Connected Devices Weekly Update/2016-11-17

This week's update

• Engineering team on track towards doing Sprint 1 demo Friday, Nov 18:
  1. Firmware team: will read data from temperature sensor and transmit it to a test server
  2. Cloud team: to show api implementation and show logs of a sample Sensorthings interaction
  3. Client team: state of the web implementation with wireframe options
• Continued discussion with legal on documenting SensorWeb Data Collection on Mobile app and Website.
• Initial Visual design with exploration look and feel on App & Website. Refer to SensorWeb MVP Design doc under App & Website section

Upcoming Tasks

• Nov 28 - Share out of findings and Next Steps of User Study #1 in Taipei
• Mid/Late Nov - Running Pricing sensitivity study in Taipei. Results to share week of Nov 21st.
• Dec 16 - Perform Rapid Risk Analysis (RRA) review with Enterprise Security

Software Schedule - Dec 16 Milestone:

• Firmware: Collect real data from sensors, transmit to cloud

• Cloud: SensorThings API, receive and store sensor readings

• Web: Ability to browse collected sensor readings

Hardware Schedule

• Nov 22 Working demos on demo boards
• Nov 23 Printed Circuit Board Assembly (PCBA) ready for test
• Dec 01 Sample boards ready for Mozilla

• Voicemail study currently in progress. To complete this week
  • User interviews next week and findings due Nov 29th
• Team working on Product Specifications document
• Proof of Concept in progress
  • Completed voice and emoji transfer over WebRTC Data Channel
  • Completed capture and playback of audio
  • Prototype hardware under evaluation
  • Cross compilation of WebRTC over Beaglebone in progress

Updates

• Monthly Leadership review done on 11/16. Discussed progress in
  • Finding our platform story
  • Design/architecture around WoT
  • Extraction process
  • Challenges/mitigation
• Feedback being worked on (e.g. Deeper dive on security, data ownerships, Designing with a platform experience in mind, avoiding a headless platform, exploring partnerships etc)

Hawaii work week targets

• Finalize Design Paper: for wider review in Mozilla
• Alignment between Platform & Track1 products: Around architecture & extraction process
• Developer story: Track1 developer facing contributions (tie to developer events in 2017)
• 2017 goal planning

• Engineering team had a VERY successful work week
  • Completed the first working prototype of DeepSpeech last week. (for more info, please reach out to Gregor)
  • Discussed OKRs for 2017
• doyouunderstand.me is pulling in real data now but still a WIP
• Continuing to implement the English language model
  • this will drastically lower our WER
• Goal planning for 2017
• Blog post for the community

Project VoiceBank

• open source speech corpora
• Working on revamping the UI of the web app where we could be collecting voices
• Planning on kicking off in Hawaii with stickers!

Concept Validation Testing

• Data from the study to test alternate Abigail product concepts is in the process of being analyzed along with the results from the Facebook campaign.

Engineering

• Working this week on minor bug fixing for the reminder patterns.
• UX consultant to work with the team to define high level user experience. Collaboration is to begin the week of November 28th.

Abigail Study with Prototypes

• Testing three scenarios:
  • Setting reminders for yourself & other members of the family
  • Alerts notifying the users of activity on a door/window
  • Alerts indicative of when your loved one is home
• Study includes consumer interaction with the prototype and collection of feedback regarding usage and responses to questions asked of the consumers.

• 5 sessions have been held so far during the week of November 13th, with another 4-5 to be completed by the end of the week.
• Response has been very positive.

Schedule

• Value proposition validation - Oct. 24 - early November
• Prototype development: month of October
• Assembly of devices: October 31 - November 12
• Bug fixing: October 31 - November 12
• Testing in homes: November 13 - 18
• Data synthesis from pilot and video/landing page studies: November 21 - 29
• Gate 1 pitch: December 21

Experiments/User Studies

• FB Landing Page campaign to drive downloads of the app and validate interest: finished a series of experiments that confirms that the problem of people wanting to rediscover their city or discover hidden secrets exists, but we did not achieve high levels of conversion (registrations or downloads) from the landing page.
  • This could either be poor execution on the Landing Page with messages that don't connect to end users, or just lack of interest in getting an app for this problem but not qualitative data around.
  • We need to go deeper and look for other options as the twitter campaign and also get information through user interviews that may offer better insights.
• User Validation for 10 people in SF office on the week commencing Nov 28th is ongoing. This study, based on a mock-up, will provide us information to further validate the main assumption “The Magnet experience attracts users to try the app and return frequently”.
  • Wilson will be with Francis at the SF office to work on the mock up protype, it will also be a good opportunity to work with David on the product definition.

