How reconciliation works
Three systems usually claim to count the same thing. An ad platform reports conversions it can attribute to its own clicks. An analytics property reports events fired on the website. The business reports orders, invoices or qualified enquiries from its own database or CRM. Reconciliation is the disciplined comparison of those three, done to explain the differences rather than to remove them.
The work is mechanical. You fix a single date range and time zone, define exactly which action counts on each side, then walk the gap down cause by cause: conversions credited to the click date instead of the purchase date, estimated data filling gaps in observation, duplicate firing, blocked tags, refunds and cancelled orders, enquiries that were never genuine.
A useful sequence when the numbers refuse to line up:
- Match the date range, time zone and currency on both sides before anything else.
- Confirm the two systems define the action identically, including whether repeats count.
- Check for duplicate or missing firing on the website itself.
- Allow for attribution timing, then for estimated and unobserved data.
- Remove refunds, cancellations and rejected enquiries from the business figure last.
Why reconciliation matters
Unreconciled numbers destroy trust faster than bad results do. Once a client or a finance team sees two figures for one month, every later report is met with suspicion, and good decisions get delayed while people argue about which screen to believe.
It also protects the money. A gap can hide a genuine fault — a tag that stopped firing after a website release, a payment step that fails on certain phones, an enquiry form silently rejecting long messages. Reconciliation is often how those are found, because the mismatch appears before anyone notices the drop.
Common mistakes with reconciliation
Expecting a perfect match is the big one. The systems measure different things by different methods, and a stable, explained gap is the realistic goal. Chasing zero difference wastes days and usually ends with someone quietly changing a setting to force agreement.
The other mistake is reconciling only when something looks wrong. A gap you have never measured has no normal, so you cannot tell a routine difference from a broken one. Comparing time zones carelessly causes more false alarms than any real fault.
How to act on it
Reconcile on a fixed rhythm, monthly for most small businesses, and record the size of the gap each time rather than only the verdict. A gap that holds steady is a known quantity; a gap that moves is the signal worth investigating.
Decide in advance which system is the source of truth for each question, so a difference triggers an explanation instead of a debate. Keep the definitions, filters and time zone written down beside the report, and build the comparison into your reporting and dashboards rather than doing it from memory each time.
