IS TRANSLATION MEMORY LOSING YOU MONEY?

 

This article isn't about whether TM works (which it does), but whether buying into it is a viable business proposition for the freelance translator.

 

In theory, TM is a great investment. It may cost twice as much as Microsoft Office and put the brakes on your productivity while you learn how to use it, but you'll then gain from increased consistency, no duplicated research time and an overall increase in productivity of perhaps 15-30%. Sounds great, doesn't it? But, as many translators have found out to their cost, higher productivity doesn't necessarily mean higher income.

 

Fuzzy matches and fuzzy pricing

 

Most TM packages come with a function that calculates the number of repetitions in a text and the number of 100% and fuzzy matches you obtain from your database/s. This makes it easier to calculate how long it will take to complete a translation. At the same time, translation memory has been marketed both by some TM companies and many translation companies not so much as a quality assurance and project management system, but as a productivity and cost-cutting tool. As a result, a whole new approach to pricing has become established, with many clients only paying a percentage of your agreed rate for repetitions, full and fuzzy matches (which may, of course, come from translations you yourself supplied in the past).

 

Leaving copyright issues aside, that might seem quite fair. After all, it's the TM system that "translates" the 100% translations and repetitions, just as it saves you research and maybe typing with its fuzzy matches, so why charge full price? The answer is because a 100% match or repetition may need to be translated very differently in different contexts, and it can easily take longer for an experienced translator to "fix up" a fuzzy match than to just type in a new translation.

 

In short, however scientific all these percentages look, they don't accurately reflect the productivity increases (if any) a translator will actually achieve. What's more, if the emerging context of a translation makes it necessary to change certain terms, you'll end up having to edit each translation unit individually irrespective of whether it's a repetition, and maybe need to make changes to the memory, too. Add the time you spend maintaining your databases, plus the fact that several systems have formatting and stability problems that can lose you hours, and things start to look even more depressing. With many translation companies paying nothing for 100% matches and 25% for repetitions, you can actually see your income fall using translation memory unless you increase your rates to suit.

 

How are translators responding?

 

Well, the most extreme response is the most obvious. Many translators are keeping their use of TM to themselves. As far as they're concerned, it's nobody else's business. It's something they invest in to increase consistency, make their research more efficient and generally help maintain a competitive edge. They don't see TM as primarily a productivity tool and consider a sliding scale of discounts as anathema to a communications professional. Above all, they refuse to enter into a situation where their work is being recycled to the economic benefit of their clients but not themselves. The only exceptions they make are for revisions of work they have already done in translation memory.

 

Others do offer discounts on jobs where they achieve a significant increase in productivity, but only after they have amortised the time invested in building up their databases.

 

A small minority of translators quite simply charge for their services by the hour. That way their clients pay more while the databases are being built up, but subsequently reap the quality and productivity benefits of the system. But there are differences of even here. Some will work this way for "direct clients" but not translation companies. They contend that the use of TM with its recycling of past work is acceptable in an in-house context, but not between a freelance and translation company.

 

Many freelances, on the other hand, are quite happy to offer discounts to translation companies, but limit them to no more than 50% for repetitions and 100% matches, and won't give any for fuzzy matches of less than around 75% if at all (although there's less consensus over the latter threshold and discounts).

 

Finally, a sizeable number of translators concede these discounts but are no longer happy with them. After all, why invest in a system and how to use it if you don't get any of the benefits? It looks as if change is in the air.

 

What about the translation companies?

 

A small but significant number of translation companies sell the advantages of TM to their clients without offering them discounts or expecting any from their translators. Their numbers are, if anything, increasing. Some translation companies - newcomers to translation memory in particular - are reducing the discounts offered. While others find themselves in a quandary, faced with a backlash against discounts by translators on the one hand, and pressure to continue offering them from their translation-memory-using clients on the other. As you can imagine, all jealously guard their databases, and to my knowledge not one has accepted the offers of clients to buy them.

 

So, once again, when you look at things closely, there isn't such a difference on either side of the "great divide". Translation memory offers us many important benefits. We shouldn't lose sight of them because initial infatuation with the systems led them to be marketed principally as a magic lamp for saving time and money - something which has led many end clients to have an even more limited understanding of what is actually involved in the translation process.

 

Tips on how to make TM increase your income as well as your productivity

 

* Negotiate: There may be more scope for a mutually attractive solution than at first appears.

 

* Communicate: Let "discount" clients know if the system and memory are generating unhelpful fuzzy matches and inappropriate 100% matches and/or repetitions. Offer an alternative solution.

 

* If you don't work for an hourly rate, consider offering a variable discount that depends on how each job actually proceeds. The final price will thus, for example, reflect appropriate formatting by the client that allows you both to benefit from TM (e.g. no carriage returns and tabs mid-sentence instead of automatic indentation). This will help improve their understanding of what is involved and, ultimately, your business relationship.

 

* Never buy translation memory on the assumption that it will help you get new clients. You'd be surprised how many translators have bought systems on this premise and never used them.

First published in tranfree, 2001.