The numbers do not match between your systems
The sales report says one thing, the accounts system says another and the spreadsheet used for the board pack has a third answer. The team can make them agree, but only after exports, lookups and manual adjustments.
That is usually a specific disagreement about definition, timing, ownership or movement between systems. APIwise traces each important figure to its source and scopes the smallest controlled fix.
Start here
Take one disputed record, align the definitions and dates, then trace it through each source, transformation and report before rebuilding anything.
- Finance and operations repeat reconciliation every week or month.
- Meetings start with an argument about whose number is correct.
- A spreadsheet becomes the unofficial source of truth without ownership.
- Automation acts on stale, duplicated or differently defined data.
- Exceptions are fixed in the report but return because the source is unchanged.
- 01Two systems define the same metric differently.
- 02Reports use different event dates, statuses or refresh times.
- 03Refunds, fees, VAT, currency or cancellations are treated differently.
- 04A sync is delayed, one-way or failing on particular records.
- 05Identifiers do not match, creating duplicates or missed joins.
- 06Manual adjustments never return to the source.
- 07A dashboard uses a stale extract or transformation.
Start with one real example.
Preserve the current data before bulk replay, overwrite or clean-up. One traceable example is more useful than a broad description.
- Choose one disputed number and write down its exact business definition.
- Record date basis, status filters, VAT, currency, refunds and exclusions.
- Identify the system that creates the underlying event.
- Check refresh or sync times for every report.
- Find one record included on one side but not the other.
- Trace manual adjustments and the identifier used to join systems.
The Integration Health Check maps definitions and data movement from source systems through transformations to the report. It separates source-data, mapping, timing and reporting-rule issues.
The fix may be field cleanup, revised ownership and sync rules, validation checks, a controlled reporting dataset or exception monitoring. A spreadsheet can remain when it is the sensible output.
Define
Agree what the disputed figure means and when it is final.
Sample
Collect records that agree and records that do not.
Trace
Follow records through source, integration and report.
Own
Assign each fact and each exception decision.
Repair
Correct the responsible mapping, rule, timing or flow.
Monitor
Alert when totals, counts or freshness cross a threshold.
Mechanism-only copy. APIwise offers data-flow mapping, matching and reporting-ready integration, but no approved client case study showing measured reconciliation improvement is currently available. Some differences are legitimate because systems answer different questions.
A fit when
A team repeatedly reconciles systems and reports where definitions or data movement are part of the problem.
Try the smaller route first when
An accountant or reporting analyst may be the better first route when the flow is sound and the issue is solely accounting treatment or metric definition.
Why does Xero not match our sales report?
The reports may use different dates, statuses, VAT treatment, fees, refunds, currency or source records. Compare one transaction through both definitions first.
Can you create a single source of truth?
We can define which system owns each fact and design a trusted flow. That does not always require one new database.
Do we need a data warehouse?
Not necessarily. We first fix the flow between existing systems. A warehouse is appropriate only when the reporting and scale require it.
Can Excel remain part of the process?
Yes. The aim is reliable inputs, definitions and exceptions, not banning the right analysis tool.