It’s relatively old news now that the Translation and Localisation industry is making great strides, evolving into an industry full of promise. The industry’s progress, however, goes beyond its growing size and revenue and is characterized by continuous innovation. In such exciting times, Slator, founded in 2015, has been doing a remarkable job of helping LSPs stay on top of things with its concise language industry intelligence.
With us this month is Slator’s Co-Founder and Managing Director, Florian Faes, who is himself a language professional. With his thorough research on the language industry, he walks us through the language industry as it stands today and gives insights on what to expect from it in the near future.
Within a very short period of time, Slator has managed to not only create a huge name for itself within the industry but the reporting also carries a lot of weight. What would you attribute this success to?
I think it’s about relevance. We are very careful about the stories that we choose to cover and we place a lot of importance on our editorial independence. We review around 50 potential stories everyday for publication. Secondly, we make it about business which is something that perhaps resonated with the community. We cover “news you can use” which is actionable for the businesses that are competing in this highly competitive industry. We provide a lot of financial information for companies and see the sector as the thriving business sector that it is. And since there are around 10 thousand LSPs in this industry, there is definitely a demand for business news.
Based on the research that you have done and the insights that you have thereby got, what would be the 3 main things changing in the industry?
The first would definitely be the rise of Neural Machine Translations. Far from being a buzz word, I think it’s going to have a tremendous impact on the industry. In the languages that I am personally competent in, for example translation from English into German, there has been an amazing jump in quality over let’s say the past twelve months. So Neural MT is going to be productized and will become a part of the offering.
The second trend would be the continued consolidation among LSPs. There is a fair amount of M&A and investment going into this space that has implications for providers because they either have to buy or sell to someone.
Lastly, the aggressive market entry of the very big tech players really accentuated when Amazon announced the launch of Amazon Translate. Much more interestingly, Amazon targets their solution not only at big enterprise buyers but also specifically singled out LSPs as a client segment. This is unique so far because most of the bigger outlets have produced their own engines but they never really targeted LSPs. I think this is somewhat game-changing in the mid to long term.
Could you tell us a little about the effect that the rise in neural MT will have on buyers and sellers of language services?
LSPs which are slow to adapt and integrate neural MT in their supply chain will find it very hard to maintain profitability going forward. They may be able to retain pockets of business with clients who are themselves slow to adapt this new technology but will not be able to grow much. It’s very comparable to where CAT tools stood 15 to 20 years ago. Any LSP that wanted to grow simply had to adapt those tools. As for buyers, it’s a very simple equation. They will get more for less or even more for the same which is a very positive development.
Do you think translation prices will go down significantly owing to MT and will that mean a lower pay for human translators? Or do you believe they will earn the same but increasingly as post-editors?
Yes, per-word rates will definitely fall over the next couple of years. But I do see a transition towards per-hour rates. So, translators will have to increase their per-hour output to maintain their current pay. As for post-editing, it could very easily be a thing of the past soon and people will begin working with interactive MT solutions like Lilt.
What do you think about Machine Translations in the context on Indian Languages?
There is a lot of work going into what they call Low Resource Languages, which, unlike English, French, German and Italian have smaller corpora of data. Since so much work going into Low Resource Languages with Neural MT, I would assume some of it will be directed to India languages. So I feel there is going to be a big change. In the past there was hardly any MT for these languages but with Neural MT, these models should provide decent output even for Indian languages.