Business metrics and KPIs for localization

Making a business case for localization is a lot easier today than it may have been in the past. The advantages of winning new customers or satisfying existing ones with products and content provided in their native language are obvious. That so many successful enterprises use localization to generate more than 50% of their revenues internationally is an argument that is hard to beat. However, many localization managers may still struggle in supporting their case with concrete numbers and metrics that would stand as solid proof as they present their plans and budgets internally.

That is not say that there is a shortage of metrics being used. But many of these are, understandably, operational, focusing on the efficiency of localization operations and measuring the various aspects of the age-old trifecta of quality, time and cost. Measuring the return on investment (ROI) of globalization efforts is notoriously difficult, especially the return part of the equation. But it is not impossible, so having a good set of practical business metrics that can serve as key performance indicators (KPIs) is useful. Sound business metrics can also go a long way toward changing the perception of localization from cost center to revenue enabler.

Currently, there are no generally accepted standards. It’s not that there have been no efforts to develop a set of industry-wide metrics. Those long enough in the industry may recall the Localization Metrics Initiative (LMI), spearheaded by the Localization Institute back in 2003. Perhaps as a proof of the industry dynamics, of the then 14 participating companies, only five — Canon, Cerner, Cisco, IBM and PTC — exist today as independent organizations. GALA, for its part, has developed the concept of supply-side metrics via its Vendor Metrics Initiative.

When considering localization metrics, it makes sense to distinguish between three areas: presales content localization (marketing, websites, social media or advertising); actual product or service globalization; and after-sales content localization (such as customer support or knowledge bases).


Presales metrics

While some organizations may view localization as a cost, more forward-looking and mature companies consider it an investment capable of enabling international revenue growth. There is a range of leading and trailing indicators that look at the localization’s contribution toward generating sales in international markets. International market growth rate and relative market share in a given local market are two that provide the most relevant information.

However, the devil is in the detail when it comes to understanding how product localization actually influences the extra revenue. If an original nonlocalized product was available in the given market, it makes sense to establish a baseline revenue and calculate the increase in sales, adjusted by the “organic” market growth rate, meaning the rate of growth that would incur regardless.

In calculating this ROI in localization, it is necessary to include other nonlocalization investments into new markets. On the flip side, the “return” part of any calculation should include the decreased cost of doing business in a given locale, such as lower support costs.

Most investments into product localization are hardly discreet events. Companies invest in ongoing local product releases, expanding the amount of content made available in a given local language over time. As a result, many localization ROI calculations may fail to account for the ongoing nature of such an investment in localized assets.

So an alternative option is perhaps to calculate the return on localization assets, as a parallel to the financial ratio of return on assets (ROA). This indicates the efficiency of using the assets companies have at their disposal. In the context of localization, this metric would use the income derived from localized products or services divided by the total or average value of the existing localization assets in the same period.

The beauty of this approach is that it forces companies to understand the current value of such assets, which includes the past investments in multilingual content as well as its reuse and unavoidable degradation over time. This metric is especially useful for making comparisons over time.

Where hard data on revenue is not immediately available or relevant, organizations may measure a number of other quantitative or qualitative indicators, and web and the impact of multilingual marketing content is one area where a wide range of metrics is readily available.

Customer online engagement is a set of metrics that serves to show how engaged customers are with your company’s online presence — websites, blog, online support pages, discussion forums and so on. There is currently no one widely accepted measure of online engagement but, in the context of localization, the following are the key metrics:

Number of unique visits from target locale

Number of conversions and conversion rate by target locale

Percentage of visitors from target locale accessing content in local language

Click-through rates in local content pages or email campaigns

Bounce rate on localized pages

Average time on site by target locale

Percentage of new versus returning visitors per target locale

Search engine ranking for target keywords in target locales using preferred search engines

While these are useful leading indicators, the trailing indicator of the impact localized marketing content has is in the number of qualified international leads and the cost per such leads. Among other objectives, localized web presence serves to generate incoming leads from the given target locale, which can then be qualified and, ideally, converted to a customer.

The number of thus generated leads from language-specific marketing is a useful benchmark, as is the conversion ratio — how many eventually get converted into actual sales. In calculating the cost per such leads, it is important to consider the total cost of local marketing, including the web content localization, online advertisements, pay-per-click campaigns and social media, as well as internal costs. You can then drill down to the quality and cost of leads by specific sources.

One frequently overlooked insight is provided by comparing the behavior of customers from a given locale in the localized web content with how they navigate and how they convert in nonlocalized web content.

Online share of voice is an extremely useful proxy for how much your company is discussed online and in social media — the conversations people are having about your brand or your products, and how positive, neutral or negative their perceptions are. The beauty of it is that this enables you to narrow down on the share of voice by target markets or by language use.

While such a sentiment analysis is built into a growing range of commercial tools, free tools do a great job as well. Watching this data over time will give an indication of the impact a localized approach to a given language market has, as well as comparisons with competitors and their products.

The online share of voice is calculated as the number of mentions of your brand or product in a given market relative to the total number of mentions of all the brands or products in a given segment, and it is possible to analyze specifically positive, neutral or negative mentions. The actual share of voice would then consider only the positive and neutral mentions.

