The Translation Engine Supporting Global CX

What do we mean when we use the term “customer experience” or CX?

CX is the sum of impressions associated with a brand for a particular customer. CX is built over different stages of the customer journey and encompasses all the interactions the customer has with the brand. It may include both positive and negative feelings, incidents, and expectations.

Customer experience can be linked with the direct experience and overall perceptions of:

• Convenience
• Responsiveness
• Speed
• Quality
• Aesthetics
• Value alignment
• Pleasure
• Ease of access
• Omnichannel capabilities

While CX is sometimes equated with making more information available online, with customer surveys, or increasing digitization of customer interactions, it is more. Much more.

Commonly used snapshot measurements of CX like CSAT (Customer satisfaction) and NPS (Net promoter score) give executives a sense of it, but we should understand that true CX "goodwill" is not so easily measured and put on KPI dashboards.

CX is a continuous journey that begins with the first contact or impression from a brand or enterprise and continues for the life of the customer engagement.  

Why CX Matters

As more of the world gets more accustomed to a digital-first buyer journey, especially with the challenges created by the pandemic, companies have focused on more engagement with their customers via digital channels. In an increasingly digital world where products and services are scrutinized in the court of public opinion, customer experience (CX) is king.

What truly makes for a good experience? Speed. Convenience. Consistency. Friendliness. And one big connector: human touch - that is, creating real connections by making technology feel more human, and also giving employees what they need to create better customer experiences.

People are increasingly more loyal to the retailers, products, brands, and devices that consistently provide exceptional value with minimum friction or stress. Value is created from the complete CX.

Every interaction is a direct reflection of the company’s brand. Convenience—seamless transition from tablet to smartphone to desktop to human—is a baseline expectation.

The benefits of providing superior CX are increasingly clear:

• Consumers will pay a 16% price premium for a great customer experience AND are more likely to be loyal to the brand. – PwC
• Among US consumers, 63% say they’d share more information with a company that offers a great experience. - PwC
• “Customer Experience leaders grow revenue faster than CX laggards, drive higher brand preference, and can charge more for their products.” – Forrester’s Rick Parish.
• #1 – By 2021, customer experience will overtake price and product as the key brand differentiator – Walker
• Consumers are 3.7x more likely to recommend an organization after a positive customer experience. - Qualtrics
• Maximizing satisfaction in the customer journey can lift revenues by up to 15% and lower the cost of serving customers by up to 20%. - McKinsey

The consequences of providing what is regarded as inferior CX are also quite clear:

• One in three consumers says they will walk away from a brand they love after just one bad experience. This figure is even higher in Latin America, at 49%. – PwC
• 92% would completely abandon a company after 2 or 3 negative experiences. - PwC

Thus, we see increased attention on what it means to get CX right. Done right, technology can help companies create phenomenal customer experiences and reap the resulting benefits: however, a large majority of the top-performing companies report paying close attention to the human experience around digital and tech.

Companies won’t be able to solve their customer experience problems with technology alone—it’s just the enabler. Employees also need to be empowered, supported, and allowed to make mistakes, and a majority of customers around the world report that a positive experience with a brand is more important than advertising.

Nearly 80% of American consumers point to speed, convenience, knowledgeable help, and friendly service as the most important elements of a positive customer experience. It is important to note that it is the combination of technology working together with competent, friendly humans that matters.

According to a PwC survey, only 49% of US consumers say companies provide a good customer experience today. So, there is much room for improvement.

Cultivating a customer-centric culture is a critical component of building the brand CX. This calls for targeted efforts to build employee capabilities across the enterprise. Truly customer-centric companies train employees at every level of the organization and in every function, from sales to accounting, to make sure they understand the role they play in maximizing the customer’s experience.

The Pandemic Impact

The pandemic has accelerated the steady growth trend in digital interactions with customers, and many believe that the opportunities and paradigm shifts that have emerged will persevere postcrisis.

The increase in digitization has affected both the B2C and the B2B market. Consumers who grow accustomed to superior B2C digital CX also expect the same kind of user experience in B2B scenarios. The rate of change in digitization was dramatically influenced by the pandemic. McKinsey estimates that the pandemic impact across the globe was to accelerate digitization by an average of seven years.

