The cost of developing content (part 4)
In previous articles, I’ve talked about the cost of developing content in Word, FrameMaker or InDesign, and in topic-based authoring tools. The numbers show it’s cheaper to create content in a topic-based authoring tool because of the typical ease of content reuse in this paradigm. But what about localization?
Localization
Localization is the process of translating content to be appropriate in another culture. This includes translating from one language to another but also includes things like changing measurements and examples to be culturally appropriate.
Not everyone localizes their content. When I ask content developers, about half say they are localizing now and about half expect to be doing it in the near future. A small percentage say they have no need, now or in the future.
Not everyone needs to localize their content – I had a client who did training documentation for a US organization and the language of the workplace was English.
The cost of translation
Localization is expensive. It requires people who are fluent in both the language the content was written in and in the target language. There are tools that help but humans are involved in the translations. Currently, machine translations can only get you so far.
The languages you need to translate to also influence the cost of translation. The fewer speakers of that language, the more expensive. But a good starting point for costs is to estimate $ 0.25 US a word (including some to a lot of page layout and/or graphics recreated). Some languages will be more, some less, but it’s a good place to start.
So, what do the different sort of tools cost us? The following tables show you.
Remember, though, because we are comparing apples to apples, the total cost of the more efficient tools looks like they cost more to translate. In reality, because you have a lower per page creation cost, you can get more content created for the same price. You have more content to translate, which means it costs more overall, although at least 30% of that content will only need to be translated one time, because you smartly reused content.
That’s why I show you a cost-per-page-per-language as well. That cost drops dramatically per tool.
Word
| Year | Number of Writers | Fully burdened hourly cost per writer ($50 an hour) | Total pages (new and updated) | Total cost per page to create | Total cost per year to develop | Localization cost per year ($.25 per word for one language) |
|---|---|---|---|---|---|---|
| 1 | 10 | 500 | 55,738 | $ 154 | $8,575,000 | $ 2,786,875 |
| 2 | 10 | 500 | 55,738 | $ 154 | $8,575,000 | $ 2,786,875 |
| 3 | 10 | 500 | 55,738 | $ 154 | $8,575,000 | $ 2,786,875 |
| 4 | 10 | 500 | 55,738 | $ 154 | $8,575,000 | $ 2,786,875 |
| 5 | 10 | 500 | 55,738 | $ 154 | $8,575,000 | $ 2,786,875 |
| 200 words per page = | 55,737,500 | Total= | $42,875,00 | $13,934,375 | ||
| Average page cost (200 words per page * .25)= | $ 50 | |||||
FrameMaker or InDesign
| Year | Number of Writers | Fully burdened hourly cost per writer ($50 an hour) | Total pages (new and updated) | Total cost per page to create | Total cost per year to develop | Localization cost per year ($.25 per word for one language) |
|---|---|---|---|---|---|---|
| 1 | 10 | 500 | 69,417 | $ 124 | $8,575,000 | $ 2,776,667 |
| 2 | 10 | 500 | 69,417 | $ 124 | $8,575,000 | $ 2,776,667 |
| 3 | 10 | 500 | 69,417 | $ 124 | $8,575,000 | $ 2,776,667 |
| 4 | 10 | 500 | 69,417 | $ 124 | $8,575,000 | $ 2,776,667 |
| 5 | 10 | 500 | 69,417 | $ 124 | $8,575,000 | $ 2,776,667 |
| 200 words per page = | 69,416,667 | Total= | $42,875,00 | $13,883,333 | ||
| Average page cost (200 words per page * .25) with 20% reuse= | $40 | Localization costs drop by 20% per page from Word | ||||
Topic-based Authoring
| Year | Number of Writers | Fully burdened hourly cost per writer ($50 an hour) | Total pages (new and updated) | Total cost per page to create | Total cost per year to develop | Localization cost per year ($.25 per word for one language) |
|---|---|---|---|---|---|---|
| 1 | 10 | 500 | 111,475 | $ 77 | $8,575,000 | $ 3,901,625 |
| 2 | 10 | 500 | 111,475 | $ 77 | $8,575,000 | $ 3,901,625 |
| 3 | 10 | 500 | 111,475 | $ 77 | $8,575,000 | $ 3,901,625 |
| 4 | 10 | 500 | 111,475 | $ 77 | $8,575,000 | $ 3,901,625 |
| 5 | 10 | 500 | 111,475 | $ 77 | $8,575,000 | $ 3,901,625 |
| 200 words per page = | 111,475,000 | Total= | $42,875,00 | $19,508,125 | ||
| Average page cost (200 words per page * .25) with 30% reuse= | $ 35 | Localization costs drop by 50% per page from Word and 38% from FrameMaker or InDesign | ||||
Cost per page comparison
And for you visual learners out there, a picture, showing the cost reduction at 200 words per page in each tool.
