*
AMS Adviser *
Volume 4 Issue 3 - May/June 1999
Welcome to a new issue of the AMS Adviser.
Ascent Capture 3.0
Article - The Dynamics of Cost in Document
Capture.
Plus we have all the usual bits (AMS
Services).
AMS
Click here for the information on
Ascent Capture 3.0
Go to Top.
The Dynamics of Cost in
Document Capture
This article has been updated and reprinted in full as it now
includes the just released Version 3.0 of Ascent Capture.
The Dynamics of Cost in Document and Data Capture
Kofax Image Products
April 1999
Putting
Document and Data Capture in Perspective
Understanding the Elements of
Production Capture
Document Preparation
Scanning
Recognition
Indexing and Data Validation
QC and Rescanning
Release
Analyzing Capture Costs
Strategies for Reducing the
Cost of Capture
Use batch processing to speed up
scanning
Use OCR, ICR, or OMR to
automate data extraction
Use database lookups to fill in
data fields
Reducing prep time #1: Use
automatic form ID
Reducing prep time #2:
Use auto contrast adjustment
Use
image cleanup to make images more readable and increase OCR/ICR accuracy
Use validation scripts to
reduce manual checking
Use bar codes to automate indexing
Use OMR for bubbles and checkboxes
Use a separate station for rescanning
Set up indexing and
validation for maximum keying speed
Define document
characteristics ahead of time
Use the right batch size to
speed rescanning
Use
remote scan stations instead of shipping documents to a central site
Reduce the number
of index fields
Reduce the length of manually
keyed fields
Adding It All Up
Table 1. Capital and Labor Costs of Different Production Capture
Operations
Table 2. An Example of Labor Savings via Automation of Production
Capture
Putting
Document and Data Capture in Perspective
Production scanning sites are typically designed with
either document capture or data capture as their primary purpose. The difference between
the two is straightforward:
- In a document capture installation, documents are scanned and permanently
archived. Only a small amount of data is extracted from each image, and its purpose is to
serve as a way of indexing the images for later retrieval.
- In a data capture installation, the documents are usually forms and the images
are not always saved after scanning. Instead, every field on the form is extracted and the
resulting data is then stored in a database or line of business application.
 In general,
capture subsystems of both types are fairly inexpensive to purchase (consisting of little
more than scanners and PCs) but extremely expensive to operate due to the ongoing labour
cost of operating the scanners and validating the data extracted from the images. This
makes them prime candidates for automation and cost reduction.
In both document and data capture, savings are measured in seconds, multiplied by
thousands of documents. How much can one second save? Consider:
- A clerical worker who is paid $10 per hour costs .28¢ per second.
- If you cut one second off the time of processing one document, and you handle 10,000
documents per day, that one second can save you $28 per day.
- At 5 days a week, 52 weeks per year, this amounts to $7,280 per year.
And that’s just for one second. The savings can be truly enormous if you are able
to streamline your operation even more.
Of course, operating cost is only part of the story since the long term costs of
inaccurate data can also be quite large, involving both business losses and legal
consequences. Guaranteeing the integrity of the capture process, while at the same time
reducing its cost, is the goal of production capture software.
Understanding
the Elements of Production Capture
Production capture encompasses a complex flow of processes that includes
scanning but extends much further. In general, production capture includes six operations,
namely, document preparation, scanning, recognition, indexing and data validation, QC and
rescanning, and release.
Document Preparation
Document preparation is an important first step in assuring a well-functioning
production capture process. Key manual tasks include inspecting and separating documents,
grouping documents into like categories, and designating the beginning and end of
documents and batches.
Scanning
Scanning refers to the actual transformation of paper documents into digital images.
Alternatively, existing image files can be imported into the system. Effective scanning
requires precise control over a wide variety of scanners and scanner settings, including
resolution, contrast, simplex or duplex operation, advanced thresholding options, etc.
