What Is a POS Data Export? File Formats and What You Can Actually Take
By Marcus Rivera · Restaurant Systems Analyst · 9 years experience
July 26, 2026 · 10 min read
Definitions matter here because the word "export" gets used to describe four very different things, and operators find that out at the worst possible moment.
A taqueria owner in Sacramento told me he had "exported everything" before his cutover. What he actually had was a folder of 38 PDF sales summaries — human-readable, machine-useless. His new vendor could not import a single line of it. He rebuilt 214 menu items by hand over two nights.
So let's be precise about what an export is, what the formats mean, and what you can realistically expect to walk out the door with.
The Four Things People Call an Export
When a salesperson or a support rep says "we can export that for you," they mean one of these four, and the gap between the first and the last is enormous.
- A self-service download. You click a button in the back office and a file lands in your downloads folder. Fastest, cheapest, and the only one you fully control.
- A support-generated file. You open a ticket, someone runs a query, and you get a file back in one to ten business days. Frequently comes with a fee.
- An API pull. Your new vendor's onboarding team connects directly to your old system's interface and reads the records out. Cleanest result, but requires credentials and an old vendor that still allows the connection.
- A screen scrape or manual retype. Not an export at all. Someone reads a report and types it into a spreadsheet. This is what you fall back to when the first three fail.
Before you ask about formats, ask which of these four your vendor is offering. The answer changes your migration timeline by weeks.
POS Export File Formats, Explained Without the Jargon
Here is what each format actually is, and what it is good for when you are moving systems.
| Format | What It Is | Best Used For | Import Friendliness |
|---|---|---|---|
| CSV | Plain text, one record per line, values separated by commas | Menu items, modifiers, employees, customers, item-level sales | Universal — every POS reads it |
| XLSX | Excel workbook, supports multiple tabs and formatting | Multi-sheet exports where menu, modifiers and pricing live on separate tabs | High, but usually converted to CSV first |
| JSON | Structured text that preserves nesting and relationships | Complex modifier trees, combo logic, API transfers | Developer-friendly, not spreadsheet-friendly |
| XML | Older structured format, verbose but precise | Legacy enterprise systems and accounting handoffs | Moderate — often needs conversion |
| A picture of a report | Accounting archives, audit trails, tax records | None — cannot be imported | |
| SQL dump / .bak | A copy of the underlying database | Forensic recovery when nothing else exists | Requires a developer, but contains everything |
Notice the pattern? Four of these six are importable and two are archives. Yet PDF is the format most restaurants have the most of, because it is what monthly reporting emails you.
Why CSV Wins Almost Every Time
CSV looks primitive next to JSON, and that is exactly why it works. It has no version drift, no proprietary encoding, and no dependency on the software that created it. A CSV your old system wrote in 2019 opens perfectly in 2026.
It also lets you do the one thing that saves migrations: editing the data before it lands. You can open a 400-row menu CSV, sort by category, spot the six duplicate entries, fix the four prices your last manager fat-fingered, and hand your new vendor a clean file. Try that with a JSON blob at 11 PM.
Ask for CSV. Accept XLSX. Take JSON only if your incoming onboarding team requests it by name.
What Actually Travels Between Systems
This is the part that surprises operators. Not all records are equally portable, and the difference has nothing to do with how important they feel to you.
| Record Type | Typical Portability | Notes |
|---|---|---|
| Menu items and prices | Very high (95-100%) | Cleanest transfer of anything you own |
| Modifier groups | High (80-95%) | Nested and conditional modifiers often need manual rebuild |
| Employee records | High (90%+) | PINs and permission levels usually reset by design |
| Customer contact records | Moderate to high | Verify opt-in and consent status carries over |
| Loyalty point balances | Moderate | Often needs a manual balance load, not a file import |
| Item-level sales history | Moderate | Cloud systems: usually 12-24 months. On-prem: varies wildly |
| Gift card liability | Low | Usually a provider-to-provider balance transfer |
| Stored card tokens | Near zero | Tokens are bound to the processor that issued them |
| Open tabs and unsettled tickets | Zero | Close everything before cutover, without exception |
The last two rows are where real money hides. If you have 300 customers with cards on file for a recurring catering program, those tokens do not travel. Your customers will need to re-enter cards. Plan the communication for that four weeks out, not four days.
KwickOS publishes its migration approach in a walkthrough on switching POS systems without losing your sales history, which is worth reading alongside this piece because it covers the receiving side of the same transfer.
Do You Legally Own Your POS Data?
Short version: you own the information, but you may not own convenient access to it.
There is no general federal statute in the United States requiring a POS vendor to hand a restaurant a machine-readable file on demand. Your practical rights come from three places:
- Your contract. Look for a "Data" or "Termination Assistance" clause. Well-written agreements specify a format and a delivery window, such as "CSV within 15 business days of written request."
- State privacy law. If you operate in California, Virginia, Colorado, Connecticut, Utah or one of the other states with a comprehensive consumer privacy act, your guest records may carry portability obligations that flow through to your vendor as a service provider.
- Card network and PCI rules. These generally restrict what can be exported — specifically full card numbers — rather than expand it. Do not expect, or ask for, raw PAN data.
Here is the practical takeaway that matters more than any of the above: export first, give notice second. Nothing about that is adversarial. It simply removes any chance that a routine account status change slows down a file you are entitled to. Operators who run their exports while their account is in good standing almost never have a problem.
