If you’re here. I hope you’re one of us. One of those who really, but REALLY enjoys live shows.
Let’s say I enjoy live shows so bad I’m saving so I could tour for a complete year with one of my favorite bands, what?! Yeah, crazy dream but whatevs.
If you’re one of us, stay with me.
What you will see next is my final project for Product School.
I’ve been a Tech Product Manager for well over 8 years but never really had any education in the field.
It was more of a “Startup CEO” turned PM. I just had to.
No certifications, no classes, just lots of reading, failing and learning quicker than the rest…
Alright, back to the point.
Quick numbers, to support my idea.
Yeah, what about that?! We’re not alone. Shit loads of people went to a live show last year!!
I know I know. COVID situation will kill these numbers, but might help with my product idea.
Why am I showing you revenue numbers and Instagram stories? Well… do the math… I think Live Nation is missing on massive revenue.
Same for instagram dudes, come on, it’s time to bank on this while helping others.
Let me expand on this. I hate to be one of those… I do it responsibly tho.
Every time I go to a show, I try to bring with me a small memory of what I just witnessed. For me and my friends following me on Instagram yes.
I kinda do it for them.
I think “Dude this show’s insane, I wish more people could experience this right now” then I grab my phone and post an IG story.
I feel like the Robin Hood of concerts.
Robinhood to all my 0.42 followers haha.
But think about it, theres no way you could go to every show. Or could you?
Sometimes you’re too far. Or you live in places like Chile, that are so far that not all artists care. And shows are usually crap here, without all the smoke and lights like in EU or EEUU.
Sometimes you’re just short on money. Shows are getting more expensive by the minute. They’e good yeah, but I feel a good chunk of people are missing out for being short on cash.
Sometimes you just can’t! I feel the pain for handicapped people that struggle with enjoying a live show. No one’s thinking of them.
Actually. 15% of the world’s population has some type of disability.
And ADA (In the US) regulations suuuuuuck!! Dude, they literally only ask for a shit 1% of the seats to be available for handicapped people.
That is not 15% not 10% not 5% dude, 1% come on.
How to re live a concert, today.
Ok. You missed that show.
Now you want to see what happened, what do you do?
Grab your phone, check Instagram, type #thatconcertimissed and… good luck my friend!
Today’s searching experience sucks. Let’s face it.
You can get to that hashtag, yes, but can you really enjoy hundreds of out-of context videos, photos and more?
I could upload a photo oh a hot dog, #museconcert, and yeah, Muse fans will be looking at my hot dog instead of those videos/photos they were looking for.
Multiple angles, no context, no timeline, just a dump of content under that hashtag.
But what if…
What if, we could actually get these videos, analyze them, understand what they are, what are they recording, what are they listening to, and put them together into a continuous/logical stream of events?
What if all videos of the same show, are put together in a way that allows us to re-live the show, like for real this time?
Stay with me, I think we’re onto something…
Let’s just grab all the videos and parse them with my special formula.
THE MAGIC, EXPLAINED
Video Processing Layer
Great. We have all the videos of a show, I mean we’re Instagram, we don’t have to scrape them from anywhere, we have them.
We could figure this out because of:
- Hashtags added
We get these videos and we analyze them. We put our best “Machine Vision” minds in project to come up with an algorithm that will help us out understanding:
- Video Stability
- Device Information
- Video Quality
- Perspective: How much of the show is the video showing vs showing useless stuff
- Face%: Let’s say the video is a selfie one, we might not want your face, but only the stage so, next!
- Distance from stage: Yeah we could figure this out to only stick with the best angles and distance.
You get it right?
We’re analyzing every video to end up with the very best ones o have the best representation of the real show.
Audio Processing Layer
In parallel, we analyze the audio of the videos.
We want to make sure we give you the very best of the best right?
Our algo analyzes the audio of every video.
This is important because we will use the BPM and audio to link the videos together.
With no audio, we won’t be able what part of the show are we recording right?
Cool thing about this is that, as we have sh*t ton of videos. We could be using audio from one video on top of another video.
This is critical.
We will look for:
- Sound quality
- Noise level
- Overall Quality
Here we try to stick with the best audio quality. We throw away the audios with too much screaming, or stuff that won’t be the best representation of what really happened.