Figure 1. Production Capture Process Overview
Recognition
There are five primary types of recognition performed by production capture
systems:
- Form ID is used to automatically recognise different form types
- OCR (optical character recognition) is used to recognise machine printed characters
- ICR (intelligent character recognition) is used to recognise handprinted characters
- OMR (optical mark reading) is used to recognise check boxes, filled-in bubbles, etc.
- Bar code recognition is used to read and extract information from pre-printed bar codes
OCR is the most common type of recognition and is generally broken into two types:
zonal and full-text. Zonal OCR is typically used on forms, where only specific
fields on the form are of interest. Full-text OCR is used on free-form documents,
such as legal briefs, to read the entire document and then prepare a searchable, full-text
index of the document.
Image cleanup is also performed in the recognition step. Techniques include:
- Deskewing, despeckling, deshading, streak removal, and other basic cleanup functions
- Line removal and character reconstruction for use on forms
- Edge enhancement, which sharpens character edges to increase OCR accuracy
The purpose of image cleanup is not usually to make the image more readable, but rather
to remove unwanted noise that can decrease the accuracy of automated recognition.
Indexing and Data Validation
Data can be extracted from images automatically via some type of recognition
process or manually by a keyboard operator (an operation known as "key from
image"—typically used when the accuracy of automatic recognition on a zone is
too poor to be useful). In either case, the data must be validated and verified, sometimes
by a second independent operator and sometimes via automated processes such as database
lookups and built-in business rules.
QC and Rescanning
Quality control entails systematic reviews and checks to ensure that the
scanned images are readable. QC includes methods for flagging bad images and explaining
why or how images should be rescanned, and can be performed either by a dedicated QC
operator or by a "key from image" keyboard operator.
Release
Release is the final stage of the capture process, and consists of handing
off batches of in-process images and data to the back end business application. Typically,
this is when the document images are written to optical disk or other long-term storage,
and the associated data is merged with the document database of the larger system. In
addition, the release of a document might trigger a workflow process, initiate the
foldering and filing of documents, etc.
Analyzing Capture Costs
The most straightforward way of determining the costs of production
capture is to examine both the initial capital costs and the ongoing labour costs of each
step in the capture process:
- Capital equipment is the most visible aspect of the cost of capture. Capital
costs include high-speed scanners costing anywhere from $5,000 to nearly $80,000 each,
PCs, specialized image processing accelerators, high-resolution monitors, and so forth.
- Direct labour consists of the people who prepare the physical documents for
capture, scan the documents, check for quality, perform manual keying and data validation,
and integrate the resulting information into the back end system.
The table below summarizes the capital and labour costs involved in each of the steps
of production capture.
Table 1. Capital and Labor
Costs of Different Production Capture Operations
| Capture
Operation |
Capital
Cost |
Labor
Cost |
Comments |
| Document Preparation |
None |
High |
Purely clerical task, but can be automated
using advanced technologies. |
| Scanning |
High |
High |
Capital costs include the scanner, controller
board, and PCs. Labor costs include scanner operators to operate the scanners. |
| Recognition |
Medium |
None |
Capital costs limited to PCs and, sometimes,
accelerator boards. Usually unattended. |
| Indexing and Data Validation |
Medium |
Very High |
Capital costs limited to PCs. Very high labour
costs (typically 2-3 data operators per scanner), but considerable scope for automation. |
| QC and Rescanning |
Medium |
Medium |
Capital costs limited to PCs. Labor costs
include QC operators to inspect images and some additional scanner operator cost. |
| Release |
Low |
None |
Capital costs limited to PCs. If done after
hours, can reuse PCs used for scanning and indexing during the day. Usually unattended. |
Scanning is by far the most expensive capital cost, thanks to the scanners themselves
and the scanner controller cards. The capital costs of the other operations are generally
quite low.
However, as the rest of this white paper demonstrates, labour costs normally dwarf the
initial capital expenditures of a capture system. The biggest culprit is the ongoing cost
of the operators at the scan and indexing stations, and this is the area that should be
targeted most aggressively for cost reduction.