The pandemic drove B2B customer behaviors to begin to shift dramatically, favoring video conference interactions with sales reps over in-person meetings. eCommerce marketplaces have emerged as a preferred mode of interaction over direct sales rep interactions in many B2B settings in the last year (2021).

The Key Requirements to Enhance CX

Much of the customer journey today involves a buyer interacting independently with content related to the product of interest. CX leaders today increasingly understand that on digital platforms, useful, and relevant content is how this journey is enhanced and improved.

Understanding and providing relevant content that matters to the customer is a prerequisite for providing superior DX and CX and enabling customer success.

Customers want personalized, relevant information to guide their purchase decisions, and also want self-service support content to be able to be as independent as possible after they buy a product.

Thus, brands need to provide much more content, both in terms of ongoing volume and relevance, than they traditionally have provided.

Gartner and other analysts have identified that the ability of a buyer to access and gather relevant information related to a purchase decision independently is a critical requirement of the modern buyer in both B2C and B2B scenarios.

There is now also an acknowledgment that customer reviews of unique and personal CX experiences of previous customers are very important to new customers who are in the final stages of their purchase evaluations.

Customers across industries have shown a strong preference to see authentic, validated reviews of shared independent customer experiences that may inform a buyer of potential problems before purchase, and also identify possible support issues after the purchase.

In many eCommerce marketplaces, customer reviews can be THE most influential driver of purchase behavior.

The transparency and trustworthiness of these reviews are often more impactful to a buyer than much of the organization’s own product marketing content.

“Customers spend much more time doing research online -- 27% of the overall purchase evaluation and research [time]. Independent online learning represents the single largest category of time-spend across the entire purchase journey.”

-- Gartner

 Developing better CX also involves building better collaboration and knowledge and data sharing models across the organization.

Customer problems are not merely sales issues. The root cause of customer problems often runs across departments, and developing structural solutions, and improving the CX requires collaboration, cooperation, and knowledge sharing across the organization.

Experts have noted that leaders develop “digital agility”, that enables cross-function collaboration focused on mapping and optimizing customer journeys which in turn leads to gathering the right data to drive predictive analytics to enhance and improve CX.

While technology is a critical enabling tool for improving CX at a structural level, delivering better CX also requires a significant investment in the employee experience, empowerment, training, and a cultural reset to develop customer-centric and customer-friendly solutions. Alert, motivated employees are needed to identify and reduce the many friction points that exist in most organizations.

Some experts say that the human development efforts are as critical as the core-technology foundations, and the understanding of the buyer journey.

Reducing friction for customers requires that the organization empower employees to focus on higher customer satisfaction, which can only happen in a supportive, and empowered culture of forgiveness that encourages learning.

This may require new ways of working, with more focus on the employee experience, and a sophisticated view of the human-machine interactions, to ensure that the foundational technology supports and enhances BOTH the customer and the employee experience.

Good customer experience leaves consumers feeling heard, seen, and appreciated, and this is most often based on human interactions.

Automation matters, but making sure customers can reach a human when one is needed is often critical. Additionally, automated solutions should “learn” from human interactions so the automated experiences also improve.

The Translation Implications of CX for the Global Enterprise

In a nutshell, the implications include:

• New types of content
• Both internal and external content
• More volume at varying levels of quality
• Use of new tools beyond the capabilities of current TMS systems and generic MT
• More dynamic, real-time, and integrated into core IT infrastructure.

Language translation in the enterprise is most often managed by localization departments that focus on relatively static content.

Even though most localization departments have been seeing a steady increase in demand for translation, traditionally this department has focused on getting critical customer-facing documentation, marketing materials, and surface-level web content translated into the languages relevant to target international markets.

Much of the focus of localization has been on what is deemed mandatory content. This includes packaging material, user documentation, marketing collateral, and software interfaces and manuals. The top-level content of the website might also be translated, but much of the product, support, and deep marketing content is beyond the scope and budget of most. Even for the largest organizations, the translation volumes rarely exceed a million source words a month.

With a CX focus, the picture changes drastically. CX is much more focused on dynamic, fast-flowing, continuously changing, real-time content. It is also broader and deeper in terms of content coverage than is typical for localization. To provide the global customer the same CX as a US customer, significantly more volumes of content need to be translated.

