A weekly routine, not a monthly scramble
Left unattended, POS batch and deposit reconciliation drifts into a monthly scramble — a controller pulling three weeks of statements at once, trying to match batches to deposits from memory of which nights were busy. That approach works until it doesn't, usually right when a genuine discrepancy needs to be found among a pile of ordinary timing differences.
This guide describes eight steps to do instead, on a weekly cadence, so each period's reconciliation stays small enough to actually finish.
Why this drifts without a routine
A single week's reconciliation is a manageable task. Three weeks stacked together is three times the volume with none of the context — by the time someone gets to it, the person who worked a specific busy Friday may not remember what happened, and the original documents have to speak for themselves.
A fixed weekly cadence keeps the volume small and the context fresh, which is most of what makes this reconciliation fast rather than dreaded.
Two decisions before the first batch
First: who owns this weekly, by name — not “whoever has time,” which quietly becomes nobody the first busy week. Second: what counts as a discrepancy worth escalating versus a timing difference that resolves itself within the processor's normal settlement window.
Both decisions take five minutes to make and prevent the two most common ways this routine falls apart — nobody doing it, and every minor timing gap being treated as a crisis.
Picking a day and sticking to it
The specific day matters less than the consistency of picking one. A Tuesday reconciliation, done every Tuesday, means the prior week's batches have had time to settle and the following week hasn't started yet — a natural gap that doesn't compete with either the weekend rush or the next week's planning.
What matters more than the exact day is protecting it. A reconciliation that gets bumped “just this once” because a busier task came up tends to get bumped again the following week, and the routine quietly slides from weekly to monthly without anyone deciding that on purpose.
The eight steps
Pull the POS batch report for the period
Every batch closed during the week, from every terminal or location involved — exported directly from the POS system rather than retyped from a printed summary, since a printed summary is one more place a transcription error can creep in.
Include void and refund detail if the export offers it separately; those lines matter later when reconciling against the deposit.
Pull the processor deposit statement
The corresponding statement from your card processor for the same window, including the fee and chargeback detail most processors bury several pages into the statement rather than surfacing up front.
If deposits arrive on a rolling basis rather than all at once, pull the statement a few days after the period closes so the last batch's deposit has had time to actually land.
Match each batch to its deposit
Use the processor's reference number where the statement includes one — it's the strongest available signal. Where it's absent, match by amount net of a plausible fee and by date proximity within the processor's usual settlement window.
Keep a running note of which batches remain unmatched after this step — that list is the input to step six.
Isolate fees, refunds and chargebacks
For each matched pair, confirm the gap between the batch total and the deposit is fully explained by a processing fee, a refund from a prior batch, or a chargeback — not just approximately close, but accounted for line by line.
A gap that's close but not exactly explained is worth a second look before being accepted — an unexplained few dollars in a dozen batches adds up to a real discrepancy hiding inside what looked like rounding.
Reconcile tip totals against the tip pool
Check the credit-card tip total read from each batch against what was actually distributed through the tip pool for that period. Cash tips, tracked on their own log, get combined at this step — not earlier.
A mismatch here usually traces to a tip pool built from a printed total before the batch fully settled, rather than from the confirmed, reconciled figure.
Investigate any unmatched batch or deposit
For each item on the unmatched list from step three: first confirm the batch actually closed and submitted successfully. Second, check for a split settlement, where one batch produces two smaller deposits. Only escalate to the processor after ruling out both.
Most unmatched items resolve at this step without ever needing outside help — they were simply still in transit or split across two deposits that hadn't both landed yet.
Document the exception
A short note — what caused the discrepancy, how it was confirmed — attached to that week's reconciliation sheet. This is the step most often skipped, and the one that costs the most time to skip when the same question resurfaces months later.
See documenting an exception below for what a good note actually looks like in practice.
Close the period
Export the reconciled batches and deposits, with every line traceable back to the source document, and archive it. This is also the moment to note anything structural worth carrying into next week — a supplier fee change, a new terminal added.
Handling discrepancies with method
Not every unmatched item is a real discrepancy — most resolve within the processor's usual settlement window once given a few more days. What separates a genuine discrepancy from a timing gap is whether it's still unexplained after that window has fully passed.
Sharing a specific, concrete discrepancy with your processor — not a vague “something looks off” — tends to get a faster, more useful response than a general inquiry, and often surfaces a process fix on both sides.
Keep a short running log of confirmed discrepancies and their resolutions, even the small ones. Over a few months, that log becomes its own useful reference — a pattern that repeats twice is worth fixing at the source; a pattern that happened once, three months ago, is worth remembering the next time something similar shows up.
