Best Practices for Paper- Based Form Design

Best Practices for Paper-Based Form Design 1

Best Practices for Paper-

Based Form Design

For Improved Document Capture ROI

Author: Scott Maloney, Senior Project Manager

Published September 2017

Best Practices for Paper-Based Form Design 2

TABLE OF

CONTENTS

Introduction................................................................3

Preparing for Form Design........................................4

Questions to Consider..........................................5

Field Type Considerations.........................................6

Character Boxes...................................................6

Choice Fields.......................................................7

Image Zone.........................................................8

Barcodes.............................................................9

Focused Recognition and Data Validation...............9

Other Optimization Recommendations....................9

Limit the List of Acceptable Characters for a

Given Field.........................................................10

Put a Form Identiﬁer in the Same Location on

Every Page.........................................................10

About the Author......................................................10

About Pyramid Solutions, Inc.................................10

References...............................................................10

Best Practices for Paper-Based Form Design 3

n recent years, online activity exploded, but despite

all of the rosy predictions about how we will become

a paperless society, we still generate more paper than

ever!

So, what is getting in the way of a faster movement to

a paperless society?

A lot of interaction with customers is still driven by

paper-based forms and there are challenges that

come with this type of interaction.

The person ﬁlling out a form is out of your control. No

matter how well you design the form there will always

be responses that cannot be read automatically. You

can encourage the form users to write neatly and keep

their responses within the allotted space, but there

will always be people who don’t read instructions (or

don’t care), and assume the form will be read and

processed by a human, not by a computer.

They do things like write a character by mistake then

draw a big “X” or scribble over it to “delete” it. In a

day and age where more and more communication

happens online with a keyboard, mouse or touch

screen, handwriting is getting more and more messy.

However, this challenge creates an opportunity for

organizations to re-design forms so they can optimize

their capture investment.

Not only will this make the process easier, but as

an organization redesigns more and more forms, it

will reduce the most expensive part of the process:

manual intervention required during validation to

make corrections to captured information before

exporting the document to the corporate repository

for processing.

This white paper seeks to provide meaningful guidance

to companies that want to reduce the human capital

costs associated with manual veriﬁcation of incoming

document images by re-engineering their forms.

By following these guidelines, you will notice more

accurate written information, a smoother capture

process, and a cleaner corporate repository.

PREPARING FOR

FORM DESIGN

here are many factors to consider when designing

a form to collect handwritten responses. Your

target audience needs to easily understand the form

which means where the user is supposed to write their

responses is obvious.

The industry average for intelligent character

recognition (ICR) is about 70 percent. You can never

expect 100 percent accuracy, but shooting for 85

percent is considered good (even though it’s still

15 bad characters out of 100). Good form design

planning can usually exceed the 70 percent threshold.

When designing new forms, try to maintain a balance

between aesthetics and readability. Every form is

unique and should be approached with a cost-beneﬁt

analysis. The goal of form design is not to eliminate

all data entry but to create cost-efﬁcient forms. There

will be at least a small degree of veriﬁcation required

on any form.

Take time to plan how to design a form. You should

identify the ﬁeld types required for layout prior to

development. The speciﬁc purpose for each form

plays a signiﬁcant role in how to develop it.

Readability is important when designing any new form.

Always keep in mind the intended audience and the

form’s intended purpose. Simple user instructions can

signiﬁcantly improve recognition rates.

• Indicate on the form that a computer is going

to process it

• Stress the importance of writing plainly,

carefully and clearly

Best Practices for Paper-Based Form Design 3

Best Practices for Paper-Based Form Design 4

• Ask for block letters on the form

• Put instructions in bold at the top of the form

• Show examples such as how to write an “A”

or “2”

Planning Steps

1. Identify and meet with the form owner. If this is

for an existing form, determine whether you can

redesign the form.

2. Obtain copies of all versions

of an existing form.

3. Identify the goal or intended

purpose of the form and the

form’s audience.

4. Work with appropriate

technical personnel (DBA,

application developer, etc.)

and representatives from

the business (business

analyst, subject matter

expert, etc.) to obtain a list

of all required ﬁelds and

ﬁeld properties.

• Field properties should

include maximum ﬁeld lengths, ﬁeld types,

and required ﬁelds.

5. Identify the capture software features and

functionality that the form will go through.

6. Involve all parties that are directly connected to

the form manipulation before it gets to the OCR

system.

