The Vision and The Reality
by Renato Beninatto
In April 2021, João Graça, the co-founder of Unbabel published an article in Forbes magazine proposing LangOps as a new paradigm for the language industry, rightfully arguing that terms like localization and internationalization have run their course.
The article has sparked debate among industry leaders. Luminaries like Britta Aagaard and Jochen Hummel have hit the road sharing the vision of the end of the siloed approach to language services in the enterprise and its inefficiencies and the dawn of a cross-functional tech-enabled collaborative environment called language operations or, LangOps.
The vision makes sense from a logical and pragmatic perspective. Who wouldn’t want to have all their content and communications aligned and managed in a consistent manner across organizations, platforms, and geographies? What enterprise wouldn’t want to have their customers’ journey through offerings and systems be integrated to allow for rapid and appropriate responses to customer inquiries, no matter what language the customer speaks?
After many years of working with enterprise customers in multiple industries, I cannot count how many times I have worked on projects to centralize and consolidate language services inside organizations. Here are three examples that, in my opinion, hinder the dissemination of LangOps.
- Organizational structures. A company like Microsoft is present in 190 countries and has over 220,000 employees, and Volkswagen has over 660,000 employees in 153 countries. Just try to imagine coordinating all the translations required in product, sales, marketing, legal, HR, training, support, events, and any other departments.In analyses that I have done in various audits, a frequent finding is that the only place in the organization where all these departments meet if an escalation is required is the CEO. And if the CEO of a large enterprise knows the name of an LSP, it is usually because there was a major failure in the process that exposed the company to some liability. Not necessarily a good place to be. An added element is that companies are in constant reorganizations, and all business functions are constantly shuffled from one department to another, making it virtually impossible to standardize processes across areas.
- Original content and budgets. The first exercise I engage in when auditing localization practices is what I call “follow the content, then follow the money.” It is necessary to look at who creates content for whom and in whose budget that content is. If original content is paid by many different stakeholders in the enterprise, so are language services. The person who owns the marketing budget might have zero influence on the budget for customer support or human resources. Training might be paid for by the product organization or by the sales department. To add complexity, some costs are covered by the headquarters while others are assigned to in-country operations. Philosophically, it makes sense to try to centralize language services and offer them as a LangOps inside corporations. Many have tried, and many try still. But the truth is that organizations are dynamic with constant modifications in policies and priorities dictated by changing economic, environmental, and leadership changes. In order to accomplish the vision of LangOps, we would first have to be able to consolidate communications in the original language.
- Platform proliferation and new technologies. One of the key elements of the LangOps proposition is that language technology enhanced by machine translation and artificial intelligence can eliminate the barriers to cross-functional collaboration. I tend to agree in theory, but in practice this is something that might happen at a specific moment in time. It is a condition, not a state. The way content is delivered — and therefore the demand for language services — is in constant change with the proliferation of file formats, platforms, and technology solutions that crop up daily. In my experience, we tend to look at technology change almost linearly towards the future, when in reality we are dealing with change coming from all angles, all the time, and at different speeds. In the language industry alone, we went from approximately 500 technology solutions in 2017 to over 800 in 2022, according to Nimdzi’s Language Technology Landscape report. Each business problem has several alternatives that can be implemented in multiple scenarios.
The challenge now becomes: Which of these 800 technologies are going to compose the LangOps deployment? Which ones prevail when companies get acquired? Who is going to be the arbiter of these decisions?
LangOps is here to stay. But it is not a solution — it is a promise. The growth and survival of the language industry hinges around the fact LSPs manage complexity. They can deal with multiple variables, randomness, nonlinearity, and even irrational behavior. In a world that loves entropy, LangOps can put some order or predictability in language processes for parts of an organization during a certain period of time, while the whole gradually declines into disorder.
For the world at large, after a thorough explanation of LangOps, their reaction might just be the classic: “Oh, you are talking about translation!”
Traditionally, marketing teams were focused primarily on activities that took place mostly offline, including branding, messaging, public relations, design, advertising, and field marketing. As more…→ Continue Reading
Subscribe to stay updated between magazine issues.