So why the reduction? Because of content reuse. FrameMaker lets you reuse content to some extent (generally I see clients reusing perhaps as much as 20% over Word and that’s the calculation I’m using here). The real bang for the buck happens when you move to topic-based authoring and leverage content reuse in a systematic way. My clients see at least 30% content reuse (many see much more), so I’m using that number here.
But we’re not localizing
If you’re not localizing now, a best practice is to act like you will. In my experience, localization happens because a market opened and the content must be translated to the market language. Typically, a large contract depends on this. And the deadline is, well, now.
Now is too late to get the content ready. So you spend more money than you should and everyone gets upset at how much this cost.
You have to write as though you are going to translate, even if you’re not. Files that are not consistent in content, terms, styles, and so on will cost you twice as much as I’m estimating here because the translators have to solve those problems as they translate.
As a happy side effect, even if you don’t translate for several years (or ever!), files that are ready are also generally easier for your customers to use. They have consistent language, terminology, structure, and are just overall of better quality. You don’t lose anything by making sure your content is ready, even if you never translate.
The widget that lets you put in your own numbers is almost ready. Stay tuned!
4 Comments
Trackbacks/Pingbacks
- The cost of developing content (Part 3) | Sharon Burton - […] Part 4 […]
- Стоимость разработки контента (часть 4) | Разработка технической документации - […] Источник: The cost of developing content (part 4) […]
- The cost of developing your content (Part 2) | Sharon Burton - […] Part 4 […]
- The cost of developing your content (Part 1) | Sharon Burton - […] Part 4 […]
Great information, Sharon. I would, however, separate Frame from InDesign because Frame content is much easier to localize than InDesign, particularly when it comes to DTP.
The reason for this is that Frame content typically can be extracted into the translation memory software, and then pulled back out relatively easily. With InDesign, text, graphics, and other elements are frequently in separate flows, which requires manual extraction pre-translation and re-positioning post-translation. This can add significantly to the DTP costs, which are about 50% of the localization costs per language in traditional publishing.
With structured authoring (e.g., topic-based content, XML, CMS), much of the DTP process can be automated, reducing the localization DTP costs by 80% or more. (You still have to do a QA edit, but it’s more like a blueline edit from the old days– just making sure that pages output correctly.)
Kit, you’re right. I always like to assume a perfect world where people used InDesign well and not stupidly. Sadly…
When it comes to localization, you’re the expert. If people want to use their tools stupid, that’s going to cost in so many ways.
Thanks for the good analysis, Sharon. Do you think that Frame/InDesign and topic-based authoring might yield cost savings in other ways besides reuse? Specifically, many of these tools can generate standard formats, like XLIFF, that could expedite the process. (I’m not talking about machine translation, but about methodologies that streamline the workflow for human translators.) However, I’m not an expert so I can’t prove this. What do you think?
By and large, Larry, all the major tools can be imported by all the localization tools. For example, Frame can be imported into SDL’s tools and so on.
Some tools have their own localization tools, like Madcap Flare. Author-it has a manager that helps (it exports out to a format the localization tools can use).
However, by sending topics out when they are done instead of waiting for the entire chapter or document, you can also see a cost savings. Killing localizers costs more.