A week, worked through
A single location, seven batches, one weekend combined into a single Monday deposit due to a bank holiday. Six of seven batches match cleanly by reference number in under ten minutes. The seventh — Saturday's batch — shows up unmatched at first glance.
Step six resolves it: Saturday's batch settled together with Sunday's in the combined Monday deposit, which the reference number confirms once both batches are checked against it together rather than separately.
Common mistakes
Treating the POS batch total as the deposit amount without accounting for fees.
Distributing a tip pool before the batch has actually settled and reconciled.
Letting three or four weeks pile up before reconciling any of them.
Escalating a timing difference to the processor before the settlement window has passed.
Skipping the documentation step on a resolved exception.
Best practices
Reconcile on the same day every week, so it never competes with a busier day's priorities.
Keep the unmatched list visible until every item resolves, rather than letting it fall off the page.
Note structural changes — a new terminal, a processor fee change — the week they happen.
Review three months of reconciliation sheets together once a quarter to spot slow drift.
None of these four are complicated on their own — together, they're what separates a reconciliation routine that survives a staff turnover or a busy season from one that quietly stops happening the first time the usual person is unavailable.
Turning it into a routine, not a chore
The eight steps stay the same every week; what changes is how much time each takes once the historical backlog is cleared and a document reader handles steps one through four automatically. A routine that used to take an hour settles into fifteen minutes once it has nothing but the current week to process.
The monthly check
Beyond the weekly routine, a monthly review comparing total fees paid against total volume processed catches a fee-rate creep that a single week's reconciliation, focused on matching rather than rate analysis, would never surface on its own.
A processing rate that drifts up half a percentage point over six months rarely shows up as a single dramatic discrepancy — it shows up as a slow decline in the deposit-to-batch ratio that only a monthly comparison makes visible.
What you actually need
None of this requires specialized software to start. A spreadsheet works fine for the reconciliation sheet. What's genuinely slow by hand is reading each batch report and statement consistently enough, week after week, to keep that sheet current.
A document reader that extracts batch totals, deposit amounts and fee lines the same way every time, whether the source is a clean export or a scanned statement, removes exactly that bottleneck — see POS batch matching for how that reading feeds directly into the matching step.
Running this across multiple locations
For a restaurant group, the same eight steps apply per location, run in parallel once the routine is established — not sequentially, which would turn a fifteen-minute weekly task into an afternoon.
A shared reconciliation sheet format across every location makes it possible to spot a pattern that repeats at one location but not the others — often the first sign of a processor or fee-structure difference worth investigating.
See finance for restaurant groupsfor how this routine changes shape once it's running across a full portfolio of locations rather than a single one.
The very first time, step by step
The first reconciliation under this method usually covers more than one week of backlog, which means more unmatched items than a normal week and more time spent on step six. Budget an hour rather than the fifteen minutes a steady-state week will eventually take.
Once the backlog clears, the routine settles into its normal, much shorter rhythm — the first week is the investment that makes every week after it fast.
It's worth resisting the urge to rush the first pass just to clear the backlog quickly. A discrepancy investigated properly the first time teaches something about how that specific processor or POS system behaves — knowledge that pays off on every reconciliation that follows.
The reconciliation sheet, column by column
| Column | Purpose |
|---|---|
| Batch date & reference | Ties the row back to the original POS batch report |
| Batch gross total | What the batch reported before any deductions |
| Matched deposit amount & date | The net figure that actually landed, and when |
| Fees / refunds / chargebacks | What explains the gap between the two totals |
| Card tip total | The figure the tip pool should be built from |
| Status & notes | Matched, flagged, or resolved with a short explanation |
The rhythm that repeats every year
Holiday weeks bring combined deposits and higher volume. A slow season brings smaller batches that make a genuine discrepancy easier to spot precisely because there's less noise around it. Neither pattern is a surprise once it's been seen once — the same eight steps apply, just with different expectations for what a normal week looks like.
Keeping a short note of which weeks each year tend to bring combined deposits — the specific bank holidays that affect your processor's schedule — turns a recurring “why does this look different” moment into an expected, already-understood pattern the second year running.
Building a clear escalation path
Define in advance who gets contacted for what: a processor support line for a genuinely missing deposit, an internal manager for a POS terminal that isn't closing batches correctly, and nobody at all for a timing gap still inside the normal settlement window.
Without that path defined ahead of time, every discrepancy defaults to whoever happens to notice it first, which produces inconsistent handling of what should be a routine, well-understood process.
