Transcript for Keynote Presentation: The On-going Evolution of the Localization Business

Good morning, I hope to wake you up

I hope you did not have too much traffic coming in

The topic on my presentation today is evolution in the translation business

I love the word evolution because it indicates movement, progress

And I think that in the translation and localization industry we are living in very interesting times.

I would like to start with a quote from

two authors who talk about futurism and forecasting.

And one of the great things about this quote is that

when you are talking vision, about the future

it is very easy to make mistakes

If your vision is good and you have

a good picture of the future

You probably are just describing the present

so a lot of the things that I said and I say when I talk about the evolution of the industry

I have been very glad to see them happening in the last few weeks

This is how fast things are changing

So if I go and say things that seem to be fantastic and impossible

I am probably wrong but thats ok

I can live with that

But lets talk a little about evolution and innovation

I am fascinated by the topic of innovation, how things change

in the world, not only in the translation industry

I have been doing a lot of reading and research about this topic

One of the things that I found out is that

in markets that are very competitive

Yesterday Mike Anobile presented some data from the industry

some of the data from the industry that I have seen suggests that are probably 5,000 LSPs

translation companies around the world

so its a very competitive, very dynamic market

a market with low barriers to entry, anybody can start a translation company quite easily

so in markets that are very competitive like this

people don't like to, uhh, embrace change very easily

they love the way things are and are very comfortable with the status quo

and they feel safe in the way they do business and the things that they do

so what has happened in the translation industry essentially

for the last 25 years is very little change

Things are done the same way over and over again

I am a frequent participant to industry conferences

and its fascinating how often I hear the same things over and over again

and I must say that you in this audience are very lucky

because I probably attended some 20 conferences

in the last 12 months and

this is the first time that I have seen something really different

something new, some good interesting technology conversations going on

and breakthrough happening

you are among the lucky few who are seeing something really different

because most of the time when you go to industry conferences all you really hear about is quality

uhh quality, uhh quality

and sometimes standards

and then if you are in Europe you will hear about the EN15038 standard

which is a fantastic standard

that tells you how do translations the way they were done in 1950

and then you go to Canada and they are having the same conversations

talking about establishing standards the same way that they do in Europe

so they are going to copy the European standard

add some more bureacracy

and then you are going to have translations the way you did in in 1949!

and we heard yesterday from our colleagues from the TAC (I don't see them here) in China

and China wants to imitate exactly the same thing

yesterday he told us that they have published 3 translation standards

standards for/in services, I am not talking about technology

in software standards are very good, they allow for interoperability, they allow communication

they allow people to build on common platforms

but in the services industry standards are a waste of effort

it's trying to make people, everybody do the things the same way

and this is not very helpful for our industry because

it freezes the way things are done

and everybody feels safe because everybody knows what is going to happen

everybody knows how things are going to be done

but we suffer from several problems because of that and I am going to share

with you what are some of these issues

One of the things I have identified, and yesterday I confirmed it

Age and bad habits really affect the translation industry

and when I talk about age, I think that people who leading and driving the conversation in the translation industry

are old and there I include myself

How many of you here were born after 1984?

1, 2, 3, 4 ... right ... OK you can lie, ladies can lie

not too many, so 1984, 25 years old

These people who were born after 1984

were born with the internet, they didn't use typewriters

they did not use Wordperfect

They didn't use, they don't know what Ventura Publisher is

They don't know what a 3½ floppy disk is

Why? Because they are the new generation

The leadership in our industry

thinks; these are common statements I hear every time

I talk about innovation in translation: It's always been done like that

It works! I don't have time to change

my processes because I am too busy

I'm losing business, I'm doing stupid things

making a lot of mistakes but I don't have time to correct them because I am too busy

Oh I heard this last week from a large software company in Silicon Valley

I invested $250,000 in SDL TMS

I can't change my processes now

OK, you're throwing away money, but that's your problem

The other part is skepticism

machine translation will take so long to catch up that it is never going to work

This sentence about MT, I heard it for the first time 25 years ago

MT is between 4 and 400 years from perfection

Well until 5 years ago, 3 years ago, I would have said we are closer to 400 years

Now I would say that we are closer to 4

and actually I would say we are closer to 2 than to 4

People are too busy being busy, and they are not really thinking about the changes going around them

The way that translation has been done forever

and some of you might have heard of the Septuagint

which the way the Bible was translated into Latin

it was translated by 70 translators

and the Pope decided, where there was agreement, that would be the standard

and was the first model of crowdsourcing I guess

but anyway the way we have been doing translation is the same way we have been doing it for hundreds of years

It is following the publishing, the printing model

somebody does the translation, somebody reviews the translation, another set of eyes reviews the translation

and what is the process?

the process is all about catching errors

I am going to read this material because there must be a mistake here to find

How many of you have ever reviewed a translation?

