Learn 5 Reasons Why the OCR Software is Very Expensive

Written by: Gonçalo Caeiro   (4-min reading)

At DocDigitizer we are on a mission to Unburden Humans from Robotic Work. That translates in building great tech that help companies to eliminate robotic work. It also translates in making the tech simple and accessible to every single business no matter its size or pockets.

Ever thought why OCR software is not in every single company? Ever thought why only very large corporations have access to it? The reason is simple: the Total Cost of Ownership of an OCR project is huge. We will navigate step by step in the cost breakdown. The cost depicted below are an annual average. This means that considering a life span of 4 year, the percentages depicted below are the annual contribution of each item. We laid out the steps sequentially from the perspective of launching a new project although the costs refers to the annual contribution.

Step 1: License the OCR Software

The cost of license OCR software starts in the thousands. It can easily reach hundreds of thousands depending on the volume and capacity to process, scan and extract information from documents. Unlike DocDigitizer which is a service based only on a pay-per-use basis.  When deploying an OCR project you must license immediately for the full max capacity you will run. The vast majority of OCR software providers are not able to deal with fluctuations in volumes from the license cost perspective.

Typically, this cost is about 5% of the project.

Step 2: Acquire or Rent hardware / virtual machines to run the OCR Software

The software by itself does not do anything. You need to deploy the OCR software into an environment. Now you are adding some thousands of dollars to the TCO. It’s not the servers alone, either cloud or on-prem. You will need to have someone manage them, adding them to your secure infrastructure, guarantee operations, backups and so forth. 

Unlike DocDigitizer where you pay by document extracted fully validated, with an OCR software deployed project you are already at this stage spending thousands of dollars and still don’t have anything running.

Typically, this cost is about 10% of the project

Step 3: Consultancy / Development services to implement OCR Software

At this point in time your OCR project cost is already above 100k USD but most likely exceeds 500k. Your need consultancy services to configure the software develop templates and alike taking into consideration the type of documents you need to handle and the type of information you need to extract.

With DocDigitizer you are paying by document extracted and validated. But with an OCR software project you will end up paying for building and tweak each document template.

It is frequent at some point the project to learn that the consultancy team just spent fine-tuning a template during 2 days that you only get 50 documents per month of that type, but the bill for that document template is about 2k. You may say this could be poor project management, but when your project is dealing with dozens or hundreds of templates, we can guarantee that these scenarios are far from rare.

Typically, this cost is about 30% of the project.

OCR Software Project Cost Breakdown

Step 4: Build and manage the OCR editors / revision team

The project is already in the third month and running late. Multiple meetings and headaches regarding servers, installation, development and additional consultancy fees are making the project already into the red line. At least you are ready to turn the lights on the process some documents right. Wrong!

After all the OCR software is not that accurate. In an optimistic view, you will get documents (meaning with all the fields rightly extracted) at most with 80% accuracy. At this point, you know you must to deploy a revision team. The financials at this stage become fuzzy. To which cost center you allocate the yearly cost of the OCR revision team?  After all, it is OPEX and maybe you allocate that to the OPEX center of the business unit to which the documents refer. 

Soon you understand that the revision team in an operational nightmare. You need to account for uneven quality from different editors, holidays or absences and worst of all, peak demand. But you’ll learn your team size cannot easily expand or contract.

Typically, this cost is about 40% of the project.

Step 5: Maintenance and Support of the OCR Software Project

Considering a project lifetime of 4 years you will pay for the software maintenance fees and changes required to document template configuration and so on. A myriad of costs will have to be still to be account for, between managing teams, HR and IT.

Typically, this cost is about 15% of the project.

Bottom Line: The True Cost of an OCR Software Project

When starting to compute the costs of an OCR software deployed internally is easy to be attracted by a potential low cost of licensing or renting an OCR software solution. After 4 years running such projects, customers usually reach the conclusion that the TCO is 20 times the cost of the OCR software itself. When you take the cost of opportunity into account, ie, the time you took to launch the project and the limitations it as, that cost raises from anything to 40 to 80 times. 

With DocDigitizer you can be up and running in hours, with zero setups, full guaranteed accurate extracted data and full pay-per-use. You can scale up or scale down, no troublesome deployment and maintenance projects and you have the guarantee to run on the best productivity ratio enabled that AI/ML delivers.