The Rising Value of Human Translation

Ever since I began working in the language services field, I hoped to shine a light on the work of human translators. That dream came true in the form of a book I co-authored called Found in Translation. After nearly a decade of hard work and pitching publishers, it felt surreal to land a book deal with Penguin Randomhouse. It was a chance to put translation and interpreting in the spotlight, raising broader awareness of the impact of these professions.

While that book is still very relevant today in highlighting the contributions of many incredible individuals who work in the profession, the business of translation has changed a lot since it came out. Are human translators better off than they were before? Here are six reasons I have great hope that society, and businesses, will place a higher value than ever before on human translation.

1. More people Understand the limits of machine translation

Pretty much anyone and everyone knows what Google Translate is these days. Machine translation (MT) has become so commonplace that many people consider it, not human translation, the default. Today you frequently hear phrases like:

  • We auto-generated the translations, but users are complaining about the quality.
  • This is for a marketing campaign; we can’t just use free generic translation on it.
  • Our SEO team is saying if we just auto-translate, it could harm our site’s overall ranking.

One very real, and well-known problem, is that a lot of what gets translated really shouldn’t be. Companies often assume their content has value without any proof of it. It’s hard to hear that the content your team has produced might actually offer limited to zero value to people in another country and language. Many buyers of translation find it easier to simply send content out for translation than to take a step back, assess the value, and really refine the plan to make it more strategic.

The benefit of this past mentality of “just translate everything” without any real concern over quality is that eventually, if the content does actually matter, humans take notice. Either the original content owners start to notice the content isn’t performing well when machine-translated, or human end users and audiences begin to complain or point out errors. However they make this discovery, society is finally waking up to the fact that there is a spectrum of translation quality, and even if machine translation keeps improving, human translation and human creativity remain on the “highest quality” side of the spectrum.

2. Content marketers are leveling up on international SEO

In the early days of SEO, marketers would often translate web pages, blog posts, and keywords in the hopes of ranking higher for keywords in many languages and gaining traffic. This actually worked during times when SEO was less competitive in certain local markets. Today, those strategies are no longer that helpful. In the current content environment, SEO has become far more competitive in all languages. Sites need to be reliable sources of deep content that is structured and interlinked just right, cited and linked to frequently, and mentioned on social media in association with the topics in question in each language, among many other factors. Content needs to resonate natively in each market and language, and must be in keeping with local search terms. That’s a much taller order.

As more companies keep going global, international SEO has also become both more complex and more specific in nature. Gone are the days when marketers went for a volume play only, and could leverage much of the same work simply by localizing it. Today, SEO requires extensive research into the top search terms for each country and language combination, prior to authoring, which often needs to happen independently and natively in each language. Localization still plays a part in the content marketing and SEO playbook, but it’s no longer considered the primary means of creating a net new blog in another language, as it once was.

Because SEO will remain a high-value activity at many companies for the foreseeable future, keywords won’t be translated per se. Content created for one market can be leveraged for another with totally different keywords, but only so long as the content is translated with utmost care. Marketers simply won’t trust that type of high-value work to machine translation. It’s far too risky, given what they are investing in the content, and the value it has in terms of traffic, lead flow, and revenue.

3. Google has changed its algorithm (again) to prioritize helpful content

We’ll see how the latest changes from Google play out, but I’m always happy when I see any reinforcing of the “helpful” or human-centric aspect. Google used to penalize sites that published machine translated content, but as neural MT improved, and companies began using MT with human post-editing, companies experimented a lot more with MT on their websites and in many cases saw good results.

Now, Google is stating that for sites to be helpful, they should not include significant amounts of auto-generated content, and we would assume that includes MT. But even if there are no penalties, the bar should be raised on ensuring the content is helpful. Helpful content is likely to change frequently, to have human actors working on it and updating it and improving it, on top of humans reading it and sharing it, and so on.

On this front, I’m excited because I believe more companies will prioritize working with human translators than ever versus pure MT, specifically because it will help them improve their SEO. While this is related to the point above, I mention it separately, because international SEO was already becoming harder for marketers, but now, with Google’s latest changes, there are additional elements at play, making human translators an important part of the solution.

4. Buyers care about Human Translators more than ever before

I’ve always believed that what makes or breaks a brand’s voice in another language are the human translators who work on its content in that language. More companies are hiring for internal linguistic roles directly at buy-side organizations. And today, buyers of translation want assurances that once they have gone to the trouble to train translators, they can continue to work with those translators who are familiar with their brand on a recurring basis. In those situations, many language service providers (LSPs) are really not providing much more value than glorified project management, merely connecting buyers to freelancers and paying them.

