Subscribe
Sign in
RefineAI — Audience insights with the power of AI
RefineAI is an audience insights platform for videos.
It uses emotion recognition AI to understand the Engagement, Emotional Appeal, and Actionable Feedback within minutes instead of weeks or months as with traditional focus groups.
Try with your own videos today on our platform. We would love to hear your feedback.
14
Replies
Best
Tom Uhlhorn
@bharatkrish We have tried testing this in Australia across multiple sites and everyone has received the same message (copied below).
It appears your internet connection is not stable and has caused the video to stop. Would you like to restart playback?
Report
Share
@thomasoscareric Sorry to hear that. I think I know what is going on. We are hosting the videos in AWS we need to enable the cdn/cache for Australia. Thank you for bringing this up. We are looking into it. I will message you when fixed.
@bharatkrish just so you know, we tested with some colleagues in Singapore as well so your AWS fix might need to include that if it's separate.
@thomasoscareric so this is what we found. Looks like it was not an issue with the streaming in your region... But there might have been an issue while uploading the video. The campaign is created but the video seems to have not uploaded well. Can you try creating another campaign and upload a new video? Please also send a link to the new uploaded video. My email: b@refineai.com. Thank you!
Hello Product Hunters,
Thank you for taking the time to check out RefineAI’s Audience Insights product.
During my long stint in the Media industry one thing that we were not measuring well was how do our customer feel about our content. 50% of the content created ended up not resonating well with the intended audience. We spent millions on Market Research. But the results were rarely actionable as the content was already created.
The idea for audience insights gathering using computer vision and machine learning was born from the notion that we should be able to build and test video content rapidly. Similar to how software is developed and tested iteratively today.
During the past year we have worked with global companies such as Red Bull, Beiersdorf, and SAP and have tuned our platform for the world to use. Today we are proud to announce that RefineAI is ready for a public beta! We hope you take the time to try it out with your videos. Please provide feedback on everything as we want to improve our product for you.
Here’s a list of some of our use-cases:
- A/B test your ad, promo, product intro videos.
- Test your story line/arc with your target audience segment and prove your hypothesis.
- Understand your ideal audience segment for your videos by testing with a generic group.
- Test your video content to make sure it is culturally sensitive.
I would love to hear from you if you have other use-cases you would like to explore. How would you use the service?
I hope you enjoy it.
--Bharat
@aaronoleary Thanks for checking this out. At its core its a computer vision and deep learning model that takes the video of the tester and predicts the emotions evoked. The service then correlates this information with the video frames of the content shown.
This provides the quantitative feedback including the emotion and engagement scores. We then ask them to rate the sound quality, storyline and video quality. We also ask open ended questions to understand how they felt in their own words towards the content. This provides the qualitative feedback.
Please let us know if you have any other feedback.
Congrats - testing emotional reactions is something we had to manually do with human facilitators before. This is a great use of ML to automate, scale and standardise this form of market research and user feedback
@black_bear We developed our own emotion recognition models to understand faces and expressions in video. In the future we would like to combine that with text sentiment analysis to understand the qualitative feedback (ie. the questions we ask after the video is watched).
