Count during a quiet patch of the shift if you can. A blind count while three pickers pull from the same bay will disagree with the book for reasons that expire ten minutes later. The GAO locations that kept this up did it as a repeating slice, not as a fire drill after a bad month. Warehouse management in OneChannelAdmin can hold the count against the bin without a shutdown. The book quantity you are testing lives in inventory management. Open orders that explain a gap are in order fulfillment. If you later count with a reader instead of a clipboard, that is RFID inventory. How do you run one cycle count? Choose a short list. One aisle, or the fast SKUs due this week. Do not count the building because the calendar said so. Freeze that scope. Pause picks and putaway for those bins. Leave every other aisle alone. Count blind. Record the quantity on the shelf before anyone opens the book quantity. Compare. If it matches, release the bin. If it does not match, check open picks, unposted receipts, and unit of measure before any adjustment. Recount blind with a second person when the paperwork does not explain the gap. Adjust only if the two counts agree, and store the reason. Stop the program for the day if step 5 is getting skipped. Adjustments without a cause make next month's count worse. Questions about cycle counting What are cycle counting best practices? Count a portion on a schedule while the rest of the building ships, hide the book quantity until the physical count is written, and do not adjust until an open pick, a receipt, or a unit mix has been ruled out. Cover the catalog over the year, and bring missed bins back sooner. How is a cycle count different from a wall-to-wall count? A cycle count does a slice daily, weekly, or monthly until everything has been counted. A wall-to-wall count does the entire inventory at one time, which usually means stopping the floor. The GAO guide describes both. Should counters see the expected quantity? No. Showing 40 before the count pushes people to write 40. Record the shelf first. Reveal the system number only to investigate the difference. When do you adjust inventory after a count? After a second blind count agrees with the first, and after open picks, receipts, and unit-of-measure errors do not explain the gap. An open pick for the same quantity gets confirmed. It does not get written off. What accuracy number should you track? Correct counts divided by counts taken. Federal Prison Industries uses that ratio and aims for 95 percent or better. Treat 95 percent as their published goal, then track your own misses by reason so the next count list follows the failures. OneChannelAdmin Team writes about cycle counts, bin quantities, and the orders that keep shipping while you count.