Much of this content is unstructured and related to communication, broader, deeper customer support, listening to customer feedback, customer reviews, internal communications, knowledge sharing between globally dispersed internal workgroups, social media analysis, and much more.

Thus, in terms of monthly word volume, this can often mean that hundreds of millions of words or even billions of words need to be translated to provide the same buyer journey to global customers.

Given the sheer volume of content that has to be translated by machine, also requires that it is increasingly served to customers without comprehensive PEMT or even any editing.

The translation production mode used for a million words a month does not make sense for the production scenario which requires a billion words a month.

Localization processes often use a Translate-Edit-Proof (TEP) approach where all the translated content goes through one or more levels of human processing and refinement. Research shows that there is limited use of MT by localization-focused teams, and on average, in 2021, the top 170 LSPs use MT in less than 15% of their customer translation production workload.

Also, generally, the use of MT post-editing (PEMT) focuses on virtually all the MT content, and the motivation for any limited use of MT is primarily for efficiency reasons (cheaper and faster).

Global CX cannot be improved without using much more MT. MT acceptance is increasing because of the combined impact of the following:

• Market leaders show the benefit of translating CX-related content with sophisticated and expert use of MT,
• Human linguist evaluations show that MT output is often indiscernible from human translation,
• Direct customer feedback on the usefulness of MT content suggests that customers are willing to accept “imperfect” MT to get broader content access and faster response,
• To ensure that the global CX is equal across languages to minimize revenue loss,
• The increasing importance of community and customer-created content (UGC),
• The growing importance of self-service for most digital-first customers,
• Emerging fast-growing markets in Africa, the Middle East, and South Asia will need more MT.

The inconsistent translation practices across languages result in sub-optimal CX for customers of less-translated languages who prefer to see content in their own language.  This in turn can dramatically affect revenue from underserved global markets. 

Since CX content volumes involve a steady flow of hundreds of millions of new words, not only is specialized MT technology necessary but there also needs to be a close and tightly integrated feedback loop with linguists who provide corrective feedback to enable continuous improvement of the MT output quality.

The optimal translation production mode for CX scenarios needs highly responsive MT that continues to evolve in quality with ongoing feedback. As it is simply impossible to deliver a billion words at TEP quality, organizations must analyze the content and determine what the right mix of MT and human input is for different kinds of content. The higher the volumes the greater the need for automated translation.  

Thus, MT output for internal communications, user reviews (UGC), or social media analysis will have less human oversight, but customer support content can get more human input to ensure the MT output quality is higher for knowledge base and technical FAQ material.

The translation production process in the current "digital-first" reality means that digital content and software interfaces are in continuous update and evolution cycles.  Thus, the translation production process frequently involves tens of thousands of small (50 -100 words) translation tasks per month and is often called “continuous localization”. The two most widely used tools are translation memory (TM) and translation management systems (TMS) which are important to enable this kind of localization at the required scale.

However, while TM is valuable to the MT development process and is “AI-ready”, the same cannot be said for TMS systems in the CX translation-needs context.

TMS systems are instrumental in working within the structured localization context, which needs the translation workflow and project management, and monitoring capabilities that typical localization content requires.

Current TMS systems, however, are not the optimal tool for the less structured, more varied, faster-flowing, CX content. TMS systems are optimized for relatively low-volume high-touch, project-management-heavy, localization workflows. They were simply not designed to handle the sheer volume of UGC, and real-time dynamic communications-related content around CX, that needs little or none of the core TMS functionality.

As translation focus moves to service high-volume, fast-flowing social media, UGC, real-time chat, email, and collaboration around knowledge content, it makes much more sense for the systems containing this content to connect directly into adaptive MT with integrated corrective feedback (learn-fast, improve quickly) systems, or data interchange layers that are built for 100X+ the scale of most current TMS systems.

Both automated and human translation services need to connect to Enterprise CX infrastructure much more seamlessly, dynamically & directly to deliver purpose-ready translation where needed.

This explains the increasing interest in concepts like “LanguageOps”, “TranslationOS”,  "Language Platforms" and the notion of a Translation Layer for the Enterprise OS.

As translation penetrates deeper into the core IT infrastructure of the enterprise, there are a growing set of use-cases that have little or no need for the core functionality of TMS systems, especially for CX and eCommerce-related content.