Both of you, right?

How many of you have reviewed a translation and sent it back with no correction?

I don't believe you ..(laughs).. it was your translation

very small ..... one paragraph...

anyway human nature is, especially in translation, which is something that is subjective

forces us to want to change things

They say that the biggest drive in human nature is not love, not hate

it is the uncontrollable urge to change somebody else's text

right? if you have a red pen in your hand you will change text

its human nature, fortunately we don't do anything with the red pens anymore

so you can let some things go

The problem with catching errors is that

there is an imbalance, sometimes the reviewer has not the same set of information

or less information than the translator

so the reviewer is actually adding mistakes to the translation

The reviewer goes there and says I don't like six, lets put half a dozen

well same thing, BUT the glossary says six not half a dozen

and then you suddenly incorporate a mistake into the system

The other problem is that we in the translation industry follow

a sequential process, so somebody will translate, somebody else is going to review

somebody else is going to proof it

and very often there is no communication between the parties

You will say OK small project, 10 pages, 50 pages, people will communicate

but when you are talking in a project where we have 50 translators

and 12 reviewers

there is very little communication between the translator and the reviewer

so teamwork is not a natural part of the process

and I call this the blind leading the ignorant

nobody knows what they are doing and everybody prays that somebody else is going to catch

the mistake, if you are not sure of the translation, the reviewer will find it

the reviewer will catch my error

so what we need to do now is move to a new paradigm

the technology as Dirk showed us yesterday in his presentation

allows for collaboration, allows for input from multiple parties into the process

so we are in a period in technology and development where multiple people can work together

and they can do it right the first time

we heard in several presentations yesterday that when multiple people work on a project

mistakes are caught naturally, because of re-use

because of cross referencing and because of the editing process

that goes on as the translation is being done

So what has prevented the implementation of innovation in the translation industry?

There are three dogmas - a dogma is a Catholic church term - it is something that you don't discuss

you don't disagree, it doesn't 'matter

so in the Church they will say: God Exists

nobody can challenge that, it's a dogma, right?

And we have dogmas, we have these topics that nobody can discuss

in the translation industry because everybody agrees on them

and they have been the main reason that innovation has been hard to happen in the translation industry

the #1 dogma has to do with translation memory

What I liked about our conversations yesterday is that nobody talked about translation memory

every once in a while there was a question when Phillipp was making his presentation

we did not touch the subject and the whole industry has heard the story

invented by Trados by the way, a very smart marketing ploy

that translation memories are assets

That they have economic value

I contend that translation memory has zero value

For me a translation memory is the same thing as a dictionary

a dictionary - the value of the dictionary is the $50 that you paid for the book

or the access to the website

the content is free

what costs money is the package, right?

translation memories are not an asset from the economic point of view

if they were an asset they would be depreciated and they would be amortized

Do you do that at Adobe, Dirk?

You don't think so, right? you don't know how much

in your books you have - I know, its a dogma, I see the smirks

I am challenging a dogma

My contention, go ahead Michael ......

translation memory will generate savings

it will be an investment but they are not an asset

It's like a coupon - when you go the pharmacy with a coupon and you get

a 25% discount on your drugs or your shaving cream

cause you have a coupon - that's what translation memory is

a coupon - its an opportunity for you to save money in the future

You may save it or you may not

It might be a good deal, it might not

I go to a pharmacy in Boston where I live and I

every time I buy something I get coupons

and I always get coupons for discounts for diapers

cause I used to buy diapers

four years ago when my son was a baby

I don't buy them anymore, he is four years old, he doesn't

wear diapers, but I still get 25% discount on diapers, thats like translation memory

its the same thing. Sometimes it has absolutely no value, OK?

The 2nd dogma that has prevented change and innovation in the industry is

the concept that more eyes improve quality

The more you review, the better the quality

and that seems .. natural .. makes sense right?

so I write a text and I ask two people to read it

Yes if the people that are reading it and reviewing are qualified

to add value

The reality is that it has become a process

People only get reviews because you are supposed to have reviews

Its part of the workflow

Somebody translates, somebody reviews

check, check, check

There's no real value added in this process

There is just a check mark done there because we assume

that other people reviewing will improve the process

If I did a good job selecting the right translator

I don't need a reviewer

A reviewer will only add noise

add impurities to my translation process

So I don't say the process is bad

But the process for the sake of the process is not valuable

So it is a dogma

We don't need three reviews to make a better translation

As I said before

The 3rd dogma has to do with consistency, right?

