By Sramana Mitra and guest author Shaloo Shalini
Sramana: Is your architecture a private cloud, or is it hosted on a vendor’s site?
Scott: Well, we have a mix of both depending on the applications themselves. In the case of some new applications that we may implement in the future, we are having that debate as well. For example, a document management system. There are nice ones out there available for us to use from vendors that offer hosted solutions. However, in my opinion, these vendors want higher fees to run it on their cloud, whereas I can throw down one of my own within my private cloud and buy the software and its updates, right through the year, in a cost-effective way. It makes more sense from a cost standpoint to run my document management system on a virtual private cloud. For future applications, we will have to go through a similar decision-making process.
Sramana: What is the cost differential between running an application such as a document manager on the vendor’s cloud and on your own private cloud?
Scott: In that example, it was pretty darn expensive; they wanted an actual price per page, per document, which is much higher than the storage cost per document. So the decision depends on the number of documents; that really is the one that tips the decision against that vendor in my opinion because if you are talking about a lot of documents you might as well get your own storage yourself. That way you are not paying that high per document cost. However, if you are talking about fewer documents, it makes more sense to throw that one in the vendor’s cloud.
Sramana: What about when you buy the software, what are you options to put it into your private cloud? How comprehensive is the functionality including, archival, backup, retrieval, and all the technologies that you need to do manage documents efficiently?
Scott: That is a good question. On that particular one, I would say that most vendors have been selling their products as software products that you buy and run on your own network and hardware. All the main functionality is in there in the software. It is usually more limiting on the cloud because they generally try to do things via Web pages and in dot.net, which is definitely getting there. But they don’t always put in all the same functionality, to be honest with you. You have to consider both of those things. In the case of document management, it is relatively straightforward with the information and the queries you need to pull; it works well in the cloud environment. You just have to look at the costs in the cloud environment in case of document management system.
Sramana: Have you found a vendor that has the capabilities and is willing to sell you the software that you can put in your own private cloud?
Scott: Yes, in that particular realm, there are quite a few packages out there, such as Laserfiche. They sell the entire package, and it does a lot of the things one would require for optical character recognition (OCR), document recognition, and tasks like that. If you look at some of the hosted vendors like Esker, it does full body text OCR. It is like nothing I have ever seen before. You just have to take a look at the different options that are out there and see which ones make the most sense for your specific use case.
Sramana: When you are talk about OCR, does that mean you are digitizing a lot of different paper documents into electronic format and then storing that data in this document management environment?
Scott: That is correct. It is best to take it even a step further. All you really need to do in our situation is this. For example, there are many documents out there. Let’s say we are talking about bills of lading, signed receiving documents or invoices and things like that. Each of those documents has a primary key on it; this could be something like the invoice number or a routing number. All you need to do is get your document management system to recognize the document through one of those keys fields and then pre-populate all the index fields from your own database based on that primary key. That way you get the ship dates and other similar information populated automatically without having to read those from that document. You don’t need to offer that entire document; with one invoice number you can pre-populate all of the other index fields.
Sramana: How much manual work is involved when you move this particular document management workload to the cloud and transfer paper documents to electronic form? What is the amount of manual work that you are going to have to invest into or budget into this process?
Scott: That is a good question as well. You have to consider this, because you are looking at the cost and the benefits of this move to digitization. Typically, if you are going for another paper-sense system, then there is a lot of manual labor involved there anyways. You are offsetting that with some other manual labor instead. In this situation, someone has to feed the document in, first. If you are receiving it via fax, then the document is already scanned, but if you receive a paper document in the mail, somebody has to physically scan it and try to run the OCR, and recognize where it is supposed to automatically go instead of routing the rules and all that fun stuff. All of that takes a lot of setup and maintenance work. Once we get through some of that, we are working only with the exceptions. With paper documents, we have to talk about 10% exceptions. Probably someone has to look at it, populate the primary key manually, and get that moving to the right approval party or however the arriving process is supposed to work from there.