Written by: JJ Tan, Founder, Jelly
Key Takeaways
- Scanning invoices into Excel works for low volumes, up to around 20 invoices a month, before it turns into a manual burden.
- UK VAT and HMRC compliance require structured digital records, so data in Excel must flow into MTD-compatible accounting software.
- Microsoft 365 Data from Picture offers a free OCR route for occasional use, but it lacks line-item extraction, VAT mapping, and accounting integration.
- At 50+ invoices a month, manual processing consumes hours of admin time each week and error rates rise, so dedicated invoice platforms become the practical next step.
- Jelly automates the workflow from photo or email capture through to dish costing and real-time margin reporting, removing the spreadsheet entirely.
Before You Begin: Tools, Roles and Time Cost
Set up the basics before scanning any invoices so the process runs smoothly and the data is usable.
- A smartphone camera or flatbed scanner. For standard office documents such as invoices, a resolution of 300 DPI is recommended, with 400–600 DPI for fine print and 600 DPI or higher for engineering drawings, photographs, or archival records.
- A Microsoft 365 subscription for the native Excel method.
- The invoices themselves, either paper originals or PDF files.
- A clear list of the columns you need. For supplier invoices, a line-item model with these columns works well: Invoice Number, Invoice Date, Supplier Name, Item Description, Quantity, Unit Price, Line Total, Payment Status, and Notes.
In most hospitality businesses, the owner or finance manager owns the process, while the head chef uses the output for dish costing. At multi-site level, finance usually manages data capture and the kitchen teams consume the results.
Set expectations before starting. Scanning five invoices on a flatbed scanner takes roughly 5 to 15 minutes, at 1 to 3 minutes per invoice. That per-invoice time is what makes the approach break down at scale. Manually processing 50 to 500 invoices a month costs roughly 2.5 to 25 hours of labour per month and $50 to $500 in direct labour at benchmark rates. This guide focuses on that threshold.
Why Invoice Scanning Matters for Food Margin
Invoice data underpins dish costing, supplier price tracking, and gross profit visibility. Without it in a usable format, margins stay invisible until the accountant’s monthly report arrives, by which point a supplier price increase from three weeks ago has already eroded GP on every affected dish.
Getting invoice data into a structured format enables four things: faster reactions to supplier price changes, fewer manual errors in costing, stronger evidence for supplier negotiations, and consistent visibility across sites. The spreadsheet is the first step toward those outcomes, not the destination.
How to Scan Invoices to Excel: Step-by-Step
OCR converts an image of text into editable data, and every method here relies on it. These five steps apply regardless of which tool you choose.
- Prepare the invoice for accurate OCR. Lay the invoice flat on a clean surface. Use even lighting and avoid shadows across the page. Scan or photograph at 300 DPI or above. At 150 DPI, character-level recognition begins to fail on small text, dense tables, and closely spaced numbers. Crop the image to the invoice edges before processing.
- Choose your method. The Microsoft 365 Data from Picture feature in Excel handles occasional use at no extra cost beyond a Microsoft 365 subscription on desktop, and is available on the Excel mobile apps for iOS and Android even without a subscription. Free web OCR tools such as Smallpdf and iLovePDF impose usage limits, with daily task caps and file-size restrictions, and they do not extract line items or table structure. Paid invoice scanning platforms extract line items, handle VAT fields, and integrate with accounting software, which this guide covers later.
- Extract the data. Insert the image or PDF into your chosen tool and let it read the document. In Excel, go to Data > From Picture > Picture From File or Picture From Clipboard, then review the detected cells in the preview panel before inserting.
- Review and correct. OCR is never perfect, and complex fields such as individual line items and multi-row tax breakdowns drop to 95–97% accuracy even with the best systems. Check quantities, unit prices, and VAT lines against the original invoice before saving.
- Structure the spreadsheet. Set up consistent columns such as supplier name, invoice date, invoice number, line item, quantity, unit price, VAT rate, VAT amount, and total. Save as .xlsx and keep the same column structure every month so you can compare data across periods.
