My Community Offline dispensary counts when two devices edit the same stock

Blog Information

  • Posted By : Victoria Hodgson
  • Posted On : Nov 20, 2026
  • Views : 6
  • Category : Soccer
  • Description :

Overview

  • Two employees count the same dispensary package while their devices are disconnected. Each saves a different result. After both devices reconnect, the stock screen shows a quantity, but the manager still needs to establish which observation it represents and what happened to the other submission.

    A final balance does not explain the recovery process. The review needs to connect the physical counts, any movements between them, and the records the application accepted when communication resumed.

    When evaluating offline inventory management software, ask for a demonstration of this exact situation. Establish which inventory actions work offline and how competing submissions are handled. Matching screens after reconnection do not prove that the chosen quantity reflects the stock.

    Confirm which inventory actions actually work offline

    Ask the provider to distinguish viewing cached inventory, recording a count, proposing an adjustment, and changing the accepted balance. Support for offline checkout does not establish support for every inventory workflow.

    Identify the prerequisites for the supported task, including account access and any records that must already be available on the device. If inventory editing is unavailable without a connection, document that boundary and establish the store’s alternative counting procedure.

    Separate a count observation from a stock adjustment

    Determine what the employee’s action saves. A full physical count describes an observed quantity. An adjustment changes a recorded balance. A discrepancy submission may require another person’s review before anything changes.

    Keep those stages separate in the investigation. Otherwise, staff may mistake a locally saved observation for an approved correction.

    Establish a common baseline for both devices

    Before testing, identify the same package, room, and unit on each device. Record their displayed balances and the last confirmed server state available through the application.

    Establish the counting scope. Employees looking at separate shelves may be counting different portions of a package, while employees counting the whole package are making competing observations. The records need to distinguish these situations.

    Record when each observation occurred and whether goods moved between them. A sale or room movement can make different results understandable without either employee having miscounted. Do not rely on submission time alone to establish when the physical count happened.

    Ask how competing offline edits are resolved

    Request the application’s documented rule for conflicting submissions. Find out whether it rejects an outdated change, chooses a result automatically, retains separate observations, or sends the difference for review. These possibilities create different recovery tasks for the manager.

    The Firebase offline data documentation provides a concrete technical example: Cloud Firestore synchronizes local changes when a device reconnects and uses “last write wins” for multiple changes to the same document. That describes a database mechanism, not the behavior of a particular dispensary application. Ask the POS provider how its inventory workflow handles competing counts. firebase.google.com

    An automatically selected result still needs a business explanation. Establish whether staff can inspect both observations and see why one affected the accepted balance.

    Do not add together independent full counts

    Full counts of the same goods are alternative observations, not quantities to add together. Averaging them also leaves the actual stock unresolved.

    Combining partial counts requires verified boundaries between the counted areas. Record those boundaries before treating submissions as complementary.

    Test conflicting results in a controlled environment

    Use an agreed test setup with designated inventory records. Preserve the starting state, disconnect the devices through the supported testing process, and record different counts against the same package.

    Keep evidence from each device before reconnecting. Capture the task reference, employee, observed quantity, unit, and displayed submission status. Ask the provider which history or support records can confirm the outcome.

    Also record the application version and relevant settings so a later reviewer can determine which tested behavior applies to the devices in use.

    Reverse the reconnection order

    Run separate tests with different reconnection sequences. Compare the accepted quantity, device displays, and available history with the documented rule.

    Do not assume the device that reconnects last must determine the result. Record what the application actually does. If it produces an unexplained outcome, retain the test evidence for support.

    Include a movement between counts

    Add a separate scenario with a documented stock movement between observations. Check whether the resulting records allow a reviewer to reconstruct the sequence.

    The question is whether an older observation can overwrite a balance that reflects later activity. Ask how the system identifies or reviews that situation rather than assuming synchronization will resolve its meaning.

    Inspect what synchronized and what remained unresolved

    After reconnection, compare each original observation with the accepted inventory record. Identify submissions that remain pending, were rejected, or were replaced, using the statuses and history the application provides.

    Check whether the employee can retrieve a result that did not become the final balance. If that evidence is unavailable in the interface, establish what the provider can supply and include that limitation in the recovery instructions.

    Network access, completed transmission, and accepted inventory changes are separate checkpoints. Verify the outcome at the relevant record instead of treating the disappearance of an offline indicator as sufficient evidence.

    A support record should make the unresolved question specific: which submission was accepted, which quantity remains disputed, and what evidence is missing.

    Reconcile the physical stock before correcting the record

    Assign a reviewer to compare the observations with receiving, sales, and movement records for the relevant period. Where the evidence cannot establish the correct quantity, arrange a targeted recount with a defined cutoff or a record of movements during counting.

    Explain the chosen correction. Preserve the original observations and the reason one was accepted, rejected, or superseded by the recount.

    Do not retry uncertain submissions or clear local data before establishing what happened to the saved work. Another submission may introduce a further change, while removing local records may make the original sequence harder to reconstruct.

    If the workflow sends approved adjustments to another system, verify that submission and its result separately. Resolving the local balance does not establish that every connected record now agrees.

    Assign count ownership for the next outage

    Reduce overlap by assigning packages or clearly bounded areas to named employees. Record handoffs when responsibility changes, including which observations are saved and which work remains unfinished.

    Provide a procedure for accidental overlap. Staff should know whom to contact, how to preserve their results, and which task to pause while the affected package is reviewed.

    Keep a worked conflict example with the outage instructions. Ask a colleague to recover both observations, identify the accepted record, and explain the manager’s decision from the evidence alone. Revise any step that still depends on someone remembering what happened.