Development

• This week will finish our third sprint (Sprint 8) that it's the last one before uploading the client to the stores on Nov 23th (London Pilot starts on Nov 30th)
  • Complete test Run will be executed on Nov 22th, and next week we will focus on the bugfixing
  • This week we have been refining the way to scan and notify users based on their location, POI Deep link from the notification, navigation in the app, call Google Analytics from native code and other tasks in order to have the client in the best shape.
  • Updated Platform extraction document
  • Example of Real Dogfooding, map showing when Magnet ios app was scanning for locations from Picadilly Circus to Mozilla London office (scanning every 50 meters approx and a scanning radius of 50 meters)

Other

• High level planning and Roadmap post-Gate 1
• Last week we saw in our metrics that Magnet has been used to discover URLs during the Chromefest for Polymer
• Interesting read about why BT will not be an obstacle for beacons

Landing Page Experiment

• Our seven day landing page experiment launched over the weekend, finishes tomorrow. Involves Twitter ads, landing page and a survey.
• Measuring relative performance of four Twitter ads themed around four aspects of our value proposition: cost; privacy; control and expandability.
• Out of a total of 197,738 impressions so far, the current rankings for most number of link clicks are:
  • 1. Cost
  • 2. Control
  • 3. Privacy
  • 4. Expandability
• Landing page conversion rate has consistently stayed above 30% for the whole week (peaked at 41.3%, currently 31.23%). We've had 1028 visitors so far.
• 110 completed responses to the survey so far. Results are looking very promising, currently strongly supporting our hypotheses.
• Porfirio working on a blog post about the experiment
• We've welcomed Katie to the team who will be helping out on UX

Landing page conversion rate:

Technical Feasibility Study

• Juan has push-button pairing working between our gateway and ZigBee sensors (see video demo). (We now have both WPS and ZigBee pairing working using the same hardware button on the gateway).
• Juan ordering sensors from other manufacturers to test interoperability
• Tamara has a video stream from a WiFi camera (via a standard ONVIF API) saving onto the gateway
• Video streaming (without audio) also works but currently the need to transcode audio into a codec that browsers support is causing issues. The gateway doesn't appear to be powerful enough to to handle real-time audio transcoding using FFMPEG.

Timeline

We are due to present our results to the board on 30th November.

Engineering

• Lost a couple weeks' development effort waiting on the purchase of a 3D printer. Haven't found an EE resource yet but now have a lead. Most likely won't be enough time to get up to speed on using the printer and figuring out how to print a PCB for the game boards to be used in the Hawaii All Hands studies.
• Will instead go with the backup plan of building the game base a different way. The board will not be as big as what we’d hoped. Trade-offs in how we’ll be able to position the LEDs. Working to figure out the size of the board. Still plan to build 2 boards for use in the studies being conducted with children at the Hawaii All Hands, but not much time remains to get them done.
• Test plan has been developed.

• Looking at other games that can be used with the prototype. Evaluating how checkers might be adapted and creating a new game that's based on a classic children's game.

Consumer Studies

• A member of the UX team has joined the Game ON! team and will help with the user studies in Hawaii, looking at the user experience in terms of how well it aligns with the value proposition.
• Team received advice from the research team on development of the study.

User Studies

• Completed the shopping studies to test the Lighthouse prototype with 7 study participants who are blind. Participants took photos of a variety of items in the store and recorded descriptions of the items as tags that were stored with the images.
• Image matching sessions followed. Photos were taken of the purchased items to determine how well the device could match the new photos to the original ones and play the recorded descriptions. Results were shared with the engineering team so that continued improvements could be made to the algorithm and settings.
• Response to the device by the study participants continued to be extremely positive.
• Product study with 25 participants at the SCVBC is scheduled for Friday, November 18th to gauge reaction to and to gain impressions of various form factors.

Engineering

• Engineers watched recordings of the prototype testing that occurred during the shopping trips and discussions continued around the feedback received during this testing.
• Team is planning for the Hawaii Work Week.

Schedule

• Complete development of prototype: October
• Assemble prototypes: October 31 - November 4
• Prototype pilot: November 8 - 15
• Form factor study: November 18
• Data synthesis from pilot and video/landing page studies: November 21 - 29
• Gate 1 pitch: December 8