Using Microsoft 365 Data From Picture After Microsoft Lens Retirement
The native route for scanning invoices into Excel uses the Data from Picture feature, available in Microsoft 365 on Windows, Mac, Web, iOS, and Android. On Windows, it requires Windows 10 version 1903 or later and the Microsoft Edge WebView2 runtime. It does not work on standalone Excel licences without a Microsoft 365 subscription.
On Windows, open Excel, go to the Data tab, select From Picture, then choose Picture From File to upload a saved image or Picture From Clipboard to paste a screenshot. Excel analyses the image, displays a preview of the detected cells, and lets you correct errors before inserting the data. On Mac, you can also scan directly with an iPhone camera via Continuity Camera by right-clicking a cell and selecting Scan Documents.
One key update for 2026 affects phone-based scanning. Microsoft Lens was retired from iOS and Android starting 9 January 2026, with full removal from the Apple App Store and Google Play Store on 9 February 2026, and the ability to create new scans ending on 9 March 2026. Microsoft now recommends the OneDrive app for scanning: open OneDrive, tap the + button, choose Scan photo, and save to OneDrive. The OneDrive app does not support saving scans locally to the device.
The native method suits a handful of invoices, requires no extra software, and is effectively free with a Microsoft 365 subscription. It breaks down on multi-page supplier invoices because it has no line-item extraction, no VAT field mapping, and no way to push data into accounting software. It remains a manual step rather than an automated pipeline.
Free Web Tools vs Paid OCR for Hospitality Invoices
Free web OCR tools convert PDF invoices or images to text or basic spreadsheet output and work for occasional, low-volume use. Most impose page caps on free tiers, accuracy drops on photographed receipts or invoices with handwritten notes, and none extract structured line-item data with VAT fields mapped correctly.
OCR vendors often score 95%+ accuracy on clean, digitally generated test documents, but accuracy drops to 85–90% on faxed documents, photographed receipts, or forms with handwritten notes. For a hospitality operator photographing paper invoices from a mix of suppliers, real-world accuracy usually sits near the lower end of that range.
Paid invoice scanning platforms, including Jelly, go further. They classify the document, extract every line item, map VAT fields, validate that the line-item total matches the invoice total, and push structured data into accounting software. For document volumes under 500 per month, any OCR tool works; at 500 to 5,000 documents per month, platform overhead becomes justified and human-in-the-loop costs remain manageable. For a busy restaurant or pub, the crossover point arrives much sooner than 500 invoices.
The Volume Threshold: When Excel Stops Being Practical
Excel remains a tolerable choice at around 20 invoices a month, and factors such as user count, inventory complexity, and regulatory deadlines can force a switch earlier. It becomes the wrong answer beyond that point because the manual effort compounds, even though the tool itself does not fail immediately. Each invoice requires scanning, OCR correction, column mapping, and a check against the original. At 50 invoices a month, manual processing takes about 2.5 to 2.9 hours, which is a recurring half-day of admin. At 200 invoices a month, manual processing at 5–7 minutes per invoice equates to roughly 17–23 hours of admin work per month, close to a part-time role.
The threshold is usually crossed when the spreadsheet starts costing more time than it saves. Watch for these signs, which often appear together:
- Spending 10 to 20 hours a week on manual data entry, price checking, inventory, and invoice reconciliation.
- Rising error rates in costing because prices in the spreadsheet lag behind actual supplier invoices.
- Supplier price changes only surfacing when the accountant’s report arrives.
- The spreadsheet itself needing constant maintenance instead of simply providing insight.
- Chefs and finance seeing different numbers because the sheet has been updated inconsistently.
This decision is volume based rather than a hard sell. A single-location restaurant with three or four consistent suppliers processing under roughly 40–50 invoices per month can genuinely manage food cost tracking in a spreadsheet. The weekly data entry burden in that case is around 2 to 4 hours. A pub group with 8 suppliers, weekly deliveries, and 150 invoices a month cannot sustain accruals and invoice reconciliation in Excel, because at 100–150 invoices a month manual processing becomes unmanageable.
UK VAT and HMRC: Spreadsheet Requirements and Limits
As flagged in the Key Takeaways, a plain .xlsx file is not a compliant record-keeping system on its own. HMRC’s VAT Notice 700/21 requires VAT-registered businesses to keep digital records for at least six years, and those records must be complete, legible, and retrievable.
