It’s What’s Between the

Tags that CountsControl Information Quality to reduce

translation cost and time to market

Andrew Bredenkamp (standing in for Kent Taylor)

CEO - acrolinx

Overview

• Abbreviated History

• The Result - Global Information Quality and Cost Issues

• Quality Management Principles

• Information Quality

• Information Quality Management Tools

• Q&A

The Drivers

Technology

• Analog → Digital

• Internet

Philosophy

• US only → Global

• Collaborative Scheduling → Market-driven deadlines

• Quality at any cost → Do more with less

What’s going on?

• Independent writing groups g interconnected, distributed teams

• Print g digital media (CD, Web, Help)

• Unstructured authoring g structured authoring

• Monolithic “documents” g topic-based writing (supporting reuse and

repurposing)

• feature-oriented g task-oriented

• “Writing + Translation” g global information development (also on-demand)

• “Product development + documentation” g coherent “user experience”

The Result:

Process Change

The way it was …

• Collect Source Material

• Write Draft -> Edit Draft -> Review Draft

• Write Final -> Edit Final -> Review/QA Final

• Produce Master -> QA Master

• Reproduce – QC (Statistical Sampling)

• Distribute

The way it is …

• Write -> Hit RETURN -> Global Distribution

And quality? What quality?

• Paper or web style guides are not consistently used

• Terminology is in the wrong place

• Reuse is not granular enough

• Risks and costs of quality are not appreciated

• 5-10% of global content is under QA

• 25-35% of content (including translation memory content) is redundant

• Every new language costs ≈ 30% more than it needs to...

• ... but that’s where the growth is!

So what can we do about it?

• Manage Quality

• Create content with quality built in. Information Quality Management consists

of three key components:

• Best practice writing standards (spelling, grammar, style)

• Terminology (technology and branding)

• Intelligent Reuse (consistent wording)

The Bottom Line Part 1:

Doing more with less costs more!

• We’ve effectively applied technology to increase volume and

speed of information development, production, and distribution,

but ...

• Information Quality has suffered, and the costs associated with

poor quality increase daily

• Translations cost more than necessary, and may slow your globalization

strategy

• Non-native speakers have difficulty reading your English content, and are

less likely buy your products

• Misinterpretations of your content can lead to increased support costs,

accidents, and/or litigation.

• Every shortcut taken in authoring adds a corresponding cost -

multiplied 5, 10, 15 … 30 times or more - in

translation/localization.

The Results - Readability

Case Study:

• In one 1,100,000-word software manual set:

• 15% of words were simply not needed

• “We suggest that you use…” > “Use”

• Translation: $40,000 per language

• Production: ~330 pages less (~$66) per copy (~ $66,000)

• 30% of sentences were at a graduate-school reading level

• 50% of sentences were at a college reading level or higher

• 85% of sentences were at a high-school reading level or higher

What does this mean in practice?

http://support.microsoft.com/kb/291875

http://office.microsoft.com/en-us/powerpoint/HA012303211033.aspx

Content reuse (how not to)

... and this:The Status End Date is either earlier than the Start Date

of the Assignment or later than its End Date.

Please enter an end date that is later than the start date.

Date To must be later than or equal to Date From.

The Date To must be later than the Date Received.

The actual end date must be on or after the actual start date.

End date should be greater than start date.

End Date cannot be before the Start Date.

The start date must be prior to the end date.

You entered a start date later than the end date.

Ending range must be later or the same as starting range

Please enter a new start date later than the original end date.

The ending date must be later than or the same as the

beginning date.

The date to has to be later than or equal to date from.

End Date must be greater than Start Date.

You cannot enter an "End Date" that is before your "Start Date."

End Date must be greater than or equal to Start Date.

Please enter a start date that is before the end date.

The end date you enter must be between the grade's start and

end dates.

The start date you enter must be between the grade's start and

end dates.

The projected end date must be on or after the projected start

date.

The Period start date cannot be later than the Period end date.

Please enter an actual start date earlier than the actual end date.

The Start Date cannot exceed the End Date.

The End Date cannot precede the Start Date.

End Date must be later than Start Date.

End date must be equal to or later than the start date.

The end date must be later than or the same as the start date.

End Time must be later than the Start Time.

The valid grade's end date must be later than or equal to its start date.

Please enter an End Date that is later than or the same as the Start Date.

Competence end date has to be later than or equal to the start date.

The start date cannot be later than the end date.

The appraisal end date must be later than or equal to appraisal start date.

The Effective start date cannot be later than the Effective end date.

Date from cannot be later than date to.

The start date must be on or before the end date.

The Start Date cannot be after the End Date.

Your end date must be after your start date.

The end date cannot be before the start date.

Start date must be before end date.

Your start date must be before your end date.

Enter a Start

Please enter start date before or equal to the end date.

Please enter a start date earlier than or the same as the end date.

