Reconcile daily orders with Make

Outcome and scope

Compare source orders with the internal order log each day. Matching counts alone cannot prove that the same orders are present, so compare sets of order identifiers and inspect duplicates separately. This scenario produces a reconciliation worklist; it does not ship an order, change payment status or issue refunds based on an unexplained mismatch.

Reconcile order states as well as dates. A source containing drafts and cancelled orders cannot be compared directly with a confirmed-orders-only log without an exclusion rule.

Prepare the fields and destination

Prepare these fields: Reconciliation date, source order ID, logged order ID, retrieval range, duplicate count, discrepancy reason and reviewer. Keep source values separate from a reviewer’s confirmed values, and do not reuse an existing column for a different meaning. Create the first record with practice identifiers in a destination whose read and edit access matches the intended staff roles. Inspect the saved result from that role before using live information.

Normalize cautiously. Removing punctuation or changing case can merge identifiers that the source treats as distinct. Keep raw and comparison values together and account for retrieval limits.

Build the workflow in order

Set the same time interval on the source and destination. If one system uses another timezone, define the conversion explicitly and allow a documented lookback for late arrivals. For a paginated source, record that retrieval reached its end; a first-page count is not a valid basis for comparing the whole day.

Normalize identifiers consistently and build two difference lists: source-only and log-only. Preserve leading zeroes by keeping identifiers as text. A set comparison hides repeated copies of an ID, so also count occurrences in each system. Retain the raw identifier when a normalization rule changes punctuation or whitespace.

Save retrieval times and source references alongside discrepancies, then let an owner classify the cause. Separate late arrival, cancellation, manual entry and integration failure. Notify a suitable channel with counts and a controlled link, rather than copying customer details into a broadly accessible chat message.

Work through a small example

A useful practice case has A, B and C in the source, but A, B and B in the log. Both sides contain three rows, yet C is missing and B is duplicated. Ensure the reconciliation exposes both findings. If C arrives tomorrow, close the earlier discrepancy with a reason instead of erasing its history.

Handle changes during operation

Track new, ongoing and resolved discrepancies rather than creating a fresh incident each day. A successful zero-difference run also needs a record so silence is distinguishable from a monitor that never executed.

Estimate cost and usage

Count retrieval on both sides of the reconciliation, including pagination and the historical lookback. Record orders per day, pages, days reviewed and discrepancy notifications. Choose the range around the business’s required detection time. Leave room for a recovery backfill instead of budgeting the entire allowance around an unusually clean ordinary day.

Investigate failures

A large discrepancy may indicate the wrong shop, time interval or destination, not a mass loss of orders. Verify those parameters before any repair. Re-import only identifiers confirmed as missing, then run the same reconciliation again. Keep the repair reference so staff can distinguish a recovered order from a newly received one.

Review before release

Before release, test a normal record, one missing a required value, a repeated identifier and an edited record. Use reconciliation date to trace the saved result back to the source. The acceptance condition is: Compare counts and identifier sets, and keep a reasoned history of repairs. Record the expected result before testing, compare it with the actual saved state, and resolve unexplained differences before increasing the scope.

• Log Shopify orders
• Use HTTP requests in Make
• Handle Make errors