This is where the spreadsheet approach starts to strain. For each invoice, a VAT-compliant record must capture all of the following fields, and each one has to be correct for the record to stand up:
- Supplier name and address
- Supplier VAT registration number
- Invoice number and date
- Date of supply, if different from the invoice date
- Description of goods or services
- Net amount per line item, excluding VAT, noting that simplified and modified VAT invoices do not require net amounts per line item.
- VAT rate applied, such as 20%, 5%, or 0%
- VAT amount
- Gross total
Under Making Tax Digital for VAT, which has applied to every VAT-registered UK business since 1 April 2022, records must sit in functional compatible software that can exchange data with HMRC via API. Manually retyping figures from a spreadsheet into filing software does not qualify as a digital link. Copy-and-paste between a spreadsheet and accounting software is acceptable, while manual re-keying is not.
In practice, invoice data captured in Excel needs to flow into MTD-compatible accounting software, such as Xero, to form a compliant record. The spreadsheet alone cannot carry that load. HMRC accepts digital images of paper invoices provided the image is legible and the audit trail is intact, so a scanned copy retained alongside the structured data usually satisfies the original-document requirement for paper invoices.
From Spreadsheet to Insight: Turning Data into Decisions
The spreadsheet acts as a container, while the value comes from what the data enables. Accurate invoice data supports dish costing that reflects real ingredient prices, supplier price tracking that flags increases in the week they happen, and margin visibility that does not depend on a monthly accountant’s report.
A spreadsheet stops at the download. The real operational win comes when invoice data flows into a system that updates dish costs and margins automatically, so a price increase from a key supplier appears as a red margin flag on every affected dish instead of a surprise in next month’s P&L.
Jelly’s invoice scanning best practices and invoice scanning accuracy guides cover the operational detail of getting line-item data right at scale. The Xero integration page explains how digitised invoices push directly into accounting software.
Common Mistakes and Troubleshooting in Excel Workflows
Hospitality operators who scan invoices into Excel manually tend to hit the same problems, which you can often prevent with a few process tweaks.
- Poor scan quality producing bad OCR. Cause: low resolution, shadows, or angled photos. Fix: scan flat at 300 DPI minimum, use even lighting, and crop to the invoice edges before processing.
- Inconsistent column structures across months. Cause: different team members setting up the sheet differently. Fix: create a locked template with fixed column headers and share it as the single source of truth.
- Missing VAT lines. Cause: OCR reads the gross total but misses the VAT breakdown. Fix: always cross-check the VAT amount and rate against the original invoice before saving.
- Duplicate invoice entries. Cause: the same invoice photographed twice or entered by two people. Fix: use the invoice number as a unique key and add a duplicate-check formula to the sheet.
- Price changes going unnoticed. Cause: nobody re-checks the sheet against previous months. Fix: add a column comparing the current unit price to the previous period’s price, or move to a platform that flags changes automatically.
How to Measure Success of Your Invoice Process
A correctly implemented invoice scanning process captures every invoice, keeps line items accurate, records VAT against each line, and produces data that finance and the kitchen can use without extra manual work.
Useful indicators that the process is working include:
- Reduced weekly admin time spent on invoice data entry.
- Fewer invoice discrepancies identified at month-end reconciliation.
- Faster reaction to supplier price changes, measured in days rather than weeks.
- Clearer margin tracking that does not depend on the accountant’s report.
Jelly users cut food costs by 3% on average in the first three months and save 10 to 20 hours of admin a month. Those figures come from automating the entire flow from invoice capture to dish costing, rather than from a better spreadsheet.
Advanced Tips and Next Steps with Jelly
Scaling invoice scanning across multiple sites depends on standardisation. Use the same column structure, the same naming conventions for suppliers, and the same review process at every location. Without this, finance and the kitchen see different numbers, and the spreadsheet becomes untrustworthy at the moment it is most needed.
