Why Using OCR For Invoice Management is Like Hitching a Horse to Your Buggy

Spend Matters welcomes a guest post from J.R. Robertson of Coupa

Optical Character Recognition (OCR) technology is a hardware/software tool that takes a paper document, usually an invoice, scans it, and turns it into metadata that can be used to populate fields in a database. From there the invoice can be brought into an electronic workflow. Sounds great, right?

Companies spend millions of dollars on off-the-shelf OCR tools every year in the hopes of processing piles of paper invoices better. But mid-market companies that are using OCR as a strategy are discovering that it’s not really strategic. It’s only a Band-Aid on a much more painful problem.

To be strategic, you need to think beyond just processing paper invoices better. You need to think about a holistic strategy for processing everything associated with the invoice better.

Manual invoice processing might be the most obvious pain point to address, but the problem with OCR is you're just making the work slightly less manual but you're not actually reducing the workload. Here’s why.

If you see a demo of OCR technology, it will be pre-loaded with invoices that are pre-configured to work with the system. That works because OCR does fairly well at reading familiar documents.

In reality, OCR only works well for the roughly 30 percent of your supply base you do business with all the time. For the long tail of your transactions, the one-offs and suppliers you do business with infrequently, it’s not very accurate. Sure, it can be made to be more accurate, but it would be more labor-intensive to get invoices set up and suppliers enabled at an acceptable level of accuracy than it is to process the invoice manually.

In the end, OCR typically costs about 70 percent of a fully automated solution but only provides about 30 percent of the benefit. Going from paper to OCR is like going from riding a horse to riding in a horse-and-buggy. You’re more comfortable in the buggy so there’s some marginal value, but you’re still only going at horse speed.

The car in this analogy is a fully automated system combining an e-procurement front end with a supplier network.

Getting a P.O. into an e-procurement system offloads most of the work to the person making the request, and with a high level of accuracy because AP can reuse the metadata created by the person that fills out the P.O. The supplier network provides numerous ways of allowing suppliers to send electronic invoices and get their information into the system.

This lets you automate two- and three-way matching to create a truly paperless AP process. The invoice comes in. And based on the approval rules you have set, if the invoice and the P.O. metadata match, it automatically gets approved for payment. No one has to touch or validate anything. You have a high degree of confidence that what the buyer in your organization wanted and was approved for was fulfilled by the supplier at the agreed upon price. You can pay that invoice without having to look at it.

This is what I mean by processing everything associated with the invoice better. It’s a more distributed workload. Instead of everything bottlenecking in AP, the work of getting accurate information into the system is distributed across the entities that actually know the most about their part of the transaction. That’s strategic and scalable.

Technology innovation now allows buyers and suppliers to share the work of doing business together. Many suppliers are now equipped to connect electronically and help you streamline processing in the ways I’ve described, but OCR tools won’t help with any of that. They’re really just for paper, maybe some email, so it’s just a partial solution.

A fully automated e-procurement system and supplier network also has a supplier self-service component that is really valuable to AP. With a supplier network, your suppliers have a limited view into your process, so they can see where payment sits without having to call someone on the phone.

The irony is that AP or finance is usually the group that decides to buy OCR. To get to a more strategic solution, there are other factors and stakeholders that need to be considered, primarily procurement and Financial Planning and Analysis (FP&A). In most organizations, procurement and finance are siloed, so AP gets OCR to put a Band-Aid where it feels the pain, but it’s not a scalable solution.

As companies grow, they have two choices to deal with transactional growth: throw headcount at the problem, or invest in technology that will enable the back office to grow capacity and still operate very lean. If you are considering OCR technology, think beyond just solving this immediate need. Connect with other parts of your organization and see the bigger picture. Think beyond ways to process paper better, and think about the best ways to process.

Share on Procurious

First Voice

  1. Harm Poelen; General Manager 20/20 vision:

    In an ideal world I would totally agree with all of the above. There are lots of factors to take into account here; is you supplier capable of providing e-procurement and e-invoicing in a way that is accessible and easy to integrate into internal systems, is the organization buying stuff sufficiently organized to act according to strict purchasing rules, is there any volume being purchased from such vendors and is it a commodity or strategic purchase. All these things and many more make a fully automated Purchase2Pay approach feasible.

    Also the comparison that the speed of the process will not increase (horse versus horse and buggy) is not true. In fact a horse with buggy will move much slower than a horse with man. OCR will improve the whole accounting process in a (partly) paper based environment dramatically; productivity increase of 100% is actual and proven time and time again in published business cases. Yes I do agree that it is suboptimal and in the end we should strive for the elimination of paper invoices. But we should not replace them with only a PDF. That is not electronic invoicing. We should strive towards linking the metadata of the invoice to the human readable and signed version (VAT compliance). So in that case companies can choose to process the PDF as ‘a picture of the invoice’ or import them fully automated into the financial / procurement system. Even if there is no purchase order present, the automatic processing / posting of such metadata will increase the output of the accounting process dramatically and improve on the above mentioned 100% productivity increase.

    But all being said and done; the ideal world is a world where organizations know what they are spending their money on and where they do not waist time on processing and approving invoices for stuff they already agreed to purchase. But like the book from Thomas More describing Utopia it is very difficult to create this perfect world. So every step towards the light is one to be commended.

Discuss this:

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.