Adaptive MT-connected TM tools are still valuable in this greatly expanded translation context, as they provide a means to input linguistic feedback and drive improvements in MT output quality over time. It can only be done over time because there is too much content.

This is easily understood in the context of content seen on a multilingual eCommerce site. The content on any given page is not treated the same way when considering translation production options. Many eCommerce sites will have millions of such listings. There is always a small volume of critical content that needs to go through the standard localization process, but the large bulk of the volume is better handled by different human-machine mixes that aim to produce translations that are fit for purpose and enhance the CX, and facilitate customer evaluation and understanding.

Translation in the Age of CX is different because it is more dynamic, more varied in quality, and more real-time and instantaneous. The key attributes can be summarized below:

• Scales from millions to billions of words a month
• Integrated into critical communication, collaboration, and customer data platform infrastructure
• Able to vary production modes for varying translation quality needs
• Enables pervasive but differently optimized translation capabilities across the enterprise
• Goes beyond the capabilities of any of the currently available TMS systems

Understanding MT Quality - What Really Matters?

This is the second post in our posts series on machine translation quality. 

The reality of many of these comparisons today is that scores based on publicly available (i.e. not blind) news domain tests are being used by many companies and LSPs to select MT systems which translate IT, customer support, pharma, financial services domain related content. Clearly, this can only result in sub-optimal choices.

The use of machine translation (MT) in the translation industry has historically been heavily focused on localization use cases, with the primary intention to improve efficiency, that is, speed up turnaround and reduce unit word cost. Indeed, machine translation post-editing (MTPE) has been instrumental in helping localization workflows achieve higher levels of productivity.

Many users in the localization industry select their MT technology based on two primary criteria:

1. Lowest cost
2. “Best quality” assessments based on metrics like BLEU, Lepor or TER, usually done by a third party

The most common way to assess the quality of an MT system output is to use a string-matching algorithm score like BLEU. As we pointed out previously, equating a string-match score with the potential future translation quality of an MT system in a new domain is unwise, and quite likely to result in disappointing results. BLEU and other string-matching scores offer the most value to research teams building and testing MT systems. When we further consider that scores based on old news domain content are being used to select systems for customer support content in IT and software subject domains it seems doubly foolish.

One problem with using news domain content is that it tends to lack tone and emotion. News stories discuss terrorism and new commercial ventures in almost exactly the same tone.  As Pete Smith points out in the webinar link below, in business communication, and customer service and support scenarios the tone really matters. Enterprises that can identify dissatisfied customers and address the issues that cause dissatisfaction are likely to be more successful. CX is all about tone and emotion in addition to the basic literal translation. 

Many users consider only the results of comparative evaluations – often performed by means of questionable protocols and processes using test data that is invisible or not properly defined – to select which MT systems to adopt.  Most frequently, such analyses produce a score table like the one shown below, which might lead users to believe they are using the “best-of-breed” MT solution since they selected the “top” vendor (highlighted in green). 

English to French	English to Chinese	English to Dutch
Vendor A – 46.5	Vendor C – 36.9	Vendor B – 39.5
Vendor B – 45.2	Vendor A – 34.5	Vendor C – 37.7
Vendor C – 43.5	Vendor B – 32.7	Vendor A – 35.5

While this approach looks logical at one level, it often introduces errors and undermines efficiency because of the administrative inconsistency between different MT systems. Also, the suitability of the MT output for post editing may be a key requirement for localization use cases, but this may be much less important in other enterprise use cases.

Assessing business value and impact

The first post in this blog series exposes many of the fallacies of automated metrics that use string-matching algorithms (like BLEU and Lepor), which are not reliable quality assessment techniques as they only reflect the calculated precision and recall characteristics of text matches in a single test set, on material that is usually unrelated to the enterprise domain of interest. 

The issues discussed challenge the notion that single-point scores can really tell you enough about long-term MT quality implications. This is especially true as we move away from the localization use case. Speed, overall agility and responsiveness and integration into customer experience related data flow matters much more in the following use cases. The actual translation quality variance measured by BLEU and Lepor may have little to no impact on what really matters in the following use cases.

