“When it comes to localization, what kind of consumer are you? Do you look at the translation rate alone when choosing a vendor? Or, do you look at the long-term costs associated with future updates and maintenance of your products and literature?“
In an effort to help save the environment and reduce energy costs, we purchased last year a new washer and dryer that were guaranteed to save 70% on energy costs. The washer uses 80% less water, hence saving 80% of the energy used to heat it. It also does a much better job wringing water out of clothes during the spin cycle, benefiting the dryer because the clothes are less wet.
The dryer smartly recycles the heat from the generated vapor, and shuts off automatically when the clothes are dry. These features significantly reduce the amount of energy it needs.
These smart appliances were twice as costly to purchase. But the price was well justified when we factored in long-term savings in energy costs and lower CO2 emissions. We are on track to get payback on our higher upfront investment by next year, leaving us with significant savings on energy costs for many more years to come.
With much higher energy costs today, many people are reconsidering purchasing the lower-cost but inefficient appliances. They are no longer simply comparing the sticker prices before they make the purchasing decision. Buyers are becoming more sophisticated, looking at second- and third-order parameters to calculate the real cost of ownership.
When it comes to localization, what kind of consumer are you? Do you look at the translation rate alone when choosing a vendor? Or, do you look at the long-term costs associated with future updates and maintenance of your products and literature?
With hundreds of translation vendors competing in a fragmented market, many capitalize on the limited knowledge or time that most consumers have to select a vendor. They adopt lower upfront costs per word to lure new clients, hiding the long-term maintenance costs and rates. Later, they apply unfavorable rates on leveraged translations during follow-up projects and updates.
In the short term, this may appear to be a benefit to the client and a solution to their often limited budgets. But in the long term, it is a much costlier scenario. Here are the details about what is involved and how to best calculate future update costs.
Most content publishers, software or other, update their product, online, and marketing literature at least yearly. They update the source files, and engage their localization vendor or translation staff to update all the supported languages.
Changes to the source files take the following shapes:
- Additions (new text)
- Edits (modifications to existing text)
- Deletions (removal of obsolete text)
The new text requires new translations for each target language. The edited or modified text requires updating in all target languages. The deleted text is disregarded.
When a top-down translation process is applied in the initial translation of the source text, a translation database (translation memory TM) is created. The translation database stores language pair segments or sentences for each of the target languages.
When an update to the source is available, an analysis is run on the newly-released source text one segment at a time to compare it against what is already in the translation database. The following is the result:
- Repeat or unchanged text is typically 70-80% of your initial content. These are 100%-match sentences (exact matches) found in the translation database and do not require any changes. Translators only have to verify that the translated segments in the database can be used before including them into the new files. This is a minor effort compared to the initial translation effort, and therefore under 10% of the initial cost.
- Edited or modified text results in a “fuzzy” match. This is a match in the database that is similar to an existing translated sentence, but not an exact match. It can be anywhere from 50-99% of the original (below 50% is considered a no-match). Typically, around 10-15% of the total text falls in this category. Internal changes to the sentence formatting (bold, italic, links, internal font or color change, etc.) will often force a fuzzy match. When a fuzzy match is found, the translation process is easier and the translator is able to work faster. Hence, lower costs are justified (anywhere from 40-80% savings).
- No-match or new text is typically around 10-15% of the initial total content. These sentences or segments generate a 0-50% match to sentences in the database. They require full translation and full translation rates.
- Deleted text produces no impact on the translation update effort, since the text is no longer relevant.
For instance, if you change one word in a sentence containing ten words, a 90% fuzzy match will result. If you change five words in that same sentence, then the fuzzy match is 50%. The analysis and calculations are performed objectively and accurately by the translation database search engine software.
By measuring the fuzzy match of each sentence, we can easily approximate the translation effort needed to perform the full update in any target language. Applying an appropriate weight to each fuzzy match is a process that can be used to estimate not only the cost, but also staffing and scheduling needs. This is why we at GlobalVision can accomplish more than 94% of our projects on schedule and on budget – and why project updates often cost less than 40% of initial translation costs!
When looking for a new translation and localization company, just like when you are looking at buying a new appliance, do not simply compare advertised translation rates. Ask about fuzzy-matching rates and the costs applied on already-translated sentences. Any upfront savings due to lower per-word rates will quickly evaporate when you factor in the recurring charges imposed on text you already paid to have translated!