I had asked this question countless times

What is better? To have 2 translators to do the job in 10 days?

Or 10 translators to do a job in 2 days

Everybody will say it is better to have 2 translators to do the job in 10 days

My answer would actually be, If you have 10 days

I use 10 translators and do the job in four days

I use two days for preparation

I'll use two days for cleaning up and organizing the project

The reality is that, .. in my experience in the translation industry

as a translator, as a business owner,

as a project manager

as an analyst

What I have seen is when you talk about quality

you are usually talking about 15 variables, depending on the language

90% of the errors that you find in translations

have to do with style

and style has to do with capitalization

with spelling, with choice of verb tense

form of address

informal / formal right?

language style

These variables are things that you don't standardize after the fact

Style is something that you standardize upfront

Style is something that you agree

in the beginning. It is much easier to get 10

different translators on a phone call and agree with them

we are going to translate the gerunds like this

all the numbers will be converted to the metric system

with 2 decimal points

We are going to use capital letters in all the words in the titles or not

Whatever you agree, we are going to capitalize after a colon

or anything, you can agree on those issues upfront

90% of the mistakes that exist in the translation processes can be avoided

by preparation and agreeing with the things upfront

By the way, I have a blog and I invite you to read it after the event naturally

My last entry is about the latest book by Dan Brown (of the DaVinci Code)

He wrote a book that was launched in the US on September 15th

and was launched in Sweden on October 21st

36 days after ... it is a 502 page book

36 days after it was launched in the US, it was launched in Sweden in Swedish

7 translators worked on the project

and if you talk to anybody, if you want to translate literature

You have to have a translator that understands the style of the author

and you have to have consistency in the tranlsation

What these people did is, they said that Harry Potter

lost a 150,000 copies in sales

in Sweden because it took them six months to translate the book

So the Swedes were buying the book in English

instead of buying it in Swedish

so with the Dan Brown book which was supposed to have been a big best-seller

they said we are not going to lose money

we are going to launch it as fast as possible

and they used this approach, they used more translators

to work on the project, they coordinated

and everything else, but

the driver here was not the quality of the translation and by the way

if you have seen my blog, I refer people to a story

that was published in a Swedish publication

and the Swedish author says

unfortunately, or sadly is the word she uses, "sadly the translation is quite good"

and she said sadly because she would rather have

bad translation that she could criticize and say that the process did not work

But again this for me was a clear proof that

you can have good quality, consistent quality

consistent translations, regardless of the number of translators that are working on the project

If you do a good setup, a good preparation phase

So why is it important to talk about

innovation and change

and changing the processes and defying the dogmas

Because we are in an environment

where automation has to be

a requirement

We cannot continue to doing translation the way we have been doing

and expect that everything that needs to be translated will get done

into all the languages that need to be translated

and the reasons are very clear

the number of translators is limited

it takes a long time to develop a good translator

it would take you 3-5 years to have decent translator working

while I could double the amount of content in a year

or less actually, somebody mentioned yesterday quadrupling the volume

I forget the context but

I could essentially say that tomorrow I am going to localize all the Adobe products into Burmese

Are there enough translators to do Burmese?

Are there enough people who can do Burmese?

And all of a sudden I have a concept that has never been translated before

or that never existed that needs to be translated

so formation takes time, translators are scarce

and the demographics run against us

Most rich European economies have declining populations

They say that even India and China, in the next 40 years are going to have

stable populations

The global population is expected to reach a standard stable level in 40 years

So the population is aging, people don't make babies anymore in Scandinavia and Italy

The Netherlands, I think that France is the only EU country to have positive growth

Probably because of the Algerians, the Moroccans and the other ones who are there

There is another limitation that we have

You cannot buy Norwegian in Uganda, as they say

Maybe there is one translator, a Norwegian who lives in Uganda and can do translations but

The market is driven by the country that speaks that language

Norway is a very rich country

Nobody wants to be a translator in Norway

Actually nobody wants to work in Norway

But that's another story

They're recording this right? they are going to show this in Norway

I think I am going to be banned from Norway

So this is the natural environment - you can see from this curve here

You have an explosive growth in volume

You have price that is pretty much standard or slowly declining

for the last twenty years

the number of translators is going down

So you need to improve productivity, that is the only way that you are going to address the demand for content

So we have the perfect environment for automation

Add to to that the fact that I think that translation is like toilet paper

Its very cheap and it's only important when it is not there

Nobody thinks about toilet paper until it's not there

Nobody thinks about translation until it's absent

You never see a positive translation story in the news

The only time that translation will make the news is when something goes wrong

That is the only time that translation will go to the CEO of the company

The last time I saw news about translation related to a large company it had to do

with Google in China, where they had failed in their translation efforts

and it went all the way up to the founders and the CEO and they had to go all the way to China

and address issues that had to do with Localization

The conversation I heard here yesterday and I hear over and over again is

How are we going to make translation a strategic issue?

