In a keynote address entitled The Post Localization Era at the LISA Globalization Summit in October 2000 (published in issue 9/iv of the LISA Newsletter) Roger Jeanty, President of Lionbridge Technologies, highlighted the root of the problem. Jeanty chronicled the development of the language industry, describing "three distinct eras in an increasingly complex ecology" (see sidebar).

Pre-Localization Era pre-1982 The focus of the fragmented cottage industry... was on content, and the value chain consisted of language domain expertise, resource management and possibly project management." Localization Era 1982–1999 "...translation agencies turned into localization service providers... the result was a piecemeal approach in which solutions were developed for specific problems, but their scope was narrow..." Post-localization Era 1999–present "Margins will improve as the business models migrate away from the pure price per word model to models based on time, quality, cost but also availability..." Source: The Post-Localization Era

His straightforward analysis provided an insightful and perhaps unintentional commentary on the situation. It not only accentuated the disproportionate historical emphasis on a relatively small market segment, but also illustrated how limited market analysis had unwittingly created unfavorable competitive pressures. Not only had industry heavy-weights focused on a market segment representing less than 25% of total translation spend, but in the process they had undermined overall industry profitability, creating an "intense [industry] pressure to commoditize translation." The resulting climate, as Jeanty summarized, "hampered service providers' ability to invest in and grow the industry."

Market Models

We should not let ourselves get carried away by gloom and doom, however. In fact, TRADOS' experience suggests that the solution to industry troubles might be surprisingly simple. By organizing data taken from translation-related case studies from over 75 multinational organizations, the company has created simple market models that provide insightful explanations for everything from mega-merger headaches to workflow product debacles. We give the unembellished results below as a starting point for further discussion. But be warned: before proceeding, check your experience and expectations at the door. These models are not meant to cater to ivory tower academics, and they fly in the face of much conventional industry wisdom. Rather, they represent a careful review of the (sometimes painful) real-life experiences that are part and parcel of work in the trenches.

Model 1: Translation in two dimensions

The first model is the result of lessons straight from the school of hard knocks. Many of our original business plans for new products and services were based on a traditional view that segmented the market in terms of volume and repetition. As the performance of these new initiatives faltered, however, we were forced to re-evaluate the accuracy of our cherished mental models. As a result, we collected translation-related data points from over seventy-five organizations facing the localization challenge and tried our hand at organizing them in meaningful ways. After many attempts to create a simple yet comprehensive approach that explained the trends we observed across the organizations, we finally came up with the model below.

Figure 1. Translation in two dimensions

While it's not much to look at (yet simple enough to undoubtedly provide ample fodder for the skeptics), the sharp lines of Figure 1 provide a framework for analyzing the tough questions facing the industry. The x-axis represents process sophistication, ranging from organizations that are involved in the design and execution of every step of the translation process on the left, to organizations that throw work over the wall in a completely out-sourced model on the right. The y-axis represents the organizational structure employed to manage translation. Large, dedicated translation departments sit up top while ad-hoc, do-it-'cuz-you-hafta efforts lie below. Unfortunately, the data points (i.e. mapping of companies onto the model) are not shown due to disclosure issues, but by mapping organizations that are familiar to them, industry insiders will be able to bring themselves up to speed and quickly recognize many of the same trends we point out later in this article.

Model 2: Translation in three dimensions

Figure 2. Translation in three dimensions

Like any good model, Figure 1 allowed us answer many questions, but it also prompted us to take a closer look at the details. In order to do this, we further dissected the model by trying to construct sub-sectors within the basic framework. Figure 2 shows an initial attempt to superimpose a z-axis given the generic label "Fragmentation Dimension." Much of the complexity within the language industry is due to the highly fragmented nature of the landscape, and the introduction of a third dimension is a crude attempt to represent this reality. This step, while not directly contributing to our understanding of the underlying forces in the market, is important to note because it represents the most sophisticated level of market analysis publicly available prior to our efforts. For example, the Allied Business Intelligence report gave a nod to the idea that the market is extremely fragmented, but it did little to explore that fragmentation.

Model 3: The Multi-dimensional Model

Figure 3. The multi-dimensional model

By collecting additional data on factors contributing to the "Fragmentation Dimension" of Figure 2, we created the final model in the series, shown below. Contrary to our initial hypothesis, which proposed that one main organizational attribute (like file type or IT sophistication) could account for data clusters within each quadrant of the model, we discovered that several important dimensions are necessary to accurately describe meaningful market segments in the highly fragmented marketspace. The resulting multi-dimensional model provides a powerful tool for isolating and dissecting market sectors that share superficially unrelated, yet important characteristics. (Ideally, these attributes should be represented as separate dimensions, but for visual purposes, we have represented these as planes along the z-axis as shown in Figure 3.)

Industry Trends

Typical industry market segmentation has had limited application because it failed to focus on customers. Instead, it has all too often classified organizations in terms of the market vehicles available at the time to solve the translation problem. While there is value in analyzing which type of company is attracted to a particular solution (machine translation or translation memory, for example), the series of models developed as a result of the blood, sweat and tears we poured into our efforts provide exceptional value because they focus on the consumer. Our segmentation is designed to seek out the raw customer, focusing on specific customer attributes that comprise the individual company profile. The real power of this approach can be seen by analyzing the trends that emerge as companies are mapped onto the figures. The examples below summarize three simple yet powerful concepts that provide a refreshing reality check for the industry.

Translation in Perspective

Figure 4. Translation in perspective

Table 1. Characteristics of the translation quadrants shown in Figure 4.