• For example, a customer may send a ﬁlled out

form to a speciﬁc business department where

the personnel will then review, and possibly

add additional data/stamps/other markings.

If the customer ﬁlls out the form neatly, but

the department personnel place stamps on

important data, OCR accuracy will be affected

dramatically, and will not be as good as it

should be. The personnel above should be

trained to know this, and perhaps the form

design should include ﬁelds or areas named

“Internal Use Only” where they can place any

stamps, writing, etc., so it doesn’t interfere with

other clean data on the image.

Questions to Consider

1. Who is the audience? Do they require

special consideration like larger fonts

or extra space between ﬁelds?

2. Who owns the existing form? Does this

prevent redesign?

3. Are aesthetics an especially important

feature for the form?

4. What is the form’s intended purpose

or goal?

5. Approximately how many ﬁelds will be

on the form?

6. Are there any areas on the form that

require free form entry?

7. Do previous versions of the form

exist, or is this a new form?

8. Can the layout of previous forms help

with the template for the new form?

9. Do you have to capture data from an existing form

until you can implement a new form?

10. What is the expected submission volume of the

forms?

11. Do the forms require peak processing periods

(e.g., certain months of the year will generate huge

volume while other months will be comparatively

light)?

12. Can you break a packet of forms into multiple

form types?

13. Does each form have a unique identiﬁer to link the

documents together after the capture process?

14. Is the form a multi-page document? Do the pages

have to be linked together after processing?

The industry

ave rage

for correct

intelligent

ch a rac te r

recognition is

only about 70

percent.

Best Practices for Paper-Based Form Design 5

15. Will a double-sided form be used? Keep in mind

the quality of paper can impact recognition

accuracy in double-sided forms. Standard 20#

copier paper works well for single sided forms and

24-28# paper works well for double-sided forms

to prevent the back-side content from bleeding

through when scanned. Fields may also be offset

to ensure that any bleed-through content from one

side will not interfere with the ﬁeld recognition on

the other.

16. Will you print the form internally or at an outside

vendor?

17. Do you know the ﬁeld requirements for every

ﬁeld on the form? Where can you obtain this

information?

18. Which parties will enter data on the form?

(i.e. customer, internal department ((received

stamping, other, etc.))

FIELD TYPE

CONSIDERATIONS

roper layout of the printed response areas

signiﬁcantly impacts the accuracy of hand-printed

content recognition. A common mistake in ﬁeld design

is to provide a free form area for response (i.e. a

simple blank line where people should write). Without

any character restraints, people will run characters

together and write in cursive, on top of the line and in

multiple lines which signiﬁcantly impacts recognition

accuracy.

A well-designed form has a deﬁned response area

for each character, encouraging character separation.

Character boxes, comb lines and choice ﬁelds yield

the highest degree of character recognition.

A form can use image zones to capture and recognize

free-form text, however these zones require signiﬁcant

testing and will never achieve the caliber of recognition

levels associated with constrained ﬁelds. Veriﬁers often

have to correct or input handwritten information as

free-form text into free-form zones on a form.

Character Boxes

Character boxes are the best method to encourage

character separation. A good character box design

allows users to write characters completely within each

box. Unfortunately, many forms contain boxes that are

too small and too close together. People often can’t

write small enough to keep an entire character within

each box. Pencil lead creates strokes that are usually

wider than pens, making it even harder to constrain

the character.

The following are some general guidelines:

Each box should be square in shape, and the line

width should be 1 pt. Thicker lines run the risk of

leaving partial lines behind when the document goes

through the image enhancement process. Rectangular

boxes with a height taller than the width can make the

user feel like they need to squeeze their characters

into the space. This often results in characters written

in a compressed vertical form, reducing accuracy. A

square shape encourages wide, formal characters.

Narrow boxes Square boxes

Create single character response locations, such

as for Male (“M”) or Female (“F”), in a single box

separated from other responses.

Best Practices for Paper-Based Form Design 6

Separate individual ﬁelds with enough space to easily

identify where one ﬁeld stops and the next one starts.

I recommend spacing of at least 1.5 box widths to

prevent users from interpreting the space as a valid

character location.

Be sure to separate rows of ﬁelds stacked vertically by

at least one-half the height of an individual box.

You can print boxes with either solid black lines or

dropout colors, depending on the scanning and forms

processing technology. Software-based form dropout

removes the boxes from the image after scanning.

Some forms processing systems require dropout

colors when printing forms. For example, a form might

be in red ink and a red bulb in the scanner eliminates

the red content during capture.