What good documentation looks like
A useful exception note is short and specific: what the discrepancy was, what caused it, and how that cause was confirmed — not a vague “resolved” that tells a future reader nothing about what actually happened.
“Saturday batch settled combined with Sunday in Monday's deposit, confirmed via processor reference number” takes ten seconds to write and saves ten minutes of re-investigation the next time a similar pattern shows up.
Measuring whether the routine works
Track two simple numbers: what percentage of batches get a matched deposit within the expected window, and how many days it takes on average to resolve a flagged discrepancy. A routine that's working keeps both numbers stable, week after week, without a dedicated compliance hire watching over it.
A rising number of unresolved discrepancies is usually the first sign that either volume has outgrown the current process or a structural change — a new processor, a new POS terminal — needs to be accounted for explicitly.
Who inside the restaurant should own this
In most independent restaurants, nobody has this as an explicit job title. A general manager closes the register and moves on to the next task. A bookkeeper sees the deposit weeks later, disconnected from the specific shift it came from. The reconciliation itself sits in the gap between the two roles — which is exactly why it's the first thing to slip when either person gets busy.
The routine doesn't need a dedicated hire to work. It needs to sit explicitly with one named person — usually whoever already closes out the register — with these eight steps as a concrete checklist rather than a vague expectation that “someone checks the deposits.”
What tends to make the difference isn't adding headcount, it's making the ownership specific enough that it survives a schedule change. A manager who knows they personally own Tuesday's reconciliation behaves differently than a team where everyone assumes someone else is watching the deposits.
It's worth naming a backup, too — someone who can step in for a single week without having to relearn the routine from scratch. A reconciliation that only one person can perform is fragile in exactly the way a documented, shared routine isn't, and that fragility only becomes visible at the worst possible moment: when the usual owner is suddenly unavailable.
A spreadsheet alone versus a document reader
A spreadsheet is enough to hold the reconciliation sheet itself — dates, amounts, matched status, notes. What a spreadsheet can't do on its own is read a batch report or a processor statement and pull the numbers out; that part still has to happen by hand, typing figures off a printed page into the right cell.
For a single location doing a handful of batches a week, that typing is tedious but bearable. For a location running two shifts a day, seven days a week, the typing alone can eat the fifteen minutes this routine is supposed to take before the actual matching and review even begins.
A document reader that extracts the same fields consistently, whether the source is a clean digital export or a scanned paper statement, removes exactly that typing step — leaving the spreadsheet for what it's actually good at: holding the matched, reconciled record.
The two aren't competing tools — a spreadsheet still holds the final reconciled view, whichever way the underlying numbers got into it. What changes is only how much manual effort sits between a stack of documents and that final view, and that manual effort is exactly where a reconciliation routine tends to stall out.
Training a new manager on the routine
A new general manager inheriting this routine doesn't need a long training document — they need to walk through one real week's reconciliation with the previous manager, seeing an actual unmatched batch get resolved rather than reading a hypothetical example.
The discrepancy log from a prior month, with resolution notes still attached, teaches the judgment calls — what's a genuine problem versus an ordinary timing gap — faster than any written policy could on its own.
It's worth being explicit about the one habit that's easiest to skip when someone is new and eager to clear a backlog: guessing at a resolution rather than confirming it against the actual processor statement. That single habit, called out directly during training, prevents most of the mistakes a new manager would otherwise make in the first month.
Staying ready for a processor or franchisor review
A processor can flag an account for review at any time, and a franchisor can request reconciled sales data with a short notice window. Neither event is predictable, which means the readiness for it has to be built into the weekly routine rather than assembled after the request arrives.
A reconciliation sheet that's current through last week, with every match traceable to its source batch and deposit, turns a surprise request into an export. A reconciliation sheet that's three months behind turns the same request into a multi-day reconstruction project, done under a deadline instead of at a comfortable pace.
The difference between those two outcomes isn't effort spent preparing for an audit specifically — it's simply whether the weekly routine has actually been followed consistently. A restaurant that reconciles every week is, without any extra work, already audit-ready every single day, which is the whole point of building this into a routine rather than a periodic catch-up project.
That readiness costs nothing extra once the routine is running — it's a byproduct of doing the reconciliation weekly rather than a separate task added on top of it, which is ultimately the whole argument for building a routine in the first place rather than treating reconciliation as an occasional, standalone project that competes for attention with everything else on a busy week and inevitably loses to whatever feels more urgent on that particular, unpredictable, busy day — which, in a restaurant, is most days of the year.