Model 1 creates a solid framework for analyzing common characteristics of useful customer segments (as shown in the chart below). But perhaps its most important contribution is that it allows these same customer segments to be evaluated in terms of overall size. By mapping data from various sources, we gain an interesting perspective (see Figure 4). Quadrant 2 has traditionally offered a market segment with characteristics that make building and sustaining a business easier. Localization firms have focused on organizations that treat translation as a business process. Typically, this segment has rewarded firms with high volume, high repetition, and predictable turnover. But increased supply of translation has led to competitive pressures that have had a detrimental impact on overall industry profitability. More importantly, it has fostered a vicious cycle of competition that has concentrated efforts on a small part of the market while leaving a vast amount of work unexplored. (Rough calculations have indicated that as much as 80% of the translation market falls outside the traditional market served by members of LISA.)

The Multilingual Development Lifecycle

Figure 5. The multilingual development lifecycle

One of the most interesting trends observed in analyzing the implications of Figure 4 on our original model was the Multilingual Development Lifecycle. In North America, many organizations start out selling products to home markets, requiring little or no translation (see Figure 5, data point A). As the market penetration of a product increases, however, it typically begins to find its way to foreign markets, requiring translation of documentation and marketing materials. In most cases, a Subject Matter Expert (SME)—often a project manager, marketing manager or documentation resource—is tasked with overseeing the translation process. While this can sometimes end up as a one-off project, if an individual participates in multiple translation projects, they begin to move upward on our y-axis—almost becoming (by default) the translation resource for the department or company. As their experience increases, they also gradually increases their own sophistication, in essence creating a process to handle the recurring translation work. At some point, translation headcount is appropriated to handle the problem (see Figure 5, data point B) and the organization finally moves to the stage where translation is viewed as a business process (see Figure 5, data point C). This natural development process, which we later called the Multilingual Development Lifecycle, provides important clues on how to create far-reaching solutions for a dynamic market.

The Service Provider Dilemma

Although the translation market is valued at almost $14 billion by Allied Business Intelligence, service providers have found it very hard to grow profitably. Why have economies of scale not worked in their favor? Actually, it turns out that the "Fragmentation Dimensions" shown in Figure 3 are the root of the problem. Large service providers tried to attack broad market segments, but soon found that the segments are not homogenous on various levels. In reality, there are many differentiating dimensions. The natural response within consolidated firms was to leverage economies of scale by focusing on a narrow portion of the market. While such action is designed to increase competitive advantage, it effectively mitigates much of the benefit derived from scale, committing firms to a particular market without establishing a differentiating strategic position. The net effect was a self-imposed limitation to leverage the market as a whole. The proof is in the pudding. Witness divisions in Bowne and Lionbridge that specialize in specific industries, targeting organizations with a strategic profile. While these divisions are created to leverage a common infrastructure, differentiating characteristics often require the development of specialized resources. The market fragmentation dimensions make it very difficult to serve a heterogeneous client base successfully.

The Technology Provider Dilemma

With billions of dollars being spent annually on translation, why have translation technology providers only managed to reach US $100m in combined revenues? By mapping target markets on to our series of models, we found a simple explanation. Technology providers have simply tried to treat the market as a uniform whole, ignoring more sophisticated marketing techniques that provide opportunities for differentiation. After significant spending to create a technology solution, Jaap van der Meer, President of Alpnet conceded this point by observing, "As customers will generally agree, there is no 'one size fits all' in the localization services market. Every customer has specific file formats, specific management processes and wants to interact with their service partners. While we were trying to build a generic system, we learnt that this is practically impossible."

Unlocking the Potential of Highly Fragmented Markets

Taken together, the models and trends discussed above do much to explain the current status of the industry. Essentially, many of the challenges being faced today can be traced to three important points:

1. a distorted view of the market has created unfavorable market conditions and eroded profitability;
2. the multilingual lifecycle creates variation in processes and development across industry segments; and
3. the highly fragmented nature of the market has created barriers to growth for both service providers and technology vendors.

While these factors represent a formidable challenge for the industry, a proper diagnosis of the challenges creates the necessary pre-conditions for a prescription for success.

What do we have to do?

It's obvious that the industry currently represented by LISA (as well as LISA itself) has to step out of quadrant 2 and find ways to serve translation needs in the other quadrants. Only then we can scale our businesses and begin to enjoy forecast growth rates. Based on our research, the fundamental component of an overall strategy for unlocking the potential of the translation market is the development of industry standard workflows and processes to create the critical mass necessary to construct profitable market segments outside of Quadrant 2. The Multilingual Lifecycle describes a natural trajectory that creates variation in basic processes across the industry. But taken together with additional information gained from Translation in Perspective (Figure 4), it can also act as a guide to developing effective service and technology solutions.

Standardizing Workflows

Appropriate industry standards established by a neutral third party like LISA could provide a common language for the optimization of overall translation workflows, while streamlining the processing of individual tasks within those workflows. Standard bidding, invoicing and editing conventions could provide the foundation necessary to realize efficiency gains along the value chain. Standards would enable the technology providers to develop the software to automate these processes. Moreover, when performed within the framework of the market model series, standards initiatives would further strengthen the industry by overcoming the glass walls of perception that confine the industry to Quadrant 2. In essence, proper industry standards would be the catalyst to de-fragment the industry along the Multilingual Lifecycle, driving growth and opening up a stifled market.

The bad news is that as an industry we face tougher challenges then expected in taking our game to the next level. But the good news is that it is our own perception - rather than external forces - that has prevented us from unlocking the true potential of the market. Indeed, language is hot. And the secret to our success will be found in our ability to collectively create standard-based solutions that truly sizzle.

---

Jochen Hummel (jochen@trados.com) is CTO and Chairman of TRADOS Corporation.