Brand awareness is another excellent leading and trailing indicator of your company’s success in a given local market. It is useful to conduct brand awareness research before and some time after major local-marketing or localization milestones, as well as over time, to understand the impact that your local product and content have and how your brand is known and perceived in individual markets. It is critical to test both unaided and aided brand awareness. While unaided or spontaneous brand awareness will show how much local users know of your product or brand without being prompted (brand recall), aided brand awareness will do the same while testing the knowledge of those surveyed giving them a specific list of product or brands (brand recognition).

Globalization metrics

There are myriad metrics and KPIs used in practice to measure the efficiency of the actual localization efforts. This is where many organizations routinely work toward performance measurements, such as:

Percentage of on-time deliveries

Number of linguistic or technical quality errors

Pass/fail rates

Project delays or requests for extensions

Average turnaround time

Adherence to budget

Business metrics would look at the bigger picture, consider the contribution that localization makes toward overall corporate goals and determine the efficiency of doing so. One useful measure is looking at the total localization headcount: the total number of in-house staff directly involved in localization activities. This can then be assessed as a number relative to the company’s total revenues or its international revenues — showing the contribution each employee involved in localization has to international sales, as a total or by markets and regions.

In a similar vein, another practical measure is considering the total localization cost as a sum of all the in-house and outsourcing budgets, taking these as a percentage of the total or international revenue. By the same token, companies can consider the total localization staff headcount and compare this with total localization costs, getting the globalization cost per employee involved in localization.

There are no general benchmarks in this area, given the obvious differences in the way organizations structure their localization operations, the proportion of their in-house staff and the amount of outsourced work. But these are very relevant metrics to watch over time, and to potentially compare with the performance of individual units in larger organizations.

Where meaningful data exists, companies may do the same on the product or product line level, measuring the total costs or headcount associated with localizing a given product with the revenues.

Where intercompany comparisons are possible is in the area of the cost per word. Here it makes sense to consider the fully-loaded average cost per word, which includes the overall internal as well as outsourced costs required to produce localized content divided by the total number of words localized. This measure enables an understanding of the overall efficiency of the operations and the total cost of localization on a per-word basis.

While it is useful to consider the unit word rates for individual languages, the truth is this often does not allow for comparing apples with apples. This is because they may include different activities. For instance, steps such as editing, proofing, terminology management, project setup, independent review and quality assurance may or may not be bundled into given per-word rates. In another case, falling unit word rates may belie major expenditures elsewhere in the overall production process.

Instead, the fully-loaded average cost per word considers the whole production workflow, and takes into account the internal as well as outsourced costs of any localization operation. Some organizations may also work with the average cost of outsourced words and exclude the cost of their internal operations. In this case, the measure will show the productivity of their localization outsourcing.

The average cost per word has gone down substantially in the industry over time, as a result of the years of investment into automation, translation technologies and processes that have substantially increased leverage and decreased the average cost per word. This average per-word cost has decreased dramatically — in some cases by 40-50% for many organizations, and certainly more than unit rates for new words in most target languages have over the period.

Time to market is another critical metric, looking at the time it takes to release the given product or content in local languages. A faster time to market normally indicates a better integration of the localization and internationalization functions into the overall product/release process, and a higher level of automation.

In this age of simship and agile development, a large part of product releases are multimarket and multi-language. This means product or content localization may become a critical path for global product releases. This serves to show how tightly linked localization is to the actual product development and marketing strategy.

There are no established standards for measuring time to market, but there are perhaps two meaningful ways of measuring the impact of localization in this context. The first looks at the time-to-market as a delta between the first release of a given product or content, and its release in local language. For simship scenarios, this will be zero. Particularly for first releases, you can also measure time to market in the original sense, from, say, when the product is conceptualized to the moment it is made available to local consumers. This time may differ for individual local language versions.

One aspect to always consider is the opportunity cost of a delayed release. Every week of delay in providing a given product on the market means a week of lost sales. The first to market advantage can be potentially huge.


After-sales metrics

Localization plays a huge role in the after-sales, with the main goals normally being to increase or maintain local users’ satisfaction, ongoing engagement through localized marketing and social media, and the reduction of support costs by providing local users with support content in their own language.

In many organizations, presales information is the priority, as it serves to get the client to buy and use the given product, and as such is seen as an investment. But localized after-sales support is often an afterthought. It’s expensive — customer support costs are a significant percentage of total revenue, it’s seen as a cost with no apparent contribution to the bottom line, and spread into multiple budgets.

However, change is afoot. First, more and more companies are beginning to move customer service from a cost center to a differentiator, using customer satisfaction metrics to measure how good they are. Second, there is a major increase in the provision of self-service channels, digital channels such as real-time chat and email, and in usage of communities for customer service. Finally, machine translation is dramatically reducing the cost of providing customer support content in multiple languages. And it’s finding users very grateful for local language content, even if linguistically imperfect, as a clearly much preferred option to no translation.

In this sense, the impact of providing localized support content can be measured as follows:

Support call deflection: Decrease in the number of support calls or other requests after publication of localized support content.

First contact resolution: Percentage of support requests resolved at first contact.

Customer satisfaction: When it comes to measuring local customer satisfaction, arguably nothing beats the simplicity of the Net Promoter Score.


Final recommendations

There are a huge number of potential metrics that localization professionals may use to measure the impact of their work. Focus on a few key metrics; three to five usually make the most sense. Also, be consistent. The beauty of KPIs lies in the way they enable comparisons. Make them visible. There is little point in hiding them. Promote them so that everyone knows how well you do. Finally, create a direct link to your company’s KPIs. This is an excellent opportunity to promote the critical role localization teams have in reaching the overall goals in every organization.