The business case for OCR and why it doesn’t stack up for invoices

The business case for OCR and why it doesn’t stack up for invoices

Posted by Pete Loughlin in AP Automation, e-invoicing, Electronic Invoicing 03 Sep 2014

This is how the thought process goes for AP automation:

“Electronic invoicing could save us lots of time. We could automate accounts payable.”

“Wait – this could be to be complicated. We’d need a project manager and an expert. We may have to buy in some software or work with a third party. Actually this stuff doesn’t come free. It could cost a fortune.”

“$1 per invoice must be cheaper than the cost of a paper and manual process but at 1 million invoices per year, we’ll never justify that. Let’s make an incremental step toward automation. We can scan our invoices and handle them digitally.”

Yeah, right!

Purchasing Insight logoe-invoicing, whether it costs a dollar or half a dollar per invoice, is cheaper to work with than a paper and manual process. It’s basic common sense. From a business process perspective, it is far more efficient for all parties. The problem that faces people in business is how to convince their boss. How do you convince somebody that something that is obvious to you should be just as obvious to them – especially when it involves spending money? How do you say that you can transform your accounts payable function and reduce costs by spending $1 million without sounding like you’re saying “This Chanel handbag is a steal at $1,500”.

So it’s easier to go for the incremental change. Scanning paper invoices doesn’t involve supplier adoption and it’s not such a change in business process that it will involve complex change management. There are some great business cases for document management (so the capture sales guys will tell you) and the technology has evolved over recent years such that OCR accuracy is virtually 100%

A little aside here – if you really want to know how effective document management, capture, OCR – whatever you want to call it – can be, the last people you want to speak to are the sales guys. They’ll tell you lies. I’m not saying they’re liars – they’re really not – they are just deluded. I once had a conversation with a capture sales guy. He’s a lovely man – I like him a lot. He was raving about his mobile capture software, telling me how accurate it was and how 21st century, leading edge everything it was. I downloaded the app. I tried it. It didn’t work. On a scale of diamond to horse shit, it was bad horse shit. Maybe it was me. Perhaps I was doing something wrong. So I had another conversation about it with the marketing director of the same company – this time after a few drinks. “The only problem we have with this software is” he said, “it doesn’t work”. In vino veritas.

Back to the business case. OCR works really well for a use case like a mortgage application. A 6-page application form can take what seems like hours to complete and for a data input clerk, minutes to input. It could be a minute – it could be 3 minutes but compared to the fraction of a second it takes a good document management system to interpret it, it’s a long time. Building a business case for document management for mortgage applications is a simple. Replace minutes with fractions of a second and you have an order of magnitude increase in the number of mortgage applications that can be dealt with – that means more decisions, quicker. More business. Greater market share. More profit. The business case is clear.

Same for invoices right? Actually, there’s a problem with the invoice business case. Fact is, even with a paper process, it only takes seconds to key an invoice into a finance system. Sure, OCR is pretty accurate  – and faster – but if you replace seconds with fewer seconds, the business case isn’t quite so compelling. And when you really get in to it, to replace those seconds employed in the manual process with the fewer seconds on the OCR process, you really have to optimize the life out of your system – and that takes much more than seconds to achieve.

OCR might seem like an incremental step towards AP automation but in reality, it’s a sideways step, and while e-invoicing might seem daunting, it is the only route to automation that will actually deliver any real business benefit.

Pete Loughlin can be found on twitter @peteloughlin

  • Nicole Piper September 4, 2014 at 11:27 am /

    Are you saying I should change my career to accounts payable because they are already automating purchasing and order picking???? 🙂

  • Steve Britton September 4, 2014 at 2:33 pm /

    Great article and having been in the IDC market for 20+ years and yes as a ‘Sales Man’ I have to add that the 3rd gen approach has to be ‘Multi-Channel’ so e-Invoice (PDF, XML, EDI Etc) Paper, Portal (PO Flip) etc and so on. Restrict the platform to ‘e’ only and I do not believe that this will deliver short or medium term benefits. I typically see vendor on-boarding rates below 20% and STP rates in a similar category, whereas a ‘Multi-Channel’ approach can deliver vastly superior performance levels such as 40%+ STP and >100k invoices per FTE per year and over time the platform can be used to encourage and gain high adoption rates? Also the ‘e’ approach must not charge the vendor or demand a certain file structure, this imposes a barrier for the vendor. So test the business case and see what outcomes will be delivered through either approach AND then make an informed decision on which solution to adopt, hope this helps?

  • Richard Fitzwilliam September 5, 2014 at 11:27 am /

    I agree with Steve Britton that a multi-channel approach is the optimal way to receive and process invoices today.
    10 years ago, yes, scanning & OCR was at the top of its game, the technology was just about there for invoices and einvoicing hadn’t taken off. Even today looking in isolation, scanning & OCR can provide a healthy ROI, increase STP and increase invoices per FTE. But technology has moved on, there are more options.
    Were in violent agreement that eInvoicing / EDI is the optimal way to receive invoices today. So why invest in scanning & OCR technology when your paper volumes are reducing? Far better to outsource this or have the einvoicing provider handle it for you.
    Critically, einvoicing comes with something scanning & OCR doesn’t provides; feed back to the sender. The sends wants to know the invoice has been received, is being processed, if there are errors, when they are going to be paid, etc. – a good einvoicing platform will tell them that; scanning & OCR is a comparative black hole.

Post a comment