Savvy buyers of translation today know that their outcomes are linked not to which agency is sourcing the translators, but who those translators actually are, and their skill levels. This is one reason why LSPs that offer marketplaces can provide an interesting option for buyers, so long as those marketplaces are actually populated with high-quality and vetted translation providers that meet certain minimum criteria.

I am also loving the trend of more translators being vocal advocates and influencers on social media, especially bloggers, and podcasters. This is important, because visibility matters.

5. Privacy and security concerns keep growing in importance

The translation industry might be due for a shake-up in terms of data privacy and security concerns in the years ahead. Rare is the translation agency that differentiates based on this topic, namely because no one wishes to add unnecessary blockers to gaining access to content and delivering it back. But with data privacy and security becoming greater topics of concern each day, agencies will have to go to greater lengths to vet their linguists.

Either that, or even more buyers will start to in-house their human translation work, but with the added benefit of maintaining a secure environment in which content doesn’t need to leave in order for translation to happen. This has long been the case with highly sensitive work, such as the translation of content obtained from government intelligence gathering. But as more content moves online, and as machine translation APIs become even more widespread, certain types of content will require more encryption and security measures in both directions in order to be translated. Beyond that, human translators will have an important role to play as part of a more secure solution, especially for certain content types.

6. Quick-turn, high-risk translation Needs Require human-level quality

What kind of content carries risk today, but is also needed rapidly? The answer is that the types of content to which these two criteria apply is on the rise. Here are just a few examples:

  • Legal content. More and more legal content is available online, with terms of service, e-signature of contracts, and so on. Whenever these types of docs need updates, which can be very frequent, human translators will need to be available, but with more of an instantaneous expectation, to keep up with today’s demands. Companies need to be able to iterate quickly and launch updates in many languages at once. Only human translators can help to minimize the risks.
  • Marketing content. For many companies, if their marketing engine slows down, they run the risk of not hitting their sales targets, and failing in terms of overall financial performance. Every day matters in terms of lead flow, and in some cases, a single missed day of pushing a campaign live can result in a company not hitting their quarterly targets. Yet, marketing content typically can’t be machine-translated, and requires a high level of human involvement. For this reason, expect more human translators to play key roles, and to be needed for quick-turn projects of a highly urgent nature.
  • Security-related content. With cybersecurity taking on greater importance than ever, companies will not only need human translators as described above, to have greater control over where certain types of content flow, but also, for the express purpose of communicating to end customers when things go wrong. Consider any outage of any software product you’ve used recently (Zoom, Twitter, Slack, and so on). When thousands of users suddenly lose access, companies must react swiftly, and if they are global, need to provide security updates in many languages to their global customer base. Because such incidents are extremely embarrassing for companies, and can cause their customers to leave them, they won’t want to trust the messaging to machine translation, and will require human translators who can quickly jump in to help ensure a highly accurate translation that is on brand, and on message, instead.
  • Customer-facing content. There is a whole world of content meant for customers that really shouldn’t be left to MT tools, especially anything that ties to getting customers set up properly, onboarded and active in a software product, and so on. Another huge category of translation relates to billing and invoicing, and renewal communications. When you’re asking a customer to pay you money, you can’t risk machine translation getting things wrong. Precision is important, but so is the tone you use. For this reason, humans will need to play a role whenever the content is customer-facing, but especially when the content carries certain risk, such as a risk of non-payment.

Ultimately, I believed in the combined power of humans and machines to achieve greatness in many different disciplines, including translation. But the market will ultimately dictate how much value is placed on human translation, and in which specific circumstances buyers see the greatest possible value. What’s exciting to me is that the market conditions related to buying translation are changing, and I believe evolving to better highlight where a high value for human translation is obvious. When this happens, machine translation can provide more of a familiar accompaniment, fading more into the background, so that human translation can step into its hard-won starring role.

Nataly Kelly

Nataly leads localization at HubSpot and has previously held diverse roles in marketing, international operations and strategy, research, and product development. Her latest book is "Found in Translation" (Penguin). She writes for Harvard Business Review on topics of international marketing and global business. Nataly works remotely from New England, having lived in Quito (Ecuador), Donegal (Ireland) and the rural Midwest where she grew up.

Leave a Reply