This Thursday, we will release original research of ours involving Fortune 1000 adoption of Amazon Alexa and Google Assistant.
To celebrate, we’re holding our first membership drive. This email features a button with a 50% discount on a paid annual subscription.
Paid subscribers get all letters - five a week, plus access to the archives and additional community features. (Free subscribers receive 1-2 a week.)
With 20,000 subscribers and growing, executives and technologists around the world wake up to This Week In Voice VIP - subscribe this week for the only discount we’ll offer this year until right before the Christmas season.
Para bailar la bamba
Para bailar la bamba se necesita una poca de gracia
Una poca de gracia pa' mi pa' ti y arriba y arriba
Ah y arriba y arriba por ti seré, por ti seré, por ti seré
It doesn’t happen often, but on occasion, we get a song that tops the charts in the United States despite not being in English.
Listeners acclimate to the song, and become fans, despite having only a minimal understanding of what message is actually being communicated.
Music is the rare universal language.
On eBay, items for sale often have long and wordy names.
And those long and wordy names are often accompanied by long and wordy descriptions.
In 2014, eBay rolled out something called eMT, which is a machine learning-based translation program which improved the translations of item titles and descriptions all across the massive website.
When items are translated from one language to another on eBay, they have to be approved by a panel overseeing translations on the site, in something eBay calls the Human Acceptance Rate.
This new machine learning program increased the number of items which passed this process from 82% to 90%, instantly increasing the amount of available surface area for commerce on the site.
Further, after this machine learning translation program went into effect, it was observed that for each additional translated word in an item’s title, trade volume went up an average of 1.06%.
Overall, ecommerce shot up 10.9% between pairs of countries impacted by this new translation capability.
When weighing the role that physical distance plays in ecommerce between paired countries, this is the equivalent of shrinking the world by 26%.
Incredible.
In the 2018 paper that was published in 2019 in the peer-reviewed academic journal Management Science, titled Does Machine Learning Affect International Trade? Evidence From A Large Digital Platform, researchers from MIT and Washington University in St. Louis discuss these findings.
The takeaway here is that while there are plenty of areas where AI and machine learning are discussed in ambitious, future-thinking terms, not enough discussion is happening about the areas where machine learning is creating powerful outcomes right now in various sectors and industries.
While this study focused specifically on one specific platform - eBay - the platform is large enough to provide enough data to causally link AI to process improvement and corresponding increased sales volume.
No complementary technologies were needed. No lengthy ramp-up was necessary. With eBay and machine learning, it was literally plug and play. Positive outcomes started from almost the moment the program was implemented.
This study should be and likely will be held up, moving forward, as a gold standard for what can happen when AI and machine learning are thoughtfully integrated into ongoing corporate operations.
Subscribe to This Week In Voice VIP and receive the same essential voice/AI information and analysis as 20,000 subscribers around the world.