Utilising your customers’ reviews to increase global sales

As identified in our translation trends for 2020, another great method of communication for effective customer service and support is to encourage user-generated content.

User-generated content, or UGC, can range from comments, reviews, ratings, photos, entries in forums, reviews, testimonials, and other forms of content created directly by your end-users.

“User-generated content is the term to describe…forms of media that was created by consumers or end-users of an online system or service and is publically available to other consumers and end-users. User-generated content is also called consumer-generated media (CGM).”  Webopedia

This is one way that companies can utilise the content created by their target audiences in order to:

  1. get a better understanding of their existing customers,
  2. have their customers identify areas of improvement for their products, services, or company, and
  3. drive new business through product development, improved marketing efforts, and even localisation.

In this post, we’re going to be looking specifically at reviews and how reviews and ratings can impact a global organisation’s presence overseas. Let’s look at a couple of different examples

The first one is Booking.com. Booking.com has grown a huge global community of users over the years, and encourage UGC by incorporating rewards in the form of their Genius program for those customers who regularly contribute to the community with their reviews and property ratings.

Through this active community, Booking.com have also tapped into a vast amount of data that they can use to continually evaluate the overall satisfaction of their current customers, whilst helping potential future customers to make an informed decision based on other visitors’ experiences, and ultimately make a booking through their system.

Booking.com logo
  • 25 million destination reviews from real travellers (Booking.com)

  • available in 43 languages (Booking.com)

  • over 70 million visits a month (Similar Web)

90% of respondents who recalled reading online reviews claimed that positive online reviews influenced buying decisions, while 86% said buying decisions were influenced by negative online reviews.” Dimensional Research

TripAdvisor logo
  • 760 million reviews on TripAdvisor (Expanded Ramblings)

  • 490 million unique visitors (Expanded Ramblings)

  • 160 million traveller photos uploaded to TripAdvisor (Expanded Ramblings)

  • avaliable in 28 languages (TripAdvisor)

The second example is TripAdvisor. In a similar way, TripAdvisor is now one of the go-to sites for identifying potential restaurants hotels and other activities popular with tourists around the world.

Restaurant reviews are probably the most sought-after information on the TripAdvisor site and who created them? That’s right, the end-user.

So here we enter a slightly more challenging area of utilising user-generated content. How to make it appeal to a global audience.

In both scenarios Booking.com and TripAdvisor allow their customers to write reviews in their own language. This adds a layer of complexity when it comes to a global content strategy.

Due to the sheer amount of content that is produced daily by their users, both organisations have utilised machine translation and machine learning tools in order to be able to churn out such large volumes of content.

Machine Translation?”, I hear you say. Yes, that’s right.

It’s not all doom and gloom when it comes to machine translation. This is a perfect opportunity for large organisations to incorporate machine translation into their localisation strategy.

Of course, it’s likely that top-level pages, property descriptions and other areas like FAQs will go through professional human translation to ensure high quality, but user-generated content can be quickly and easily localised using AI.

Need help automating translation? Contact us

At the end of the day, if you are looking to book, a hotel, or restaurant, or any other kind of activity when travelling abroad, or in your home country, you only really need to get the gist of the message. A five-star rating supported by a few positive words, that although may not be syntactically or grammatically correct, will at least help you to make an educated decision as to whether to proceed with your booking.

As our existing clients know, we would always recommend a layer of post-editing if utilising machine translation tools. However, due to the purpose and goal of using machine translation for user-generated content, we could safely say that put no post-editing is required.

As long as the technology is integrated efficiently and can operate in an agile, self-sufficient manner, user-generated content can be automatically translated and published, therefore, reducing time to market, costs over time, and internal resourcing.

In order to cover a wide range of languages, the machine translation technology utilised must allow for a variety of language combinations. For example, you may find that in the case of Booking.com, a Turkish speaker is looking to book a hotel in Denmark. There are several reviews of the property from travellers who have previously stayed at the hotel who originate from Bolivia, Russia, France and China. Therefore, the machine translation engine needs to be able to localise the Spanish, Russian, French and Simplified Chinese reviews into Turkish.

This might sound rather straight-forward; input one language and another automatically pops out. Unfortunately, in this example at least, it is a bit more complicated. You see, not all language combinations are readily available in commercially available, pre-trained machine translation engines.

In research conducted by Inten.to in June 2019, they mapped a series of readily available stock machine translation engines and the most popular languages online to identify which engines were best suited for each language combination at that moment in time. The diagram below shows the mapping for the “best general-purpose MT engines” (Inten.to, 2019).

Based on the languages in our hypothetical example above, the red crosses show that in Inten.to’s research, from the 11 sampled engines, there were no readily available systems to automatically translate from Spanish, Russian, French and Simplified Chinese into Turkish.

Taking this into consideration, an alternative strategy for localising user-generated content would see all non-English languages first being machine translated into English before being processed into the relevant target languages.

In the case, we would see the Spanish, Russian, French and Simplified Chinese reviews first being localised into English, and then the English being translated into Turkish. This way might be a little bit easier to implement and manage, and you’re still going to be gathering a hell of a lot of data that can be used to continually train your engines moving forward.

Best general-purpose Machine Translation engines (Inten.to, 2019)

Depending on budget, you may have to consider a more available form of machine translation, such as off-the-shelf, generic engines like those provided by Google, Microsoft, Amazon, Yandex, the list goes on.

Alternatively, if you have a relatively large localisation budget, and an internal team, as in the case of Booking.com, there may be an opportunity for you to invest in your own machine translation technology and algorithms to get the most out of your multilingual content.

I hope this has been a useful introduction to how to handle the localisation of user-generated content. If you’re currently experiencing challenges when it comes to handling your user-generated content, and have questions related to the best way to incorporate this into your global content strategy, then do not hesitate to get in touch with a member of the Ultimate Team and we’ll be happy to discuss the options available to you.