Software localization has changed. No longer an afterthought, localization has increasingly been given a seat at the table in discussions on content, product design, and globalization strategy.
Key to this change is the strategic use of data. It is now possible to show and predict the impact of localization.
During the latest edition of SlatorCon Remote, software localizers explored questions around strategy, data analytics, and KPIs.
Chi-Wei Chang, Head of Localization at work management platform Asana, and Giulia Tarditi, Manager of Global Experience at Qualtrics, talked about launching a software-as-a-service (SaaS) platform in a new market, including which metrics are meaningful, multilingual A/B testing, and the use of machine translation in SaaS localization.
So what does it take to launch a SaaS platform in a new market? In Asana’s case, it involves the localization of the product itself — a mobile and web app — along with key website pages. Localized content marketing and promotional events then spread the message that the platform is available in a new language.
According to Chang, the process involves “a lot of input coming from regional teams to determine what we’re localizing and what to prioritize, and to find out how the messaging resonates.”
Qualtrics, on the other hand, operates two layers of localization. The first involves the localization of the Qualtrics platform for the company’s direct clients in multiple markets. The second involves using localization to tune in to what end-customers are saying globally.
Tarditi explained, “These two layers are quite complex. It’s not just our own operations that are geo-inclusive, our clients also have their own global programs.” Localized customer feedback — garnered from call centers, social media, and online reviews — provide clients with insights they can use to build market-specific customer experiences.
A data-driven approach to localization allows SaaS platforms to be adjusted and calibrated for individual markets.
Chang told the SlatorCon audience that Asana, which is currently available in 14 languages, is in the “early stages” of localization and “just starting to look at data to inform decisions on what to localize and prioritize.”
She added, “We’re starting to work with the product-led growth team and look at market- and language-specific research. We have a strong analytics team for quantitative data as well as a qualitative research team.”
Asana also employs A/B testing. “If we see a market underperforming, A/B testing helps us figure out why,” Chang said. “We consider: Is it the language, the wording, the design, or just that the messaging that doesn’t resonate in the same way in that language and culture?”
Metrics That Matter
Data is also key to establishing meaningful metrics and demonstrating ROI. Tarditi told the SlatorCon audience that Qualtrics has recently pivoted to a “no-KPI-no-service” approach.
According to Tarditi, “Our internal teams approach us to help localize, but we will not get started until we have identified a way to measure the success of that initiative.”
The success of a localized support page, for example, may be measured by calculating the reduction in related support tickets. For an event landing page, a metric may be established with respect to the number and distribution of user registrations.
Convincing internal teams to set such metrics can be a challenge, however. “It’s a hard sell,” admits Tarditi, “because you need to tell your stakeholders they need to test the heck out of this thing, before they actually localize at scale.”
While Tarditi sees a place for blackbox metrics, such as turnaround time and cost per word, she believes “the more we talk about metrics as an industry and as professionals, the more we will identify and measure what really matters.”
She summed up by saying, “There’s no point deciding to localize in 25 languages if you have not proven that localization moves the needle for that specific content piece and initiative.”
MT for Usability Content
In the Q&A portion of the panel, Tarditi and Chang turned their attention to the question of machine translation (MT). How, if at all, should MT be employed in SaaS localization?
At Asana, workflows are currently human-based and localization priorities are focused on data and research, Chang said.
At Qualtrics, the largest application of MT has been on support content. “We’ve chosen an MT engine that was trained specifically on support content and we’ve seen incredible results,” Tarditi said.
She added, “We made the decision not only for cost reasons but for scale. “We wanted to get more content out there faster.”
Tarditi drew a broad distinction between content for usability (such as support pages) and content for engagement. “Where content is for usability, the applications are endless,” she said.