Strategies for
Reducing the Cost of Capture
For the balance of this white paper we will be presenting concrete
strategies for reducing the cost of production capture, piece by piece and second by
second. Not all of these strategies will apply to every installation, but by carefully
choosing the ones that can be implemented at your site you will be able to shave precious
seconds off your capture process and thousands of dollars from your back room labour
expenses.
Use batch processing to speed up
scanning
Batch processing is critical to getting the maximum throughput from a high-volume
capture system. If pages are fed manually and indexed or validated on the spot, the actual
throughput of even a fast scanner can be as little as 5-10 ppm. In a batch operation, by
contrast, entire batches are fed into the scanner, then OCRed, validated, and finally
released. This "assembly line" operation is far more efficient than manual
feeding. Some arithmetic shows why:
- If documents are fed manually, a reasonable estimate for the time to scan and validate
one page with three data fields is:
5 seconds to scan
2 seconds to switch between scanning and validation
12 seconds to index or validate
The total time per document is 19 seconds and the total time for a 100-page batch is 1900
seconds (about 31 minutes).
- In a batch operation, a good estimate for feeding the same 100-page batch is:
30 seconds to load the scanner
150 seconds to scan (assuming a 40 ppm scanner)
30 seconds to load the batch at the index station
12 seconds to index or validate each document, for a total of 1200 seconds for the batch
The total time for the batch is 1410 seconds (about 23 minutes), a savings of 8 minutes.
At a rate of .28¢ per second, you’ve saved $1.37 for just this one batch. If you
process 100 batches per day, 260 days per year, this amounts to $35,620 per year.
Use OCR, ICR, or OMR to
automate data extraction
You should always consider using OCR or ICR to automatically extract data from
documents. This is especially useful on forms, where information (such as a name or an ID
number) is contained in specific locations on the form and can be extracted directly from
the image.
 |
| Ascent Capture allows an administrator to
define zones on a form, which can then be used to automatically extract data via OCR, ICR,
OMR, or bar code recognition. |
The savings from OCR varies dramatically depending on how accurate the
OCR is. To decide if OCR is appropriate for a particular field on a particular document,
you need to figure out whether it takes less time to key the field manually or to OCR it
and validate the OCR results. Here’s how to do the analysis:
- First, figure out how much time it takes to manually key the field. For example, if a
field on a form averages 10 characters in length and your keyboard operators can type
10,000 characters per hour (the most frequently cited industry average), it takes 3.6
seconds to key the field. In real life, you should add about a half second per field, so
figure the total time would be about 4.1 seconds. For 100 documents, the total is 410
seconds.
- Next, figure out how accurate the OCR is on the specific documents you will be using.
The only way to do this is to perform tests on real pages, since OCR accuracy varies
widely depending on how clean the original documents are.
In this example, assume that we have good quality documents and the per-character
accuracy of the OCR engine is 97%. The next step is to figure out the per-field
accuracy. If each character has a 97% chance of being accurate, the chance of every single
character being accurate is .97 multiplied by itself 10 times, or 74%. Therefore, since
the chance of the entire field being correct is 74%, the chance of error is 26%.
- Finally, figure out how long it takes to check each OCR field and how long it takes to
correct OCR errors. The time to check is usually about 2 seconds, although testing might
provide a more reliable figure for your particular site. The correction time is 4.1
seconds (the same as in step 1), but this is only done 26% of the time. Therefore, for 100
documents, the total time is (2 * 100) + (4.1 * 26), or 306 seconds.
In this scenario, OCR pays off, saving 104 seconds on a 100-document batch. However,
OCR is extremely sensitive to both the character accuracy and the length of the field. As
a rule of thumb, if OCR accuracy is less than 95%, or if the field is longer than about 20
characters, you are frequently better off keying the field by hand.
Use database lookups to fill
in data fields
If your capture software permits it, sometimes you can perform a database lookup to
fill in a data field. For example, you might take a last name and a social security
number, feed it into a customer master database, and retrieve the customer’s first
name directly from the database.