Comb Lines

Comb-style ﬁelds consist of short lines between each

character. Although a commonly used format, people

rarely align the characters within each space. The

spacing between the vertical lines is often too close

together, making it almost impossible for the average

person to stay between the lines. If using comb lines,

provide plenty of space between the vertical tick

marks. Make the tick marks tall enough to encourage

people to write between them. A vertical height of

at least half the height of the expected character is

usually sufﬁcient.

Example of a poor comb line.

Example of a good comb line.

Prior to developing a form, obtain a copy of the

database schema for the backend system. You will

need to specify ﬁeld lengths, data types and required

ﬁelds. It is important that these attributes are consistent

with the backend environment. Field lengths can be

shorter than the limits deﬁned in the backend system,

but not longer.

Every field type has a confidence level (%). A

conﬁdence level measures how certain the read of

information on the ﬁeld must be before bypassing

veriﬁcation. You can set conﬁdence levels high at the

beginning and cautiously lower them as veriﬁers see

over time that the information reads are accurate and

seldom require manual correction.

Choice Fields

Choice ﬁelds consist of a list of options. Choice lists

can be set up for a single or multi-choice response.

The form recipient chooses from the options by

selecting the appropriate response mark. Examples

of response marks include bubbles, boxes, brackets

or lines.

Gender A B C

O Male 1 O O O

O Female 2 O O O

3 O O O

4 O O O

Bubble ﬁelds always yield the highest degree of

accuracy. The choices to select on the form may be

simple, but keep in mind you can conﬁgure the ﬁeld

to export detailed values based on the option selected

on the form.

Best Practices for Paper-Based Form Design 7

The form designer can tie each response mark to

a speciﬁc value. You can customize the range of

values, such as A-Z; 1-5; A-E; 1-9. Choice ﬁelds

are often used in multiple choice tests, surveys and

standardized tests. They are extremely accurate,

but may be inefﬁcient for entering alphanumeric and

alphabetic text.

Optical mark recognition (OMR), sometimes known

as “mark sense,” is the analysis of form locations to

determine if a mark is present. Whether you use oval

bubbles, square boxes, open brackets or signature

blocks, be sure the area is large enough for the user

to easily mark within the designated area.

Common OMR ﬁeld design errors include making

the box or oval too small for people to easily mark

within the zone, or printing the boxes or ovals too close

together, resulting in more than one space containing

the mark written by the user.

Bubbles between 10 and 14 points in height work

the best. I recommend using a capital “O” in an Arial

font (do not use the Times New Roman font). A non-

oval or perfectly round circle is somewhat harder to

ﬁll completely, so respondents will subconsciously

be compelled to ﬁll in the circular bubbles more

completely and neatly, leading to better recognition

rates.

It is possible to place numbers or letters inside each

shape, however the number/letter must be as small

and light as possible. Dark, thick or bold characters

may cause false positives because the OMR interprets

the space as ﬁlled during scanning.

Spacing

Allow at least 3/8 inch between any text, lines or

graphics on the form and any bubble areas, OCR text

or barcodes. Minimize potential errors by separating

bubbles from one another by at least two character

spaces.

Try to stay away from using lines or boxes around

or between the bubbles on the form. If you ﬁnd it

necessary, consider making the lines a light gray or

red that will drop out (completely disappear) during the

scanning process. Keep in mind that if people try to

make copies of forms with shaded areas, those areas

will appear darker on the copies.

Some users will circle an OMR response instead

of ﬁlling in the box. Similar to character responses,

providing clear instructions and example marks will

signiﬁcantly improve recognition results. Even great

instructions will not prevent some people from marking

a zone in error then drawing a big “X” in an attempt to

make a correction. You will need to develop business

rules to handle mark situations and manual key-from-

image operations to determine user intent and make

corrections.

Image Zone

Image zones are extremely ﬂexible and useful when

designing forms. Image zones are for capturing

information that is best entered manually. Veriﬁcation

operators have to read the written response and type

it into the capture software.

Email Address

Other Comments

Keep in mind that image zones are never as reliable as

character boxes for OCR and ICR. Use image zones

for any of the following purposes:

• Reading hand and machine print

• Reading or printing barcodes

• Calculating the percentage of an area ﬁlled

• Saving an area of a form as an image or BLOB

Best Practices for Paper-Based Form Design 8

Within image zones you improve recognition by

specifying font type and if the expected content is

going to be machine printed versus hand printed.

