Would you do this project? Would you use it?

I’ve been thinking recently that there’s a lot of untapped potential in the fact that the Internet can much better handle higher-bandwidth-sucking scenarios like video in recent years.  Vimeo was founded back in 2004.  Google bought YouTube in 2006, around the same time Amazon launched Unbox (now Instant Video).  Netflix started streaming Starz stuff in 2008.  Justin.tv tried a bunch of things and realized people really cared about watching other people play video games (now Twitch.tv).

But surely video can be used for more than business, gaming, Gilmore Girls and cat videos.

Skype, I think, is closest to my ideal application of video-based software.  It’s essentially a window into any location in the world, speaking any language.  They really need to get that thing going with HoloLens.  And I don’t mean a floating square with someone’s face on it in your living room.  I mean you can experience a stream of anyone else’s reality (they choose to share with you, not the U.S. government) and actually understand what’s going on regardless of what language is being spoken (or shown, like on signs).  That last one is not an original idea.

I have another application in mind though.  In my past, I tried online dating sites.  I’ve found them to have serious problems:

  • Somehow, Tinder has not figured out how to detect escorts offering their services through their app
  • Heterosexual men appear to be a.) lashing back at heterosexual women for not responding to them and b.) sending inappropriately forward messages – based on warnings I saw not to do so in female profiles
  • Some people have photographs professionally taken, others don’t, and others yet have a very good understanding of angles which make them appear different than they are
  • Text and pictures have continually disappointed in terms of their ability to convey personality, sarcasm, passion, …
  • Sparse data:  not enough profile, not enough pictures, …
  • None of the most popular apps appear to collect telemetry on what you click on, what you respond to (and what you don’t), even what you explicitly tell them you’re looking for through their configurations, when showing you UI like your list of messages / people who like you.  Simply put, you look at one screen, where you say you like X, then go to your messages screen, and see a lot of Y.
  • You don’t get actionable feedback on your pictures or text.  Why doesn’t the app AB test for you – like show 1 of your pictures to half your visitors, and the other to the other half, and let you know which one is working better?
  • People use other people’s pictures to pretend they’re someone they’re not.  This is a lot harder to fake when the messages are video.

Some of these problems don’t require video to solve.  However, some issues, like how you have absolutely no idea what someone’s personality is actually like after looking at some pictures of them, could benefit.

Imagine – an online dating solution which only has a phone UI, and the only communication you’re allowed is video.  Your profile is a video.  Your messages are videos.  You don’t appear to anyone else on the app until you have at least 1 video.  You’re not allowed to upload professionally produced videos from your file system – the playing field is leveled by the (relatively) minor differences among phone camera quality.

Swipe left, next video starts playing.  That person gets stats regarding at what point people tend to say yes or no.  This gives them useful feedback about what resonates with people and what doesn’t.

We automatically capture the text of what was said in the video.  If you keep saying “yes” to videos of men who like “adventure” and “… went to college …”, then we only show you those kind of men in the future (videos which contain the same text in their auto-captions).  I think there are opportunities here for image recognition as well…  maybe you really like girls who wear green?

A Skype-like function might make sense in such an environment for a “1st” date.  Safer / less time wasted?  Since you’re already used to, at that point, communicating with video.

Some people are nervous when they’re just beginning conversations with a new person.  They’re not sure what to say.  They’re not sure what to talk about.  They need some time, so something more like leaving a voice-mail would make sense.

This video solution is not without its own problems:

  • The cost of running the service is much higher.  It takes considerably more space to store everyone’s videos.  Maybe we delete them after they’re viewed, unless it’s a video that acts as your profile?  Like SnapChat for dating.  Certainly I’d expect some Vine-like limit on video length.
  • How do we avoid people videoing themselves inappropriately (think Chat Roulette).  I’d certainly expect the system to perform sentiment analysis and look for keywords to hide the video entirely or certainly behind some “I’m sure I want to view this” button.  Maybe there’s another opportunity for image recognition technology here as well…
  • The friction is much, much higher.  It’s takes little effort to swipe right / left.  It takes an order of magnitude higher effort to send a message.  How frequently do you update your pictures?  It’s going to be hard to get people to take videos.  I think integration with 3rd party services probably helped reduce friction for pictures (e.g. Instagram).  There’s probably an opportunity to connect to Vine, Facebook, etc.  “Oh cool, you like wakeboarding.  And I see, you actually know how to do it.  You didn’t just put up the picture of you staying up for 1/2 second”
  • The risk associated with a photo of you showing up elsewhere on the Internet is low.  However, if someone felt the video you posted of yourself was interesting for some reason and put it up on YouTube – this could have huge negative consequences.  We’d have to lock things down in such a way that videos didn’t easily leave the app.

That being said, I think such an app would be hugely beneficial, if for no other reason than a thought exercise in adoption and scale.

Let’s build it?