The Result – Translation Issues

Ladies, leave your clothes here and spend the afternoon having a good time. (In a Rome laundry)

Please fasten seatbelt to prepare for crash. (Tokyo Taxi)

Please do not feed the animals. If you have any suitable food, give it to the guard on duty. (at a Budapest zoo)

Caution: Blade Extremely Sharp! Keep out of children. (Multipurpose Knife Instructions)

Ladies are requested not to have children in the bar. (In a Norwegian cocktail lounge)

Drop your trousers here for the best results. (Thai Dry Cleaner)

Product Information Ecosystem

Marketing

& Pre-sales

Training

Material

Technical

Documentation

Support

Knowledgebase

FAQ

Regulatory

Information

Market

Analysis Technical

Specifications

Partner/OEM

Information

“In-product”

Information

The Result – Translation Cost

© 2006 Comtech Services, Inc.

46%

25% reduction

15% reduction

13% reduction

9% reduction

20% reduction

25% reduction

Effect on TM Reuse: example

The Bottom Line Part 2:

Quality doesn’t cost – it pays!

• Translatability/Readability

• High Quality content costs up to 30% less to translate

• Quality content is easier for non-native speakers to read

• Call deferrals

• Consistent terminology and style better indexing better search better self-help fewer calls

• Product liability risk management

• Certifiable quality manageable risk

• Internal process efficiencies and cost savings

• Less rework

• Less copy editing

• Cleaner handoffs between process tasks

• Less “friction”: fewer Writer/Editor/SME “debates”

Getting Started…General Quality Management Principles

• Formal, documented Standards and processes

• Metrics Driven

• Real-time feedback to process participants (QA)

• Test results/reports at process gates (QC)

• Quality built-in from the beginning

• Focus on continuous quality improvement

Definition…

Technical Information Quality

• Key criteria:

• Compliant with relevant standards and guides

• Correct (grammar and spelling)

• Consistent (style, terminology)

• Concise

• Result:

• Easier to read and understand

• Easier , quicker, and less costly to translate

Pop Quiz!

• Do you have Information Quality Standards?

• Do you monitor, measure, and track conformance to these Standards?

• Do you provide meaningful, actionable, real-time quality feedback to

Writers and Editors?

• Do you collect Quality metrics consistently and objectively?

• Are Information Quality metrics collected on every information product

you deliver?

• Can you conclusively demonstrate Quality improvements?

• Can you tie time and cost reductions directly to Information Quality

improvements?

• Is your CFO happy with current translation/localization costs?

Where to Start?

product data: 0.1 - 0.2

documentation design: 0.5

authoring: 1.0

editing: 2.0

acceptance: 5.0

translation: 10.0

maintenance: 20.0 ©J S

chütz

& R

Nübel, I

AI

Moving

quality

management

upstream

reduces

costs

The unit cost of fixing term errors

at different points of the Content Supply ChainSavin

gs

acrolinx - the company

• World leader in Information Quality Assurance (we invented it )

• Spin-off from German Research Centre for Artificial Intelligence (DFKI)

• Technology under development since 1997 (since 2002 as acrolinx)

• Users in 25 countries, checking around millions words a month

Some of our customers

Software Life Sciences Communications Industrial Technology

Adobe Dräger AlcatelLucent DAF Bosch

Autodesk GE Cisco HOMAG Embraer

CA Medtronic Huawei John Deere KonicaMinolta

EMC Siemens Motorola MAN Philips

IBM SonyEricsson SEW Eurodrive

SAS Institute Siemens

SymantecLeica

GeoSystems

The acrolinx Information Quality Suite

• acrocheck toolbar

• Color-coded flags

(like spell check on

steroids!)

• Auto-fix

• Suggestions and

examples

• New term harvesting and

control

acrocheck Writer/Editor Interface on-demand checking and guidance

Style

Terminology

New term

Spelling

Grammar

andelay

acrocheck Information Quality Status

Report - I

Status Results

Spelling 2 Results

Grammar 3 Results

Style 18 Results

Terminology 11 Results

New Terms

User Name kent

License ID Acrolinx Internal / kent

File Name

file://C:/Documents and Settings/Kent Taylor/My

Documents/Acrolinx/sampleDocs/Word/Topspin 360

Quick Start-Reuse.doc

Rule Set Demo

File format application/word

Document language English

Checking scope 618 word(s) / 77 sentence(s)

Date and time of check Sat Sep 22 07:16:48 MDT 2007

Version information Plug-in : 3.1.0 (build : 384) / Server : 3.5.0 (build : 1622)

Resources standard : 3.5.0 (build: 139 / 2007-09-10)

Term banks used Switches

Checking status (score: 138)

Administrative Information ▼show | hide

Settings and Result Overview ▼show | hide

acrocheck Content Quality Status

Report - II

Word Count Suggestion Context

managment 1 managementOne managment workstation, such as a PC running

terminal-emulation software

chasis 1

chais

chases

chasms

chassis

phasis

Phasis

The chasis cable kit (included)

a_an_distinction 1

avoid_duplicates 1

wrong_sequence_of_words 1

Install a additional switch blade: → Install an additional switch blade:

Tighten the screws to the on the left and right-

hand sides of the IB switch card.→

Tighten the screws on the left and right-hand sides of

the IB switch card.

Tighten the screws to the left and right-hand sides of

the IB switch card.