Are you with me?
We’re almost there!!
By now we know what are the best videos and the best audio from all the Media we’ve collected.
Just one more step!
Let’s put all this together.
We have so many videos that sum up the whole show right?
Now we have the best ones, so it’s just a matter of putting them together into a stream that makes sense.
Something that goes with the beat, something like re-living the show, on your phone.
So we grab videos and audios and merge them so we get a continuous stream.
Video 1 from S1 to S23
Video 2 from S23 to S55
Video 3 from S55 to S60
Now we have a 60 second video made by these 3 other videos.
Francisco what are you talking about?!?!
Stay with me.
WHAT YOU'RE GONNA SEE NEXT
IS THE ALGO WORKING
4 DIFFERENT PEOPLE
4 DIFFERENT ANGLES
SYNCED-UP BY THE AUDIO
Just think about it! Every live show, in your pocket.
Now everything changes.
I write this blog post right after the release of “Reels” and I can see how IG Shows can be released just like that.
Go to “Search”
Boom! The next will be history…
The Search Experience
Now we’re talking! Right?
No more silly hashtag searching.
Just search for the artist/show and fasten your seat belt.
Of course, we’d know what’s “Happening Now”…duuuhhh!
And of course we could let you know what’s coming to remind you of those shows.
Wanna get a little taste?
The next video will give you an idea of what I’ve been talking about.
Still with me?
Press "Play", this is what I'm talking about:
Multiple points of view
One continuous timeline
In your pocket
There you have it. 12 different angles. 1 min video.
Sorry for the poor sound quality but Youtube’s Copyright team is after me.
This product can be done in a hackathon. Literally.
With the right people behind the computer, we could build this. Fast.
Best thing is that we already have millions of videos to put together and start polishing the algorithm.
The pilot phase is basically testing the hypothesis, right?
We will start listing few shows and see how many people “get” the idea.
Here we’ve already proven that people want to browse for specific shows.
This is the part where we want to start pushing social features to invite friends, to “record together”, or to have “verified angles” meaning people that are great at recording live shows and their content is more likely to be shown in the final cut.
This is the controversial part.
But it will happen one way or the other.
The first versions of the feature will mostly include amateur recording + some good quality videos from attendees.
But here’s where hardware comes in place.
We must implement simple hardware in shows to get 360º takes as well as great sound quality.
This is is basically the future of live shows.
This is where you integrate with Oculus and add a suscription based plan so people anywhere in the world can enjoy a live show.
What about that?
100.000 seat arena? What about 10.000.000 arena?
What about 100.000.000 arena?
That’s the future.
V1, V2, V3, VR.
This is a very light take on KPIs I’d focus on, while leading the development of this feature.
On Launch Stage I’d focus on:
- Conversion: How many users tap “Shows” after released
- Searches: How may users search after tapping “Shows” (Important here to have another funnel if people searched for an artist/show and couldn’t find it, how does this affect retention)-
- Views: How many users watched a show or part of the show (important difference)
- View_Length: Here we expand on how long do people last. I think people won’t see a complete show on their phone. So with this we can figure an average time to maybe summarize the best takes on the show.
- Reminder Requests: How many users requested a reminder. This can tell you how many users will be interested in a show so you can step up in quality or something. This will come handy when monetizing.
I’ll leave it here for now.
“How would you measure success?” … Would be an interview question..
To make it simple, I’d think of two personas.
First, the one that goes to the show.
Second, the one that watches the show on “Shows”.
I think with enough data, we could figure a trend of which users are more likely to watch shows based on their previous content, check ins, etc.
So I would pick that demo and say “These are the users that will love this feature” and measure success against how many of these users end up using Shows, and then how many come back.
In general terms
- Frequency: How many after watching Shows, come back, when? How often?
- NPS: How many more users come based on interactions with other users and/or because of sharing tools.
- Viral Coefficient
- Retention/Rolling Retention
Well that’s it guys.
I leave many questions on the table, yes. But come on, good enough for a blog post?
If you liked it, please share!
See you around.
“System for multi-angle live video parsing and processing”