This is an especially good technique if the data fields can be used to check each
other. In the example above, if one of the first two fields was mis-keyed, the database
would reject the query since the name would not match the social security number. This
provides a double benefit: the database is checking the accuracy of the first two fields and
it’s filling in the third field automatically.
Reducing prep time #1: Use
automatic form ID
In order to use automatic recognition techniques such as OCR, different document types
must be defined ahead of time so that the system knows where to look to extract the
required data. This means that different document types must be scanned separately, and
this in turn requires time consuming sorting of documents into separate batches during the
preparation step.
Automatic form ID is an advanced technique in which the system "learns"
different form types ahead of time and then automatically recognizes them during scanning.
The system can then sort the scanned images electronically and process each document based
on its predefined characteristics. Manual sorting of document types is completely
eliminated.
Reducing prep time #2:
Use auto-contrast adjustment
Form ID helps to reduce prep time, but there are other factors that may force you to
sort documents into separate batches anyway. One of the most common problems is that
different documents are printed on different types of paper. Scanner settings that work on
white paper, for example, will not work with documents printed on light blue paper. In
this case, automatic form ID does not eliminate the need for document sorting since the
two different document types need to be scanned with different scanner settings.
To resolve this, use a scanner that incorporates VRS (VirtualReScan) technology.
VRS-equipped scanners perform automatic contrast adjustments on each page in a batch and
provide nearly perfect images regardless of the type or colour of paper used for each
document type. Furthermore, for documents that fall below a pre-defined quality threshold,
an intelligent monitoring agent stops the scanner and allows the user to adjust the
settings manually. The scanner then continues with the rest of the batch normally.
VRS produces perfect scanned images on nearly any document type, and automatic form ID
automates the processing of different document types. The two of them together eliminate
virtually all sorting associated with document prep.
Use
image cleanup to make images more readable and increase OCR/ICR accuracy
There are several techniques that can make images more readable and increase OCR and
ICR accuracy. The most effective ones include:
- Deskewing. This technique straightens pages that have been scanned slightly
crooked due to mechanical tolerances in the scanner’s document feeder. Deskewing can
increase the accuracy of OCR by 5-10% or more, which, as we saw above, can be enough of an
improvement to make OCR cost effective compared to manual keying.
- Deshading. OCR engines are unable to read words against the gray shaded
backgrounds that are common on forms. Removing shading allows you to OCR zones that are
otherwise unreadable.
- Despeckling and streak removal. These techniques remove small speckles and
streaks caused by dirt in the scanner feeder or noise in the scanner CCDs.
- Line removal. On typewritten forms, words are frequently typed so that they cross
over the lines on the form, which makes them unreadable to OCR and ICR. Line removal
erases the lines on the image and then reconstructs the characters so they can be
recognized.
- Edge enhancement. This is actually a multiple set of filters that sharpens the
edges of characters. The results are usually invisible to the eye, but they can increase
the accuracy of OCR and ICR by as much as 5-10%.
Overall, by applying the proper cleanup functions for different document types, you can
increase OCR and ICR accuracy by anywhere from 10-30%. This can easily make the difference
between using automatic recognition profitably and being forced to hand key every field.
 |
| Ascent Capture supports a wide variety of
image cleanup options that make images more readable and increase recognition accuracy. |
Use
validation scripts to reduce manual checking
If the accuracy of a field is especially important, a technique known as double key
entry is frequently used: the document is indexed (or validated) separately by two
operators, and the results are compared. If they don’t match, the system displays an
error. However, although this technique is extremely accurate and has been used for years
in the data entry field, it’s also expensive since it doubles the amount of manual
keying. Double key entry is usually used only for one or two critical fields.
 |
| Ascent Capture includes a scripting language
similar to Visual Basic that provides a powerful, but easy to learn, data validation
capability. |
Another way to increase field accuracy is to use automated
scripts instead of (or in addition to) double key entry. A flexible scripting language
allows you to perform anything from a simple accuracy check to a sophisticated database
lookup. For example:
- A simple validation might check to make sure that a telephone number consists of all
digits and is the correct length.
