A Guide to Receipt Capture and OCR for Expense Reports

Top 10 Questions & What to Expect from Your Expense Reporting Vendor’s OCR Offering

Optical Character Recognition, or OCR as it’s generally known, can be a big help in receipt management. With OCR, receipt images are read by “machine” and the information you need from them is automatically extracted. Ideally, all you have to do to record a transaction is photograph the receipt and let the software take it from there. Easy enough, but vendors differ significantly in their OCR capabilities. That’s why it’s important that you know what to look for if you want to maximize the potential benefits.

1 | Does It Matter Whether I Take a Picture Sideways or Even Upside Down? How Sensitive Is the Recognition Capability to Image Quality?

Say the image is light or there is a wrinkle in the paper. The OCR should be tolerant of expense report filers’ patience and photographic skills. Generally, the quality of receipt photography is a challenge. The OCR software should be up to that challenge. In the first example, the orientation of the image shouldn’t matter. In the second, it should still be successful unless it’s hard for even a person to read the receipt.

2 | How Much Direction Should You Have to Give the OCR? Should You Need to Tell the OCR Whether the Receipt Is a Tape Receipt or Something More Complex?

Once you start having to describe the receipt, you might as well type in the fields yourself.

3 | Can the OCR Be “Taught” to Get Better Without Additional Programming? If It Has Difficulties With a Receipt, Is This Just Something You Have to Accept?

There should be a simple, easy way to add new formats so your vendor can continually improve the success rate of your OCR process. Suppliers of Point-of-Sale systems give their customers effectively unlimited freedom in designing their receipts. The OCR system has to be built to adapt to your procurement environment if it is going to have a significant impact on productivity.

4 | Can You Map the OCR’d Receipts to Credit Card or P-Card Charges?

If you use a corporate card or a P-Card, this is a critical feature. Without it, it is too easy to duplicate entries in expense reports or miss entries entirely. The process of opening receipt images and matching them with card transactions, in addition to being unreliable, is tedious and time-consuming.

5 | Can You Scan Receipts by Mobile, Then Later OCR Them as You Prepare Your Expense Report?

You should be able to OCR receipts from any device or computer. Many times travelers simply want to scan their receipts and email them in. You don’t want to be doing your expense report while the client you took to dinner waits around. OCR should let you work in the way that best suits the circumstances.

6 | Can OCR Be Part of the Same Process You Use to Create or Add to an Expense Report?

Having separate processes and needing to manually sync them up is cumbersome, often slow and undercuts the ease
of use benefits of OCR.

7 | Can You Control Where the OCR’d Receipt Goes?

You should be able to specify a report, a report line item or your repository from which you can later select transactions for reporting. Again, OCR is not a separate process apart from expense reporting and should not impose limits on flexibility.

8 | Can the OCR’d Receipts Be Used for Validation and Auditing?

One of the most cumbersome processes in receipt management is opening receipt images and comparing them to expense report line items. Approvers and auditors either spend excessive amounts of time doing this or they do a hit-or-miss job. Neither is acceptable. OCR can make a huge difference in reducing the manual requirements of receipt checking and may be its key benefit.

9 | What Data Can OCR Capture Off a Receipt?

Typically, you’ll want “vendor,” “date,” “amount,” and “expense type.” Depending upon the type of expense, there may be additional fields. For example, you might require “room rate,” “taxes,” and “meals” from a hotel folio receipt. If it is for car rental, “gasoline” and “daily rate” could be needed. In other words, your OCR should have the flexibility to pull information appropriate to each kind of receipt.

10 | Does the OCR Use the Codes I Configure or Does It Have Its Own Set of Codes (e.g., for Expense Type) That the User Has to Translate?

If a major purpose of OCR is to save time and effort, obviously the answer should be that it uses your own codes.