Image zones combine the reliability of manual data

entry with the efﬁciency of automated recognition and

data validation. For example, the veriﬁer can draw a

band around an area to display it for veriﬁcation and

to allow for manual entry into the appropriate ﬁeld(s).

A large image zone can be set up for multiple

handwritten lines of free-form text or when a set of

ﬁelds is either difﬁcult or inefﬁcient to recognize. Input

options are tied to the image zone. For example, a

veriﬁer may manually input a social security number

that was handwritten into the middle of an image zone,

or select a value from a choice list on the validation

screen.

Barcodes

Barcodes help classify document or form types that

are scanned. Barcodes come in 1D and 2D formats

and can be horizontal or vertical on a form.

Barcodes should be at least 26 points in height.

Similar to OMR bubbles, leave at least 3/8 inch

of space around your barcode. Remember that

barcodes may vary in length depending on what you

are capturing. Leave room for the largest expected

barcode both on the form and when the region

deﬁnition is conﬁgured in the capture software.

FOCUSED RECOGNITION

and DATA VALIDATION

ome ﬁelds are designed to allow only certain

characters. For example, a date ﬁeld may allow

only digits, or only digits, dashes and slashes. A

“Male/Female” ﬁeld may only allow the characters M

and F. Ensure that your form contains instructions

or examples for each ﬁeld to ensure the user knows

what characters are allowed. You can factor the

deﬁnition of allowable characters into the veriﬁcation

software so it ﬂags invalid characters for review.

Low conﬁdence data typically requires a “key from

image” process to validate. This requires displaying

suspect characters or ﬁelds to a human for manual

data entry. Human interaction is the most expensive

part of any data capture process, so any efforts you

can take such as strong form design or additional

image enhancement processes, will easily pay for

themselves when compared to the cost of manual

data entry.

OTHER

OPTIMIZATION

RECOMMENDATIONS

Limit the List of Acceptable

Characters for a Given Field

xpected characters can be alpha, numeric or

alphanumeric. Specify whether the received text

will include hand or machine print. Deﬁne which

special characters you anticipate receiving on the

form. For example, a telephone number will never

include alphanumeric characters and should be limited

to numeric. You may also want to identify the dashes

within the list of expected characters if you did not use

a constrained-print approach for the ﬁeld.

Best Practices for Paper-Based Form Design 9

Put a Form Identifier in the Same

Location on Every Page

When scanning, indexing and verifying multi-page

forms, you can improve the accuracy tremendously

by putting a form ID in the same location on every

page with sufﬁcient white space around it. Format the

form ID in a way that facilitates easy recognition of the

pages. For example, form ABC-1 for page one, ABC-2

for page two and ABC-3 for page 3.

ABOUT THE

AUTHOR

cott Maloney has held various management

positions in marketing, operations and information

technology with ﬁnancial services companies for more

than 20 years. During his tenure with Pyramid Solutions

as a Senior Project Manager, Scott has managed

enterprise content management projects that resulted

in the deployment of new solutions across a spectrum

of state government, insurance and banking clients.

ABOUT

PYRAMID SOLUTIONS, INC.

yramid Solutions develops products and

innovative solutions for organizations in a wide

range of industries – from ﬁnancial institutions to

insurance providers to automotive suppliers and

industrial automation companies. Our experts

specialize in three core practices – Enterprise Content

Management, Intelligent Manufacturing and Network

Connectivity. For the past 25 years we have developed

visionary solutions that yield exceptional results. For

additional information about Pyramid Solutions, visit

PyramidSolutions.com.

September 2017

prohibited. Pyramid Solutions, the Pyramid Solutions

logo and Pyramid eXpeditor™ are trademarks of

Pyramid Solutions, Inc.

REFERENCES

Scanlan, Rick. “Best Practices: Improving ICR

Accuracy with Better Form Design.” Published

August 25, 2015.

https://www.accusoft.com/whitepapers/best-

practices-improving-icr-accuracy-with-better-form-

design/.

Remark. “Form Design Best Practices.” Accessed

August 8, 2017.

http://remarksoftware.com/support/ofﬁce/form-

design/

Verity. “Scannable Form Design Best Practices.”

Published April 2005.

http://www.consiliumdcs.com/tools/spaw/ﬁles5/

Forms%20Design%20Best%20Practices.pdf

Best Practices for Paper-Based Form Design 9