- A more complex validation might compare a city and a ZIP code in a post office database
to make sure they match.
- A third type of validation script might display a list so the operator clicks on an
entry instead of typing it. Not only is this faster than typing, but it is more accurate
as well.
In most cases, validation scripts are not as foolproof as double key entry, but they
are often a good substitute, especially if there are several other fields to act as
backups. They are also useful tools for increasing the accuracy of fields that are not
quite important enough to warrant the additional cost of double key entry.
Use bar codes to automate indexing
Bar codes are far and away the best way to automate the extraction of data from
business documents. A good bar code reading package can read multiple bar codes on a page,
at any angle on the page, with an accuracy of 99.5%+. What’s more, since bar codes
have built in error checking, there’s no need to have operators check the accuracy of
the data. The capture software can do it for you.
Of course, not all business processes lend themselves to bar coding. However, you
should consider bar codes if at all possible. They have proven themselves over time to be
one of the fastest, most accurate, and most fault-tolerant forms of automatic data
recognition.
Use OMR for bubbles and checkboxes
OMR (optical mark recognition) is a technique for automatically reading items on a
form that are either selected or not selected. Example include bubbles on multiple-choice
test forms, checkboxes on credit card applications, and circled numbers on reader response
cards.
Forms often contain dozens or even hundred of these types of fields and keying them
manually is extremely time consuming and error prone. OMR is a highly reliable method of
automating this process, and good OMR engines can recognise a wide variety of filled in
marks.
Use a separate station for
rescanning
If you scan a large number of documents, you should consider a separate rescanning
station for two reasons:
- It’s expensive and disruptive to interrupt the operator of a high-speed scanner. A
production scanner runs at 40-100 ppm, and that scanner is idle while its operator spends
30-60 seconds searching for an original document and then another 30 seconds feeding the
single page. This is wasteful for a scanner station that could have processed an entire
batch of documents in the same time.
- Production stations frequently have only automatic document feeders. Rescan stations
usually require flatbed capability so that poor quality documents can be rescanned with
greater precision.
As a rule of thumb, perhaps .5% - 1% of all documents have to be rescanned, and each
document can take as much as 2-4 minutes to rescan. Why? Consider the steps it takes to
manually rescan a page:
- The QC or index operator must reject the page and write a note explaining the problem.
- The scan operator must shuffle through the batch looking for the bad page.
- The page must be rescanned.
- The scan operator must insert the new image into the batch. If the software doesn’t
allow this, the entire batch must be rescanned.
- The index operator must page through the batch to find the bad page and then re-index
it.
To avoid this, be sure that your capture software has an integrated rescan queue. This
allows notes to be written electronically and pages to be inserted back into batches and
re-indexed automatically. With the proper automation, rescanning time can be cut down to
1-2 minutes per page, a significant savings if your scanning volume is high.
Set up indexing
and validation for maximum keying speed
Many fields cannot be automated using OCR, bar codes, or database lookups. For these
manually keyed fields, make sure that the data entry screen is laid out for maximum keying
speed. For example, do you want to display the entire image on the screen, or just the
zone to be indexed? Should the data entry fields be on the left or the right? Top or
bottom? Should they all be on the screen at once or should they show up one at a time?
 |
| The indexing screen in Ascent Capture allows
the operator to adjust the size and placement of both the image window and the data entry
window . |
There are no answers to these questions that apply to all
cases. In fact, not only will different document types sometimes demand different
treatment, but different keyboard operators will be faster and more comfortable with
different setups.
Your capture software should be able to handle a wide variety of screen layouts to
accommodate different tastes. The payoff can be surprising. Increasing your average keying
rate from, say, 9,000 characters per hour to 10,000, reduces total keying time by 10%. If
you spend a total of $50,000 on keyboard operators annually, this works out to a $5,000
reduction each year.
