How Translation Can Improve Accuracy in Market Research

If you’re planning to explore a foreign market or you want to reach out to customers who are non-native English speakers, translating your research materials can bring a lot of benefits. Using professional translators is more accurate than Google Translate. It can also help with market research in the form of data layout and analysis. Professional translation can even help with marketing tests.

Professional Translation Helps with Accuracy Overall

While there has been much in the news about the wonders of machine translation, it still doesn’t quite perform at the level needed for professional uses, especially for market research. For instance, according to Synced, even the most advanced neural machine translation can instantly translate texts with 60-90% accuracy, as of 2020. Accuracy tends to vary based on text length and language. These programs can often translate sentence-length structure with a fair degree of accuracy and sentences can be used as models. However, paragraph-length or full document texts tend to look more inaccurate because programs handle each sentence individually, which can lead to differently translated words in the same document and a lack of coherent flow.    

That being said, Google Translate is getting more accurate. Human assessors evaluated Google Translate and rated it on a scale from 0 to 6. Google Translate got an average score of 5.43 out of 6. The score between languages tended to be slightly different. It got a 5.43 for English to Spanish and a 4.3 for Chinese to English. While impressive, that still leaves a margin of error that could mean misrepresented data during marketing research.    

To see how that margin of error plays out in the real world, there was a study conducted that looked at 100 sets of real ER patient discharge instructions. These were translated with Google Translate. The sets contained 647 sentences in total. 

The results? Well, 594 of those sets of instructions were accurately translated to Spanish, while 522 were accurately translated to Chinese. That left 53 inaccurate Spanish sentences and 125 inaccurate Chinese ones. 15 of the Spanish sentences and 50 of the Chinese sentences had the potential to cause significant harm to the patient. 

If that’s what’s going on with patient instructions, imagine how market research could be misrepresented. That’s why it’s important to make quality a key consideration in your market research materials, as part of a comprehensive localisation strategy that focuses on quality from every angle. That strategy needs to define how your business will connect with audiences on both a global and a local basis. It should consider the customer experience at every stage, mapping out what you need to change and achieve in order to excel in every market. Relying on human localisation experts, rather than machines, is non-negotiable.   

Making Sense of Foreign Data

Digital marketing agency WordStream covered some case study examples of data-driven marketing. Below we’ll look at how each of these might need to adjust if that data was coming from foreign markets. 

• Using demographic data in your campaign planning: A lawn-care company separated price-conscious zip codes and ran an AdWords campaign aimed at those areas that showcased a $20 price point. They saw a 200% lift in click-through rates and a 30% lift for on-page conversion.  

If you run an ad campaign in a foreign area, this concept becomes far more complex. There is the language barrier to contend with, plus you might not immediately know which areas have certain income levels or consumer preferences. You’ll need to factor this into your market research. 

• Using trends from one marketing channel to inform another one: A gig marketplace owner used PCC data to inform his SEO keywords.    

One of the biggest areas of big data is information coming in from digital campaigns. There are often thousands of data points and, like all big data, the data set is ever-growing. This is something that you need to factor into your planning carefully in order to ensure you can make sense of these data sets when they come from foreign markets.  

• Measure success incrementally: A digital marketer collected data from every step of the process the user of the sales funnel took. By making the process easier where people were leaving the funnel, they saw a 135% rise in conversion rates.   

Adapting this to foreign markets can potentially show you why certain parts of sales funnels don’t work based on local consumer habits.  

• Use PPC and email lists: One marketer looked at who was clicking which links in emails and used that to make follow-up ad campaigns on Facebook.

Because those emails and Facebook campaigns might need to be in another language in a foreign market, you’ll need to ensure you move smoothly between cross-informative data and campaigns like the one mentioned above. 

Understanding consumer habits in foreign markets is key to all of this. Doing so will enable you to tailor the way you use data in order to achieve your goals more effectively.

Marketing Tests

One of the best ways to make sure your marketing efforts are getting off to a good start is to perform marketing tests. According to HubSpot, only 17% of marketers use landing page A/B tests for improving conversion rates. And companies that A/B test every email see 37% higher marketing returns. 

If the data is coming from foreign markets, it’s important to test how a market or service might perform in a new location, or to see how a campaign is going overseas. You might want to examine the data points themselves and help lay them out in formats that are easier to analyse. Doing so can enhance your use of the data in multiple ways.