Case Study

Fast and accurate video captions enhance customer experience

The Need for Speed

These days, just about everyone is a content creator.

Suffice it to say that smartphone adoption has reached cruising altitude, putting a high-quality camera in just about every pocket. Content remains king, but now more than ever it’s also a commodity; for creators in an increasingly crowded market, differentiating that content is imperative. Passive posting to text-based social media is necessary but insufficient to build durable, deep relationships with fans and followers. When it comes to personally engaging, relatable content, video is where it’s at.

On a daily basis, the average person spends more than 2.5 hours a day across an average of 7 social media accounts, and video is the most sharable and engaging medium. As a result, video (especially short form) has become central to the marketing strategies of virtually every business type, both big and small, with 87% of marketers saying video marketing has helped them drive more sales. This is where BIGVU enters the chat. BIGVU has created a comprehensive, end-to-end video creation platform that allows ordinary people to easily create and distribute speaker-driven video content right from their mobile phone or personal computer. Leveraging AI throughout, people use BIGVU’s tools to create more than 2 million recorded videos each month.

From scripting to shooting, from styling to sharing, BIGVU helps small and medium-sized businesses build their digital presence and drive engagement with their customers with richly creative video each and every day. With accessible tools for everyone, BIGVU offers a number of features their users have grown to depend on.

For example, BIGVU automatically generates captions that are perfectly synchronized with the speaker’s narration. Beyond the obvious search engine discoverability benefits and the fundamentally important goal of making video content accessible to people who have trouble with hearing, data shows that captioned video is simply more engaging. Even folks with keen, pitch-perfect hearing can consume captioned video content where full-volume playback wouldn’t be acceptable, like on the subway or waiting in line at the DMV.

Deepgram: The Solution You Can Trust With Your Customers

Across every aspect of video content creation, time is money. Time spent on production is money out the door. Every second of viewer attention is potentially worth its weight in gold. That’s why BIGVU needed a transcription solution that moved as fast as its users. Beyond the need for sheer speed, BIGVU also required cost-effective batch transcription, at scale, to accommodate their users’ creative workflows.

Legacy transcription solutions and alternative providers presented the same issues: long wait-times for transcription results harmed the user experience, leading to lower customer satisfaction which, in turn, raised the risk of churn.

Deepgram is excited to help BIGVU deliver on the promise of a seamless user experience for captioned video production. Gone are the days of minutes-long transcription wait times; Deepgram delivers 10x faster video-synced transcription that arrives in seconds. What’s the secret sauce? Deepgram delivers category-defining ASR performance through its fine-tuned hybrid conformer model architecture and bespoke GPU compute infrastructure to run blazing-fast inference. Instead of finding the optimal tradeoff between speed, accuracy, and cost, Deepgram’s models are designed to optimize all three simultaneously.

After their preliminary testing completed, BIGVU’s developers successfully integrated the Deepgram API into their production app and measured a shocking 58% reduction in transcription latency. The net result has been a significant boost in their platform’s customer experience, user engagement, and retention rates. With Deepgram, BIGVU is able to deliver a world-class experience to its customers that is simply unattainable with alternative solutions.

I was blown away by the speed…It was definitely a game changer for us.

Yoav Reisner

Data Scientist