Barometer teams up with Speechmatics to drive accuracy and inclusion in podcast intelligence

A strategic partnership focused on accuracy and inclusion

Today we are thrilled to announce that Barometer has selected Speechmatics as their speech technology partner.

Barometer, known for delivering AI-driven insights to help business understand communication strategies, has teamed up with Speechmatics to help unlock the value of the human voice.

At Barometer, inclusion and accuracy are core pillars, and understanding all voices, regardless of accent, dialect or demographic. This overlap with Speechmatics' own mission to 'Understand Every Voice' makes for a perfect partnership with an aligned focus.

As a founder and member of the MOSAIC Collective, Barometer strives to provide a platform for diverse voices in audio. Additionally, their membership in the ALIVE Network supports black creatives in the media space through the ALIVE podcast network.

Promoting inclusivity across all media forms, Barometer will leverage Speechmatics speech technology to expand across geographies, enhancing both the accuracy and inclusivity of the platform.

Lauren King, Chief Marketing Officer at Speechmatics commented, “We are thrilled to partner with Barometer, a company that shares our commitment to accuracy and inclusion. This collaboration will empower us to bring our advanced speech recognition technology to even more diverse voices around the world. Speechmatics is excited to help support the bold vision that founder and CEO, Tamara Zubatiy, has in transforming the podcast industry”.

Garbage in, garbage out

Despite the emergence of new transcription vendors fueled by the rise of Large Language Models, Barometer selected Speechmatics for its proprietary dataset, trained on millions of hours of manually transcribed text.

This unique advantage places Speechmatics ahead of competitors based on non-proprietary, open-sourced data.

As the adage goes, "garbage in, garbage out."

Higher transcription quality leads to fewer errors, and, consequently, more accurate analysis. Speechmatics achieve consistent, reliable and inclusive transcription - regardless of dialect.

"The decision was influenced by several factors," explained the Barometer team.

"We found that Speechmatics' transcription had a lower word error rate compared to other suppliers. Specifically, we considered the accuracy of transcriptions involving voices with accents, older adult voices, and voices from diverse backgrounds. Fewer errors result in better analysis, allowing Barometer to more accurately interpret the context of content."

Another significant factor in the decision was Speechmatics' international focus. Their support for a variety of languages, including the first bilingual model for Spanglish, aligns with Barometer's diverse linguistic needs.

This consideration is essential for the nuanced and individual world of podcasting, where realistic speech and diverse voices must be accurately captured.

Grant Nelson, Barometer's CTO, remarked, "This partnership allows Barometer to continue delivering best-in-class brand suitability and contextual targeting data for the podcast advertising ecosystem." CEO Tamara Zubatiy added, "It also supports our expansion beyond audio into new media channels, including YouTube, digital, and more."