Share This

• 
• 
• 
• 
• 
• 
• 
• 

Today’s post is about Hungarians startup Easyling, which uses crowds to translate websites with a lightening speed. Its founders Peter Farago and  Balázs Benedek have been running a telemarketing and transcription crowdsourcing platform Skawa, employing a “modest” number of operators: 17’000. Easyling, its offshot is a crowdsourced translation platform, which made it to the finals at the How to Web startup competition.

As I am in Bucharest this week attending the conference , I thought I’d share a post about the company, which business I find fascinating.

Skawa allows big chunks of work, such as video or audio transcription, or simple telemarketing tasks, conducted in a record time. There are three important elements in this business. First is the crowd, comprised of Hungarian stay-at-home mums and disabled people, who work from home on a part-time basis. Second is the technology, which allows to break down large chunks of work (be it audio files that need transcribing or a list of leads to call) into small packets of work, which are then picked up by many of the 17’000 workers for an immediate execution. Predictably, the crowd completes the work in record time.

Time is money, and crowd offers competitive prices, so a lot of affordable value is created right there. But the third, most crucial element, is the quality control. With that many workers to check, Farago and Benedek rely on automatic tests before they add agents to the system, but they also charge extra for quality checks.

Quality is what remains one of the major issues in making crowdsourcing a viable business. If you want to build and manage your own crowd, it is also possible with Skawa, but time overhead will be significant, (for example, using the contractor marketplaces such as Vworker). And we shouldn’t even get to the quality control issues from Vworker-type crowdsourcing, with which I have had a mixed experience.

This testing system is at the core of Easyling as well, which aims to provide an inexpensive but better quality alternative to an automatic google translation widget, enabling website owners address their audience in a great (they say, unlimited) variety of languages.

Easyling offers to connect website owners with freelance translators and translating agencies. Many translators will get to work on your website, completing the job literally overnight. The crowd element brings the cost of the translation down, whilst the speed of execution adds value and enables the service to charge a premium.

The Easyling business model is a combination of a modest translation service fee (which doubles if performed by a prossional translating agency, quality indeed comes at a price) and price per thousand impressions. Professional agencies also pay to get Easyling business. 

As one advanage of Easyling is speed of execution, I thought it it would be valuable for news sites aiming to publish on other languages. They would require translation to be done within the same day, if not within an hour. But translating a blog into multiple languages would probably be costly. 500 words per day will result in Euro 5000-10 000 per year. In addition the cost per 1000 pageviews has to be balanced with incoming advertising revenue. Easyling plans to charge Euro 1 per 1000 pageviews, whilst you can earn between Euro 0.03 to Euro 3.5 per 1000 impressions with Google ads. So there maybe enough reasons for an ad-based business to go for such an arrangement. 

At the same time the model will work best for businesses looking to sell internationally and earn significantly more per 1000 pageviews,  while having less content to translate.  In this case Easyling’s speed of service becomes less important than a simultaneous translation to many languages, but it is still a cool option to have.

At How To Web I will keep the fingers crossed for Farago, who has already won a recent startup contest with 4Issues at the Startup Live Budapest. I will also look into making GoalEurope available in other European languages.