Define document
characteristics ahead of time
Scanner operators should not have to set up scanner characteristics (resolution,
density, contrast, etc.) for each batch. Instead, the capture software should allow an
administrator to predefine different classes of documents so that the scanner operator
merely picks from a list when a batch is loaded into the scanner.
This technique reduces batch overhead from as much as 1-2 minutes to 30 seconds or less
and also helps reduce scanning errors. Instead of remembering different combinations of
scanner settings, the operator can simply pick from descriptive names such as
"Purchase Requisitions -- Parts Department."
 |
| In Ascent Capture, the administrator
predefines batch classes, which contain detailed information about how to scan, index, and
process different types of documents. The scan operator merely has to pick the
correct batch class from a list and then begin scanning the batch. |
Use the right
batch size to speed rescanning
Every time you put a batch in the scanner there is overhead associated with putting
the paper in the hopper and starting up the scanner. This tempts people to make their
batches as large as possible in order to reduce the batch overhead.
However, there’s a downside to this. If a page turns out to be unreadable and
needs to be rescanned, the scanner operator has to find the page within the original
batch. If a batch is 100 pages long, this is not too difficult. If it’s 500 pages
long, the search could take minutes. For maximum overall throughput, you’re best off
picking a moderate batch size that minimises batch overhead but also minimises search time
for rescanned documents. A batch size of 100-150 pages is usually optimal.
Use
remote scan stations instead of shipping documents to a central site
All too often, production scanning is done at a single central site even if the paper
itself originates at multiple remote locations. This creates two problems: 1) scanning is
delayed since it takes at least a day for the paper to get shipped to the central scanning
site, and 2) the cost of capture is increased since you have to pay to ship batches of
paper to headquarters every day (and possibly ship them back to the remote site when
scanning is finished).
Unless there are special circumstances (such as legal requirements that require tightly
monitored scanning procedures), it is usually cheaper and faster to scan documents at each
remote location. You should choose software that allows scanning and indexing to be done
either at a central site or at a remote site and that allows finished batches of scanned
documents to be transmitted to the central capture site over inexpensive Internet
connections (typically via FTP downloads). Remote sites that scan small numbers of
documents can usually be set up quite inexpensively using low-speed scanners, low-end
capture software, and low-cost dial-up Internet connections. It is not unreasonable to set
up a small remote site for less than $2000, a sum that usually pays for itself within a
few months by eliminating daily shipping charges.
Reduce the
number of index fields (document capture only)
This technique applies only to document capture, not data capture, in which the
purpose of the data fields is to act as indexes that allow you to retrieve the document
later. There are two primary considerations for deciding how many index fields you need:
You must have enough index fields to provide redundancy in case a field
is mis-keyed. In a document imaging system with millions of documents, the index is the
only way to find a document, so a bad index means that the document is lost forever.
On the other hand, you should reduce the number of index fields to save
cost. If an index field takes 4 seconds to type, for example, and you process 10,000
documents per day, that amounts to 40,000 seconds per day, or $112 (at our usual rate of
.28¢ per second). Eliminating one index field can save $29,000 over the course of a year!
In addition, reducing the number of indexes keeps your database smaller, which speeds up
document retrieval times.
Most records management experts recommend at least three index fields per document to
guarantee that the document can be retrieved. In some cases, if you use automated
techniques to guarantee the accuracy of the indexes, you might be able to get away with
less, but three index fields is usually a good rule of thumb. It’s enough to
guarantee accuracy but not so many as to add unnecessary cost to your capture and
retrieval process.
Reduce the length of manually
keyed fields
Can you use shorter fields? If you reduce the average length of a manually keyed
field from, say, 10 characters to 8 characters, you’ve shaved .72 seconds per field
(assuming a keying rate of 10,000 characters per hour). If you process 10,000 documents
per day, that’s over $5,000 per year.
