San Jose State University

San Jose State University

Audience Acquisition via Smart Speakers

San Jose State University is one of the 2017 winners of the Challenge Fund for Innovation in Journalism Education. See all 10 winners.

Team

  • Halima Kazem, Adjunct Professor, School of Journalism and Mass Communications, San Jose State University
  • Julie Makinen, Founder and Editor-in-chief, California One
  • Vignesh Ramachandran, Head Editor for Audio, California One; Producer, Pro Publica Illinois
  • Kathleen Masterson, Journalist, NPR
  • David Gross, App Developer

Describe your project

We will prototype best practices for making news briefings and news-related “skills” for new audio and video voice-enabled home assistant devices — also known as smart speakers — and see how such content may be used to build audiences for a new media platform. Students will create briefings and skills for devices including Google Home, Amazon Alexa, Amazon Echo Show and Apple HomePod, and publish to these new platforms as well as other channels (web, Facebook, newsletters) to see how audiences use and interact with them. Results will be used to determine how such platforms can best help news outlets acquire new users.

What is your live news experiment?

Students will learn fundamentals for producing content for digital home assistants based on platform developer kits such as Amazon’s Alexa Skills Kit — a collection of self-service APIs, tools, documentation, and code samples that allows journalists and others to add briefings and skills to Alexa devices.

Students will initially develop a daily workflow for one-minute audio news roundups or “flash briefings,” determining the best content, publication schedule, technical issues and more. In the second phase, students will produce subject-specific briefings on topics such as politics or the environment, or “skills” relevant to California news and civic engagement. In the third phase, best practices from audio may be applied to creating a workflow for video briefings for new devices such as the Alexa Echo Show.

Students will experiment with creating two-way conversations with the audience via smart speakers, seeking comments or questions from listeners who hear the daily flash briefings and want more information.

Students will publish briefings and skills under the umbrella of California One and measure audience acquisition. Students and researchers will collect metrics from Amazon and other platforms, including data on the number of users, frequency of usage, peak usage time, duration of usage/engagement, and utterance distribution.

If the experiment works, what do you think might happen?

This experiment would have a number of positive outcomes, including:

  • Audience acquisition for California One and possible future internship or employment opportunities for students involved in the experiment
  • Creation of new models of audience engagement through smart speakers, including possible two-way audio conversations between listeners and journalists
  • Development of new curricula at SJSU devoted to teaching skills for cutting-edge media devices
  • Expansion of this curricula to other journalism programs in the California State University system or beyond
  • Creation of audio and video content for new platforms by student journalists for university-based publications
  • Dissemination of analytic report on best practices on using smart speakers for audience acquisition, benefitting media organizations nationwide.

California One, SJSU students and researchers will track audience acquisition to determine how effective content for these new platforms is at building audiences for media startups.

What does success look like? We’ve set these goals for the experiment:

  • Acquisition of at least 5,000 unique users of flash briefings or skills on smart speakers over course of experiment.
  • Students successfully innovate a two-way dialogue with audience via smart speakers.
  • Content created for audio assistant devices can be monetized within the smart speaker ecosystem or via means such as newsletters, Facebook or California One’s website.

How is this project unique and innovative?

Audio content is booming and California is at the forefront of smart speaker adoption. However, there is little good content being created for and published on these cutting-edge devices. Despite having a large role in developing these devices, Californians are left to find local news via traditional means like print, radio and television.

We surveyed existing news content on the Amazon Echo and the vast majority of offerings are repurposed stories from print or recordings of television scripts simply pushed to Alexa, creating a subpar user experience as evidenced by poor ratings and customer reviews. California One would produce briefings that are native to voice platforms. We will experiment with what makes sense for a 2017 news consumer who is engaging with voice interfaces in a connected home.

It may be possible to repurpose content created for smart speaker systems for dissemination via email, on the California One website, in podcast form or on social media platforms. Our experiment will investigate these options and try to outline paths forward for efficient workflows for small newsrooms. If we are able to innovate two-way conversations between smart speaker users and news organizations, this could open up wide new paths for audience engagement.

What technology platforms will you use?

Voice interfaces are a uniquely natural user experience. As smart home devices proliferate and audio in cars innovates beyond terrestrial radio, California One will use emerging voice development platforms from Amazon, Google and Apple. (The project reduced its scope to just Amazon devices for ease of use.)

Newspapers found that it was not enough to put a PDF of their print product online and expect success. Likewise, TV news broadcasters have struggled to win viewership online with little innovation around platforms such as Facebook. For journalists, each new technology and platform demands creativity and experimentation to innovate and iterate best practices and take advantage of the unique modalities of the technology.

The barrier to entry for developing flash briefings via the Alexa Skills kit is low: We have already experimented with developing lo-fi examples of simple skills, proving the potential for a more complex news report.

If it works, how might this experiment change teaching at your school or media practices in your partner’s newsroom?

California One’s smart speaker briefings and skills, if proven popular, could easily lead to strategic changes in California One’s business plan, including a greater emphasis on creating content for these devices over other formats such as text. California One may also pursue underwriting or sponsorship for such content.

If this experiment is able to innovate two-way conversations between listeners and the newsroom, it would create pathways for story discovery, commenting and interactions between the newsroom and audiences. One could even envision creating entirely new jobs for journalists to listen to feedback, tips, comments and questions from smart speaker audiences and respond to them appropriately.

This project also could lead to a series of new classes for SJSU’s journalism program. It could be the impetus for the department to create a strong news radio and audio sequence, which it has yet to develop.


The San Jose State University team provided an update on its project in a March 2018 report.

Update: What have you discovered?

Producing an engaging news skill for smart speakers is not easy and requires a suite of skills: news judgment, understanding of how to write for audio ecosystem, understanding of how to edit and record to the specifications mandated by Amazon Alexa ecosystem, acquisition of data, programming skills, marketing prowess, etc.

The collaboration process with California 1 is giving our journalism students entrepreneurial skills such as designing products/apps, prototyping products and testing audiences. The process of learning these skills have led to students thinking about the larger context or news production and not just single news stories. Documentation of this learning process is allowing me to replicate this class in the future.

In addition to California 1, students are also working with local radio KALW and the San Jose State student news publications, including the Spartan Daily newspaper. In the case of KALW, for example, the station is supplying text and audio to students on a weekly basis to “seed” the California Speaks app/skill. In the case of California 1, content featured on the site is suggested to students for inclusion in the California 1 Alexa skill. In the case of SJSU student publications, content created by students is used as basis of local San Jose state news skill.