If your current agreement has no export clause at all, that is a negotiating point for your next one. Contract terms are one of the dimensions worth scoring during a structured POS demo evaluation, and comparison pages like the KwickPOS side-by-side system comparisons are a reasonable starting point for seeing how different vendors describe their data commitments.
How to Run a Test Export This Week
Do not wait for migration week to discover what your system can produce. Run this in under an hour.
- Log into the back office and find every menu labeled Reports, Data, Export, or Tools. Screenshot the list.
- Download one of each type — menu, employees, customers, and a one-week item sales report. Do not download a year yet; you are testing plumbing, not moving volume.
- Open each file in a spreadsheet. If the columns line up and the rows are readable, you have a real export. If you get a wall of text or a PDF, you do not.
- Count the rows. Compare the menu export row count against the item count shown in your POS. A gap means the export is filtered, often silently excluding inactive or seasonal items.
- Check for the fields that matter: item name, category, price, tax group, modifier group reference, SKU or PLU, and active flag. A menu export without modifier references is only half a menu.
- Time the whole thing. If a one-week report takes four minutes to generate, a 24-month report may time out entirely — good to know now.
Write down the result on one page. That page becomes the first artifact of your migration and it makes every vendor conversation afterward more concrete.
The Fields Most Exports Quietly Omit
Even a clean CSV tends to leave these behind. Document them manually while you still have the old system running:
- Kitchen routing rules — which printer or KDS station each item fires to
- Time-based pricing windows, such as happy hour start and end times
- Forced modifier logic and minimum or maximum selection counts
- Combo and meal-deal construction rules
- Tip pooling formulas and distribution percentages
- Screen layout and button positions your staff have memorized
None of these live in a standard export, and all of them shape how a shift feels. Photographing your POS screens is a legitimate documentation method, and it takes ten minutes.
A Realistic Export Budget and Timeline
For a single-location restaurant with a moderately complex menu, here is what the data phase typically costs in time and money.
| Task | Time | Typical Cost |
|---|---|---|
| Test export and inventory of available files | 1 hour | $0 |
| Full export pull (menu, staff, customers, 24-month sales) | 2-3 hours | $0-$1,500 if vendor-generated |
| Data cleanup and deduplication | 4-8 hours | Internal labor |
| Manual documentation of non-exported settings | 2-3 hours | Internal labor |
| Gift card and loyalty balance reconciliation | 1-2 weeks elapsed | Varies by provider |
The gift card line is elapsed time, not effort — it involves two providers coordinating, which is why it should start early even though it takes almost no work from you.
For the full sequencing of where this phase sits relative to everything else, our POS migration timeline and planning guide maps the six phases end to end, and backing up POS data before switching covers storage and retention in more depth.
Three Export Mistakes That Cost Real Money
Mistake one: exporting once. Your first export is a draft. You will find gaps, re-export, and find more. Budget for three passes over two weeks rather than one heroic session.
Mistake two: trusting the row count. A menu export that produces 187 rows for a 214-item menu is not a rounding error — it is 27 items your staff will discover missing during a dinner rush. Reconcile counts every time.
Mistake three: storing exports in one place. The back office PC that holds your only copy of a 24-month sales export is one power surge away from an unpleasant conversation with your accountant. Two locations, minimum, one of them off-site.
Get those three right and the technical side of a switch stops being the scary part. The remaining risk sits in training and scheduling, which is a much more manageable kind of problem — and one covered directly in our guide to switching POS without losing sales data.
Frequently Asked Questions
What file format should I ask my POS vendor for?
Ask for CSV first. It is the format every POS on the market can import, it opens in any spreadsheet tool, and it survives being edited by hand. Request XLSX only if your data has a multi-tab structure, and request JSON or an API pull only if your new vendor's onboarding team specifically asks for it.
Does my POS vendor legally have to give me my data?
In the United States there is no general federal law forcing a POS vendor to hand you a machine-readable export. Your rights come from your contract, so check the data and termination sections for an export clause. If you operate in California, Virginia, Colorado or another state with a consumer privacy statute, your customer records may carry additional portability rights on top of the contract.
Can I export historical transaction data from my old POS?
Usually yes, but often as summary reports rather than line-level records. Most cloud systems export 12 to 24 months of daily and item-level sales as CSV. Older on-premise systems frequently cap you at PDF reports, which cannot be imported anywhere — one more reason to pull everything while your account is active and in good standing.
How long should I keep the exported files?
Keep menu and employee exports for at least a year and sales exports for seven years, which matches the IRS record retention guidance most restaurant accountants follow. Store them in two places, such as a cloud drive plus an external disk, and never as the only copy on the back office PC.
What data almost never transfers between POS systems?
Stored payment card tokens, saved card-on-file records, and gift card liability held by the old processor almost never move. Tokens are cryptographically bound to the processor that created them, and gift cards usually require a balance transfer coordinated between the two providers rather than a file import. Open tabs never transfer at all, so close every ticket before cutover.
Start Your Free Trial — No Credit Card Needed
KwickOS imports your CSV exports free, runs on your existing hardware, and gives you a named onboarding specialist for the data phase.
Start Free Trial →Related reading: Data Migration Between POS Systems · What Is POS Migration? · POS Data Backup Before Switching · SwitchYourPOS Home