Adding It All Up
A simple comparison shows the potential savings of
applying these cost saving methods to a typical capture site. Consider the cost of a batch
capture system with the following characteristics:
- 10,000 documents per day
- 500-page batches (20 batches per day)
- 40 ppm scanner
- 4 data fields per document, averaging 10 characters each
- 1% rescanning rate
- Manual keying rate of 9,000 characters per hour
The table below compares costs for System 1, which uses little automation, to System 2,
which automates the capture process significantly using features discussed in this white
paper.
Table 2. An Example of Labor
Savings via Automation of Production Capture
| Step |
System 1 |
System 2 |
Comments |
| Document prep |
5 seconds/page Total: 50,000 seconds/day |
3 seconds/page Total: 30,000 seconds/day |
Automated via form ID and VRS preprocessing. |
| Scanner load time |
20 batches per day, 30 seconds per batch
Total: 600 seconds/day |
100 batches per day, 30 seconds per batch
Total: 3000 seconds/day |
In System 2, batch size has been reduced to
100 pages, so the number of batches has increased to 100 per day. |
| Scanner setup time |
30 seconds/batch Total: 600 seconds/day |
None |
Predefined batch classes eliminate manual
setup. |
| Batch scan time |
750 seconds/batch Total: 15,000
seconds/day |
150 seconds/batch Total: 15,000
seconds/day |
|
| Field 1 |
10 characters @ 9,000 characters/hour = 4.0
seconds. Add .5 seconds overhead for total of 4.5 seconds per field Total: 45,000
seconds/day |
8 characters @ 10,000 characters/hour = 2.88
seconds. Add .5 seconds overhead for total of 3.38 seconds per field Total: 33,800
seconds/day |
This field is still manually keyed, but the
average length of the field has been decreased and the keying rate has been increased. |
| Field 2 |
Same as Field 1 Total: 45,000
seconds/day |
OCR field. 2 seconds to check plus .4 seconds
to correct (average). Add .5 seconds overhead for total of 2.9 seconds per field. Total:
29,000 seconds/day |
Using OCR, along with image cleanup to
increase OCR accuracy to 99% (90% field accuracy) saves about 1.6 seconds per document
compared to manual keying. |
| Field 3 |
Same as Field 1 Total: 45,000
seconds/day |
50 incorrect bar codes per day (99.5%
accuracy). 4.5 seconds to correct bad fields. Total: 225 seconds/day |
Bar codes eliminate checking time (since they
are self-checking) and practically eliminate manual keying time for this field. |
| Field 4 |
Same as Field 1 Total: 45,000
seconds/day |
None |
This field has been eliminated through careful
analysis of retrieval patterns. |
| Verify Index 1 |
Same as initial keying. Total: 45,000
seconds/day |
None |
Verification is done via database lookup and
is completely unattended. |
| Rescan |
100 rescans per day @ 4 minutes per page. Total:
24,000 seconds/day |
100 rescans per day @ 2 minutes per page. Total:
12,000 seconds/day |
Rescanning time has been cut significantly
because the smaller batch size reduces the time it takes to search through batches for bad
pages. |
| Total seconds per day |
315,200 seconds |
123,025 seconds |
|
| Total hours per day |
87.5 hours |
34.1 hours |
|
| Total hours per year |
22,764 hours |
8,866 hours |
|
| Total labour cost per year @ burdened rate of
$10/hour |
$227,640
per year |
$88,660
per year |
Annual savings:
$138,980 |
Contact AMS for more
information or click here to look at Ascent Capture Version 3.0.
Go to Top.
AMS Services.
For the complete run down on what AMS can do for you, click on
the following link.
AMS Services
Go to Top.
Next Month
In the next few issues we will have some new articles which will include the following:
- A look at Kodak's Colour Scanner.
- Alchemy Version 6.
- Articles and other topics of interest.
Plus all the usual bits & pieces.
Should you want a topic covered or need an article in full, please
feel free to contact AMS.
Go to Top.
Go to AMS HomePage.
Go to Top |
