Financial Reconciliation & Payment Analysis
Reconciling EPOS tender reporting against card settlement and standalone payment sources, then investigating the variances that remain.
- Payment sources
- Multiple
- Figures in examples
- Illustrative
Overview
Hospitality and event operations typically take payment through more than one rail: integrated card on the POS, standalone terminals, and digital or third-party tenders. This work is about making those sources comparable and explaining the difference.
The Problem
A single EPOS total is not a settlement figure. Integrated card, standalone card and digital tenders land in different reports, on different timings, with different failure modes. Without a structured reconciliation, variances get absorbed into 'timing' or 'the bank' and are never actually understood.
Requirements
- Compare EPOS tender reporting to settlement reports by source
- Separate integrated payments from standalone terminals
- Include digital / alternative tenders where they exist
- Surface variance at a level that can be investigated
- Keep the method repeatable for subsequent trading periods
My Role
Owned the analysis: source mapping, comparison, variance investigation and the reporting that finance and operations could use. No live transaction data is published here.
Approach
- Map each tender type to a settlement source rather than reconciling a blended total
- Align reporting windows so timing differences are visible instead of hidden
- Treat a variance as a queue of investigations, not a single unexplained number
- Use Excel and structured extracts where that is the fastest reliable method; automate the repeatable parts
System / Architecture
KAPPTURE tender reports, integrated settlement (FreedomPay, Global Payments, Worldpay depending on the rail), standalone terminal reports and digital sources (including Stripe where used) feed a reconciliation layer. The output is source-level comparison plus an exception list.
Payment sources → Reconciliation
Illustrative data| Source | EPOS | Settlement | Variance |
|---|---|---|---|
| Integrated Card | £218k | £218k | £0 |
| Standalone | £43k | £42.9k | £100 |
| Digital | £28k | £28k | £0 |
Implementation
- Source catalogue: what each report actually contains and what it does not
- Comparison tables by tender / rail
- Variance flags with enough context to investigate a transaction class
- Working papers in Excel for analysis; automated reporting where the extract is stable
Challenges
- Different cut-off times between POS and acquirer / processor reports
- Standalone devices that are operationally 'on the outlet' but not in the EPOS tender file
- Partial refunds, voids and offline behaviour
- Keeping commercially sensitive figures out of any shared artefact — including this case study
Outcome
Reconciliation became a source-level control rather than a single unexplained difference. The example table on this page is illustrative only; it shows the shape of the control, not real turnover.
- Clearer view of integrated vs standalone vs digital tenders
- Variances isolated to a source rather than a blended total
- A method for investigating exceptions instead of re-running the same export
Technologies
What I Learned
- You cannot investigate what you have blended together
- Timing is a first-class field, not a footnote
- Operators and finance need different views of the same exception
Future Improvements
- More of the extract-and-compare steps moved into automated reporting
- Tighter linkage from variance to the underlying transaction class
- Clearer handling of standalone estates alongside integrated payments