The enterprise value-equation is much more complex and goes far beyond linguistic quality and Natural Language Processing (NLP) scores. To truly reflect the business value and impact, evaluation of MT technology must factor in non-linguistic attributes including:

• Adaptability to business use cases
• Manageability
• Integration into enterprise infrastructure
• Deployment flexibility   

To effectively link MT output to business value implications, we need to understand that although linguistic precision is an important factor, it often has a lower priority in high-value business use cases. This view will hopefully take hold as the purpose and use of MT is better understood in the context of a larger business impact scenario, beyond localization.

But what would more dynamic and informed approaches look like? MT evaluation certainly cannot be static since systems must evolve as requirements change. Instead of a single-point score, we need a more complex framework that provides an easy, single measure that tells us everything we need to know about an MT system. Today, this is unfortunately not yet feasible.

A more meaningful evaluation framework

While single-point scores do provide a rough and dirty sense of an MT system’s performance, it is more useful to focus testing efforts on specific enterprise use case requirements. This is also true for automated metrics, which means that scores based on news domain tests should be viewed with care since they are not likely to be representative of performance on specialized enterprise content. 

When rating different MT systems, it is essential to score key requirements for enterprise use, including:

• Adaptability: Range of options and controls available to tune the MT system performance for very specific use cases. For example, optimization techniques applied to eCommerce catalog content should be very different from those applied to technical support chatbot content or multilingual corporate email systems.
• Data privacy and security: If an MT system will be used to translate confidential emails, business strategy and tactics documents, human evaluation requirements will differ greatly from a system that only focuses on product documentation. Some systems will harvest data for machine learning purposes, and it is important to understand this upfront.
• Deployment flexibility: Some MT systems need to be deployed on-premises to meet legal requirements, such as is the case in litigation scenarios or when handling high-security data. 
• Expert services: Having highly qualified experts to assist in the MT system tuning and customization can be critical for certain customers to develop ideal systems. 
• IT integration: Increasingly, MT systems are embedded in larger business workflows to enable greater multilingual capabilities, for example, in communication and collaboration software infrastructures like email, chat and CMS systems.
• Overall flexibility: Together, all these elements provide flexibility to tune the MT technology to specific use cases and develop successful solutions.

Ultimately, the most meaningful measures of MT success are directly linked to business outcomes and use cases. The definition of success varies by the use case, but most often, linguistic accuracy as an expression of translation quality is secondary to other measures of success. 

The integrity of the overall solution likely has much more impact than the MT output quality in the traditional sense: not surprisingly, MT output quality could vary by as much as 10-20% on either side of the current BLEU score without impacting the true business outcome. Linguistic quality matters but is not the ultimate driver of successful business outcomes. In fact, there are reports of improvements in output quality in an eCommerce use case that actually reduced the conversion rates on the post-edited sections, as this post-edited content was viewed as being potentially advertising-driven and thus less authentic and trustworthy.

True expressions of successful business outcomes for different use cases

Global enterprise communication and collaboration

• Increased volume in cross-language internal communication and knowledge sharing with safeguarded security and privacy
• Better monitoring and understanding of global customers 
• Rapid resolution of global customer problems, measured by volume and degree of engagement
• More active customer and partner communications and information sharing

Customer service and support

• Higher volume of successful self-service across the globe
• Easy and quick access to multilingual support content 
• Increased customer satisfaction across the globe
• The ability of monolingual live agents to service global customers regardless of the originating customer’s language 

eCommerce

• Measurably increased traffic drawn by new language content
• Successful conversions in all markets
• Transactions are driven by newly translated content
• The stickiness of new visitors in new language geographies

Social media analysis

• Ability to identify key brand impressions 
• Easy identification of key themes and issues
• A clear understanding of key positive and negative reactions

Localization

• Faster turnaround for all MT-based projects
• Lower production cost as a reflection of lower cost per word
• Better MTPE experience based on post-editor ratings
• Adaptability and continuous improvement of the MT system

A more detailed presentation and webinar that goes into much more detail on this subject is available from Brightalk. 

In upcoming posts in this series, we will continue to explore the issue of MT quality assessment from a broad enterprise needs perspective. More informed practices will result in better outcomes and significantly improved MT deployments that leverage the core business mission to solve high-volume multilingual challenges more effectively.

Again, this is a slightly less polished and raw variant of a version published on the SDL site. The first one focused on BLEU scores, which are often improperly used to make decisions on inferred MT quality, where it clearly is not the best metric to draw this inference.