How are we going to make strategic discussion.

We are going to make it a board level issue the day they start discussing toilet paper

So what happens in an environment like this is that everybody is comfortable

Like I said before, this is the way have done business, until disruption happens

And disruption always comes from the outside

Disruption comes from the guys who are not thinking the same way that we are thinking

They did not go to translation school, they did not work at an LSP

They never used software in their lives

They don't know what Trados is

They probably call it tray-does or something like that

Here are some examples of disruptors who are coming from the outside

Facebook -good example - they had their website translated professionally

into Spanish in about 3 months

Then they created a tool that let people translate and they had their website translated into French

in 48 hours

We are talking about 300,000 words or something like that

48 hours and the volunteers translated Facebook

I participated just to test the tool

to the Portuguese translation of Facebook

I did 47 sentences, 430 words

And I had a 198 votes for my translations

And with this little amount I am today

35th most frequent translator of Facebook

But if I go up I will have guys who have translated tens of thousands of words

What did I get in return? My picture on Facebook

You can go there and see me on the Leaderboard

Another player coming completely from the outside, our host here, Asia Online

This guy never did translation before, he can't even speak another language

He is a monolingual translator as we heard yesterday

So the thing is that somebody who looks at the translation problem as a technology problem

rather than a language problem. He is not interested in subjects and ..

verbs and inverse order, direct order

as a style issue and a linguistic issue

He is interested in this as a technology problem, as a database problem

How do I solve this from a technology point of view

And of course Google, the big disruptor

The company that goes in and wherever they go they change the rules

And they came out with this Goggle Translator Toolkit which is very basic

It is not enterprise level, it is not something that can be used for business

It is not at the same level as Asia Online

but its free, its there, anybody can use it

its there today and these are people that are not in the language industry at all

They don't care about the language industry

The reason why this is important is that Google has unlimited funds

and this is a pet project for the founders of the company

who just happen to be Russian and French, right?

and so they understand the language issue

The program manager of GTT has weekly meetings with the CEO and the founders of Google

Not everybody has that, this is a key project for them

So just as we don't use typewriters anymore, we won't be using Trados in the future

Yes Trados is dead

This is a picture from Don Shin a vendor in LA showed at a presentation, that I borrowed from him

Yesterday I saw a demo of Asia Online, it is very close to what we have here

So it is essentially a desktop for a translator

You have WYSIWYG, you have glossary, you local memory, server memory

You have machine memory and chat with team members

so you can collaborate and finally this is a cute thing a calculator

How much money am I making as I work?

Isn't that cool? Can you imagine that?

I made $5 and I did not even work 3 minutes

Just as we did not use Google 5 years ago, today my daughter asks a question I don't know - she says Google it

It is going to be the same thing with Google translate

The discussions that we are having these days are moot

We are having these discussions because we are in a transition period

In a few years maybe months, all of this will be old history

and new generations will come with different ideas and different expectations

So to conclude I make a few predictions here

I am being optimistic (or pessimistic) By 2015 TM tools will be free or irrelevant

I don't think we will be talking about TM anymore after the next 5 years

Most large translation projects will be collaborative in nature

as they are today but more, technology makes it so easy

This I find funny, I wrote this two months ago, translator productivity will be measured in tens of thousands of words/day

instead of 2,500 or 3000 words that we expect today

The funny thing is that there was an email sent by SDL last week

where Marian Greenfield of ATA said that she did 34,500 words in 10 hours

Just by using the features of her Trados whatever

Like I said in the beginning, its going to happen in the future, but its already happening today

I see a change in the landscape in this industry

Companies that get it will grow like Asia Online, Google, Lionbridge, Lingotek, Sajan and Elanex

I think that the biggest loser is going to be SDL cause they resist innovation, they resist change

It is so easy to stay the way they are. They want Trados to be faster

Trados to be better, Trados to be the standard

Or SDLX or whatever product they have

They only sow confusion - they have 19 products, some of which don't talk to each other

and they have VERY BAD customer service

You can't compete with free Goggle Translate

or you cant compete with excellent service and excellent quality

and evolved process like Asia Online when all you do is provide client-server technology

So this is my presentation for today and this is my contact information

And I see a sign the time is up