This iteration included: the creation of a basic CMS, three weeks of piloting (starting with paper piloting), and one week of 'public' episodes. The public episodes were delivered in a variety of ways to an focus group setup and surveyed by the product team in Raleigh. The advertising team wrote test copy, for an actual advertiser.
What follows is a day-by-day recap of the 'public' week, with notes about what we were testing and highlights from the audience feedback.
MONDAY 9/11 |
PLAN:Editorial: Our internal piloting lead us to a 4-5 item format, with a utilitarian editorial approach and a friendly tone.Advertising: 5 Second "Brought to you by" sponsorship note leads the report. Based on RealTime Feedback: -na- Duration: 3m 06s Avg Segment Length: 44s |
PRODUCT: |
FEEDBACK:A summary of the days feedback... |
|
![]() |
"No one likes ads, and I wasn't a fan... ...Perhaps it could be put at the end... ...Or it could be made as short as 'brought to you by lazyboy (sic) couches'." |
|
"After the Sac bee updates I got NPR updates. Which is great, I like NPR for national news. Those could mean I would only want local news from Sac bee." |
TUESDAY 9/12 |
PLAN:Editorial: Single national storyAdvertising: 15 second sponsorship note, with a call to action leads the report. Based on RealTime Feedback: -na- Duration: 3m 06s Avg Segment Length: 44s |
|
FEEDBACK:A summary of the days feedback... |
|
|
"The ad... ...was even worse and longer than yesterday, I was very tempted to close the tab." |
|
"Don't have stuff for sponsors. Don't want to hear... (ad) ...placements in news." |
|
Mr. D. liked the "reminder that more info is available online" and said, "a little more depth would be cool." |
|
"Maybe if you started off the content stating what topics would be covered (like today we will cover local, news, politics and weather) so I know what to expect and not to be surprised if it ends early." |
|
"I was a little concerned about the segment that reported 10 deaths being attributed to Hurricane Irma... ...There were numerous deaths in the Caribbean and Cuba." |
WEDNESDAY 9/12 |
PLAN:Editorial: No national storiesAdvertising: 23 second ad with call-to-action at end of 2nd story. Based on RealTime Feedback: Tweak language to emphasize ties to text articles. Duration: 3m 19s Avg Segment Length: 50s |
|
|
FEEDBACK:A summary of the days feedback... |
|
|
|
"There was a weird echo a couple times." |
|
|
"This morning I tried listening to my news flash on a Fire Stick. For some reason the Fire Stick said there Sac bee update is not available. But the NPR update worked fine." |
|
|
"More recordings! And shorter ads." |
|
|
"I would want more than four briefs. " |
THURSDAY 9/12 |
PLAN:Editorial: With the exception of weather and the overview, the stories were not rewritten for an audio format, but read as the appear online. We also doubled the duration of each story, and included a 5th block. We did this in an effort to emulate what Spoken Layer offers.Advertising: 26 second rebranding ad at the end of the 4th story. Based on RealTime Feedback: We took one user's suggestion and lead with an tease of the upcoming stories. Duration: 8m 19s Avg Segment Length: 1m 28s |
|
|
FEEDBACK:A summary of the days feedback... |
|
|
|
"Nice job adding the "ad free" option." |
|
|
"I think i like the longer story format - although it could get tedious if I have to ask to skip too often... ...It could be interesting to get short updates on sports, concerts, business updates, other events happening in the area." |
|
|
"I liked the intro which briefly listed the topics that will be covered... ...I thought it felt a little too long... ...made me wander/only half-listen more." |
FRIDAY 9/12 |
PLAN:Editorial: We doubled down on the production effort for Friday, cramming 18 story summaries into the update and breaking them into topic blocks.Advertising: 23 second ad at the end of 5th block. This was followed by a plug for a theoretical subscription based ad-free version of the service. Based on RealTime Feedback: The decision ad more stories was influenced by user feedback, we also changed the recording setup to reduce echo. Duration: 6m 34s Avg Segment Length: 1m 16s |
|
|
FEEDBACK:A summary of the days feedback... |
|
|
|
"The lazyboy (sic) ad at the end wasn't too bad either, I didn't listen to all of it though." |
|
![]() |
"I really liked the inclusion of local events and sports... ...Overall I like this service and I hope SacBee continues to offer it... ...I suggest that the reader warn listeners that the content may be disturbing... ...and take the opportunity to let listeners know they do have the ability to skip." |
|
![]() |
"I liked that it ended with events/things to do (perfect for the weekend) and then finished with the advertisement. I honestly think the advertisement is best at the end because it doesn't disrupt the short recording but you still listen to it at the end/is more tasteful and makes sense in that placement." |
Finished V1 Prototype, tested via ECHOSIM.IO (A browser based echo emulator)
We Accomplished this, using a PHP script that reads McClatchyDC's 'all news' RSS feed.
The original RSS feed isn't sufficient because it uses a single descriptive sentence and a raw URL in the description fields. Alexa will read whatever is in this field. So the immediate goal of the script is to overwrite that field with more suitable text. In this iteration we're using the first 4 paragraphs of story, which we need to clean of special characters.
For details see the script here:
Here's a look at the output:
Then, in the Alexa dashboard we point the "news briefing" skill at the PHP file:

If continuing on the Speech To Text front:
If pivoting to human read:
Davin had the idea of adding bylines.
Ultimately our feeling, following conversations with Amazon and other stakeholders is that the product requires a more human touch. For starters a recorded voice.