0 Replies Latest reply on Apr 22, 2019 2:50 PM by Gman26

    Cycle Count Advice (metrics, management)


      Hi Everyone,


          I'm working on a rather large cycle count project covering 20+ warehouses of stuff (tens of thousands of sequence numbers). Inventory accuracy is pretty low and unrecorded movements are a chronic problem. Those are separate issues I have no control over other than reporting to our client how much of a problem we think it's causing them. We're also having some other issues with Oracle though that I'm trying to get help with. Does anyone know of some good solutions, or stuff we're obviously doing wrong? We've created the cycle counts like in this video guide (https://www.youtube.com/watch?v=Ld_s0EN35Jo ) but I'm running into a lot of issues.


      Constraints: No barcodes or scanners and limited laptops / internet connectivity.

      Background: Working on Oracle R12 EBS, regular user access, no back-end access to Oracle data tables.


      1.) Everything Cycle count related is plaintext (xml or text file copy from web browser). Is there any way to get the standard reports (cycle count entries & adjustments, cycle counts open listing, etc.) in a csv or excel format?

            I thought creating one cycle count for the entire organization would be a good idea, but got overruled. So now there's cycle counts for individual racks in individual warehouse sections since we're using paper and running a cycle count on a whole warehouse would be thousands of pages. This has led to the problem of having dozens of cycle counts created but items slipping through the cracks when they have non-standard locators (eg. if the locator for a given rack in a warehouse should be Row 10, Column 20, A level and someone puts stuff in R10C20A instead the cycle count creator might not see these items when creating the cycle count and they get skipped).


      Is there any way to write a report to export all cycle counted items from a given date range and have it export to a .csv or excel file? I'm trying to do this for progress reporting purposes since the Oracle reports I see only produce output for a given cycle count and can't be summarized. I've created a python script to convert and aggregate the reports, but it still takes hours to run and export the reports since there's so many cycle counts open. Another reason I'd like this in a spreadsheet format is so that I can automatically compare a separate QA team's counts to the cycle count team's counts to produce accuracy metrics. Since I don't see any QA modules in our cycle counting menus those counts are recorded outside of Oracle and it's impossible to compare to the text file output Oracle produces.


      Some of the metrics I'm trying to produce are things like: Overall % of cycle counting complete, % complete by item type, dollar amount gained, dollar amount lost, A-item category gains/losses, % of inventory not on location or added to location during the cycle count, % of inventory adjusted for admin reason (mislabeling, wrong SKU in Oracle vs product, etc.). Some of these I can manually create using SQL and the exported transaction history, but others are difficult because the only reports I can get are textfiles.



      2.) Zero counts freezing items

           Since picks & stows are happening frequently and without proper transactions in Oracle, often times counters will arrive at a location, report 0 qty for an item, and move on. In some other warehouse another team might report finding that same serial numbered item and try to add it to their location. But if the first team has already 0 counted the item, the second team can't count it and say it's in their location since the 0 count seems to lock the item until it's processed for a drop from the system. I've messed around with selecting the "serial number is already in the system" check-box and that seems to work well for add-on serial numbers, but I'm not sure if it'll work if that SN was already 0 counted in another cycle count. What we're having to do now is find the 0 count, cancel it, and then pull the item over to the correct location and count it there, or move all 0 counts into a temporary holding location and no 0 count them so that they can be pulled over by anyone that comes across them later.



      My ideas for the future: I think creating one cycle count for the entire inventory would work better. Then even if we just use the Oracle reports they'll provide a decent amount of metrics for progress and whatnot and we can show our client how to use their system. However, this will create the problem of having to export thousands of pages of counts at one time when they want to manually go out and count using pencil & paper. So I was also thinking that despite not having barcode scanners, we could have people logged into laptops with wifi connections go around to each location, type in the locator they're at, and then enter in the counts that pull up. This seems to work well in theory and would help prevent the data entry lag-time we're experiencing.