For growing restaurants, pubs, and boutique hotels, Jelly replaces the manual scanning workflow entirely. Invoices arrive by photo or email. Jelly digitises every line item, including quantity, SKU, price, and VAT, and the data flows into dish costing, supplier price tracking, and real-time margin reporting without manual intervention. The Price Alert feature flags every supplier price increase or decrease in the week it happens, giving operators the evidence needed to negotiate credits or switch suppliers. POS integration connects sales data to costs, so gross profit margin stays live rather than historical. Jelly works alongside Square, EPOS Now, Lightspeed, and Toast to make that possible.
Jelly onboards and generates initial value in the first week, charges a flat £129 per month per location, and integrates directly with Xero, with Sage coming soon. At Amber, a Mediterranean restaurant in East London, Jelly has delivered consistent savings of £3,000 to £4,000 per month through credits, better buying, and tighter menu controls, with Chef-Owner Murat Kilic describing it as keeping the business alive.
Frequently Asked Questions
How Can I Scan Receipts into Excel?
The same methods used for invoices also work for receipts. You can use Microsoft 365’s Data from Picture feature, a free web OCR tool, or a paid scanning service. For UK VAT compliance, the receipt must show the supplier’s VAT registration number, the VAT rate, and the VAT amount. Simplified receipts covering supplies of £250 or less require fewer fields, but the supplier’s VAT number must still be present to reclaim input tax. HMRC accepts digital images of receipts provided they are legible and unaltered.
What Is the Best App for Scanning Invoices to Excel?
For occasional use, Microsoft 365’s Data from Picture feature is the native option, requires no additional software, and is included in a Microsoft 365 subscription. Microsoft Lens was retired in early 2026, and the current Microsoft-recommended phone scanning tool is the OneDrive app. For higher volumes, dedicated invoice scanning platforms that extract line items, map VAT fields, and integrate with accounting software are more reliable than generic OCR apps, which struggle with photographed receipts, multi-page invoices, and handwritten annotations.
Can I Use OCR in Excel?
Excel includes OCR through the Data from Picture feature, which converts images of tables into editable cells. It is available in Microsoft 365 on Windows, Mac, Web, iOS, and Android. On Windows, it requires Windows 10 version 1903 or later and the Microsoft Edge WebView2 runtime, and it is not available on standalone Excel licences without a Microsoft 365 subscription. Accuracy depends heavily on image quality, with clean, straight-on scans at 300 DPI producing the best results and angled or low-resolution images generating more errors.
How Do I Automate Invoice Data into Excel?
Full automation into Excel is not achievable with file-based .xlsx libraries alone, because they do not run Excel’s calculation engine, execute VBA, refresh Power Query, or touch the COM object model, and it requires a dedicated tool such as Excel MCP Server, which drives the real Excel desktop application through COM automation. Even with that setup, Excel remains the wrong destination for true automation. The Data from Picture feature always requires manual review after each extraction. For genuine automation, where invoice data updates dish costs and margin reports without human intervention, the data needs to flow into a purpose-built platform rather than a spreadsheet. Jelly automates the flow from invoice capture to dish costing and GP reporting, removing the spreadsheet from the process entirely.
How Do I Convert a PDF Invoice to Excel?
You can use Excel’s Data from Picture feature by inserting the PDF page as an image, a free web OCR tool such as Smallpdf or iLovePDF, or a paid invoice scanning service. Multi-page supplier invoices with line items spanning page breaks are the hardest to convert accurately, because OCR engines frequently capture the first page cleanly and lose structure on continuation pages. For UK VAT compliance, make sure the converted data includes the supplier’s VAT number, VAT rate, and VAT amount for each line, not just the gross total.
Conclusion
Scanning invoices to Excel suits low volume, such as a handful of invoices a month, occasional use, and situations where a simple spreadsheet genuinely covers the need. The native Microsoft 365 Data from Picture method supports that use case without extra cost or software, and the five-step process above makes it repeatable.
Past roughly 20 invoices a month, the spreadsheet becomes the new manual burden, and, as the volume thresholds above show, the hours spent on data entry, OCR correction, and reconciliation quickly outweigh the value of having the data in a familiar format. UK VAT and HMRC record-keeping requirements add further complexity that a plain .xlsx file cannot satisfy on its own.
For growing restaurants, pubs, and boutique hotels, Jelly automates invoice scanning so the data flows straight into dish costing, supplier price tracking, and real-time margins, without another spreadsheet to maintain.