Publishers tap into audio opportunities with synthetic voice, walking tours
Conference Blog | 11 June 2023
During a recent Product Innovation Master Class, INMA Product Initiative Lead Jodie Hopperton listed five arguments for audio growth: scale, young people, screen-free, accessibility, and intimacy.
“It’s pretty vast, and it’s something we’ve seen grow a huge amount,” Hopperton said.
Media leaders from Prisa Radio, San Francisco Chronicle, Stuff, and Wake World shared how they are tapping into this rapid growth to engage audiences through audio innovations around synthetic voice, walking tours, and podcasts.
Prisa Radio
Prisa Radio in Spain found great success in combining two big goals: utilising new technologies and capitalising on their captive soccer audience.
Olalla Novoa Ojea, voice coordinator for Prisa Radio, shared insights into the launch of Victoria, a synthetic voice created completely from scratch and now known as “The Voice of Soccer.”
After workshopping different ideas for voices, they decided on a female voice since the brands felt it was important to make females more visible in the sports world. They wanted Victoria to be between 40 and 45 years old since that age range spoke well to the target audience for many of their brands. Prisa began working with a Spanish startup that specialises in synthetic voices in Spanish.
“With them, we finished this whole design process creating a data set that reflected all the use cases for Victoria,” Novoa Ojea said. “What was Victoria going to say? And in which way was she going to say it? We needed to reflect that on more than 4,200 sentences.
“Those sentences also had to reflect different sound combinations because even when this is Spanish, as you may know in soccer there are many foreign players, many foreign terms, and we needed to make sure that Victoria could pronounce them.”
Prisa ended up creating a software-related dictionary with 3,000 terms for Victoria to use. That dictionary has grown to 9,000 terms today and it keeps expanding.
Once the data was ready, they had to record the voice. They wanted an original voice, which included training the AI model with four different female training voices. It took Prisa a while to record all the sentences to train the model.
“The latest model needed 350,000 iterations till we were happy with the voice we got,” Novoa Ojea said.
San Francisco Chronicle
The San Francisco Chronicle had a robust live event series pre-pandemic, which was quickly adapted to Zoom as lockdowns began.
Facing challenges around screen fatigue, Sarah Feldberg, editor of emerging product and audio, said the team began thinking about ways they could “connect with people through immersive experiences to give them something new — something different than more time spent in front of a screen.”
The genesis of the audio walking tours came from the Chronicle’s architecture and urban design critic, John King. And, at first, Feldberg thought that would be a miserable product experience since people would need to be constantly fiddling with a phone, pressing play, pressing pause as they’re walking around.
But that original idea was the kernel of something interesting to Feldberg. She wanted to take the intimacy of a podcast and couple it with GPS locations to make it reflect what people were seeing in front of them as they walked around.
The GPS-guided walking tours are written and narrated by Chronicle journalists. Each piece of audio is tied to a physical location through GPS locations, so “tours are routed in the built environment — they’re physical experiences from our journalists.” You can’t do the tour virtually,” she said, “people have to interact with the environment” for the tours to work.
Feedback has been incredibly positive, both inside the newsroom and from people who have taken the tours. The tours act as a retention benefit for subscribers, reinforcing the Chronicle’s authority while demonstrating an ability to deliver new things. It’s also a way to introduce a new audience to the Chronicle in an interesting way.
“Our tours are the most-reviewed and highest-rated San Francisco tours in the VoiceMap platform,” Feldberg said, “and all the journalists involved in the first three tours have already pitched ideas for more.”
We often multitask when we listen to regular podcasts, but you can’t do that with these tours. They require your full attention, Feldberg said: “The superpower of these tours is that connection.”
Stuff
Long before most news media companies were adding podcasts to their offerings, Stuff in New Zealand was already well-known for its audio.
Nadia Tolich, chief audio officer, said the company launched audio vertical six years ago with the Black Hands podcast, which chronicled a decades-old murder case that had gripped New Zealand.
A Stuff journalist had written a book about the case and the now-chief content officer suggested turning it into a podcast. To date, it has had 7 million downloads and is counted among the top 1% of all podcasts worldwide.
As news media companies look to the future, Tolich noted that audio is not just about podcasts; it’s also about things like text-to-speech. There are many possibilities, and Stuff is among the companies experimenting with audio in lots of different ways as well as looking at how to deliver it.
“There’s lots of different things that we’re looking at to really improve functionality and user experience,” she said. “That is our priority. Not just podcasts, even though that’s what everybody thinks about. I also think that people want more news in short form. So I think that that’s going to be a huge part of what we do moving forward.”
Along the way, Stuff will improve things like monetisation models, data and insights, and integration opportunities with clients.
“Ultimately there’s lots of things that we want to be doing,” Tolich said. “I know we’re going to get there. It’s an exciting place to be.”
Wake Word
As chief operating officer and co-founder of German podcast production company Wake Word, Ruben Schulze-Fröhlich is in the business of helping podcasts reach their audience.
Guided by a tool called Podius that the company developed, Schulze-Fröhlich is able to provide clients with deep insights into the data beyond just the number of downloads. Using data-led practices has helped the company grow its inventory and attract three main customers: marketers, media agencies, and news media publishers.
This not only provides Schulze-Fröhlich with an understanding of German-produced podcasts, but has created a global view of what people are listening to, what works, and how companies can leverage podcasts both as a vehicle for growing audiences but for attracting advertisers.
Schulze-Fröhlich shared best practices for publishers who are looking for success in the audio space, including some about the importance of timing.
• Know the right duration. In Germany, the average duration of the podcast is 25 minutes. News programmes are shorter and religious episodes are longer, so it’s important to know what the norm for your genre is and make sure you’re staying in line with that.
• Know how to place your ads. This means evaluating what’s standard for your market. For example, in Germany, the first ad is usually placed after about 30% of the podcast has played and the second appears around the 75% mark. “The ads are quite long, 60 to 90 seconds,” he said. “They are mostly host-read and almost always contain a personal story matching the client’s product.”