Result - Grammar ▼show | hide

Summary

avoid_duplicates

Result - Spelling ▼show | hide

Details

a_an_distinction

acrocheck Project Management Report metrics and reporting I

• Summary

metrics for

entire project

• Customer

definable

scores,

weightings and

thresholds

• Overall scores

can be used to

optimize

workflow

acrocheck Project Management Reportmetrics and reporting II

• Summary report

for each metric

category

• Color-coded

status indicators

• XML or Database

storage

• Active links to

individual reports

and document

files

Content Quality Management

A Real-World Example

Excerpt from a Best Practices Presentation, 2006

• Savings of as much as $1M in a year so far

• Projected savings of $3M per year for the next 5 years

• Translation costs for MT projects cut by 50%

• Time to market for MT projects cut by 50%

• 40% reuse of content

• 75% reuse for localization

Bottom Line

Information Quality improvement doesn’t cost …

It PAYS!

The role of quality management

• By defining and deploying (a.k.a. enforcing) standards:

• All information in the product works together

(coherent product information)

• The customer knows what you do

(consistent branding)

• Costs of going to (new) markets is manageable

(reduced localization costs)

Return on Investment

• Reduced product liability risk

‣ Certifiable and documented quality

• Reduced support and maintenance costs

‣ Coherent after-sales information

• Reduced costs of entering new markets

‣ Translation costs reduced by 20-30%

Questions?

More Information:

www.acrolinx.com

US Contact: Kent Taylor

970.921.3700 | kent@acrolinx.com

European Contact: Oliver Collmann

+49 (30) 288 84 83 41 | oliver@acrolinx.com

Intelligent Reuse (consistent wording)

• What’s wrong with “fuzzy” reuse?

• ... Fuzzy matching is great for translation!

• Fuzzy matching for authors doesn’t work because writing is not translation:

• Translation memory-based solutions:

• Cannot present focussed suggestions to users

• Does nothing for quality

• Does not reduce redundancy

Reusing TMs with fuzzy matching

• Writer types:

• WARNING: Switch power off only when the fan has stopped

• a fuzzy matching will offer:

• WARNING: Switch power on only when the fan has stopped

• WARNING: Switch power off before the fan has stopped

• but will not find:

• WARNING: Disconnect power only when the fan has stopped

• WARNING: Never switch the power off until the fan has stopped

• WARNING: Do not power down until the fan has stopped

• WARNING: You must wait until the fan has completely stopped before

switching off the power

Writer support?

acrolinx IQ Intelligent Reuse™

• Reuse for writers needs to:

• Maintain or improve quality

✓All suggestions are validated (and reliable for non-native speakers)

• Bring no significant overhead:

✓Typically there is only one suggestion

• Really reduce redundancy:

✓Linguistic search algorithms offer high precision

• No more overhead than spell checking

Some examples

Overview

• Who we are? acrolinx the company

• What we do? Information Quality Management

• The problem

• The solution

• The acrolinx Advantage

• Return on Investment

• Demos

The future of documentation

• The notion of “technical documentation” will start to disappear in the next 5-10

years

• ... to be replaced by “(product) information development”

• The notion of “source language” and “target language” will continue to erode

• Information development will increasingly become part of product

development

The role of documentation

• Merging of “in-product information” and “documentation”

• Konica-Minolta: product calls the service engineer before it needs

maintenance

• John Deere: combine harvester with 15 servers, internet access, GPS

controlled

• Siemens Medical: Synchronized console and maintenance information

• ... all products are becoming like software (actually they were all along)

Terminology Workflow

Best practice writing standards

Spelling Grammar “Style”

Server-based for lower

maintenance of custom dictionaries

Custom-built component for

technical writers.Off the shelf writing standards

Integrated with terminology

managementSupport for non-native authors

TÜV Certified

ASD STE Compliant

Specifically designed for technical

information (only shows genuine

spelling errors)

Can be easily adapted for specific

applications (UI messages)

Different standards can be set for

different types of information

Highly customizableIntegrated with other checking

components

Easily and quickly customizable to

meet customer needs

Reuse Workflow

acrolinx IQ Server

Terminology

Intelligent

Reuse

Grammar

&

Spelling

Writing

Standards

Intelligent Reuse™Building your Reuse Repository

Reuse

Repository

Clusters

micro-clustering

redundancy and quality

filters

review and release

Content / Translation

repository

the cat sat on the mat

The dog sat on the rug

The elk sat on the moss

The moose sat on the elk

the cat sat on the carpet

The cat slept on the sofa

Fish swam in the blue water

The fish swam in the green water

The fish swam in the red sea.

the cat sat on the mat

this is a sentence you can’t read

the cat sat on the mat

Another small test snippet

the cat sat on the mat

This is the same as the other one.

the cat sat on the matthe cat sat on the malt

The cat ate on the mat

the cat sat on the doormat

the cat sat on the mat.

The cat sat on the mat

the cat sat on the mat

the cat sat on the mat

More useless data points

Raising your corporate IQ

• best-practice writing standards

• spelling & grammar

• terminology

• intelligent reuse (write once, use many times)

• coming soon...

• Smart indexing assistant (the secret problem)

• “Markup Checker” (for example: UI strings)

• Link checker

• Translation QA