Why Video Discovery Startups All Fail

Sad to see Cory Booker’s startup Waywire pivoting to “Pinterest for video” after starting with much bigger ambitions. Prediction: they’ll struggle because video discovery just isn’t a venture scale business.

My previous role leading product at YouTube put me in position to see a lot of video startups. Additionally, most large developers hit our radar at some point via their API usage. Many of the video curation/discovery startups were based on two hypotheses:

1. There’s an explosion in web video content (and it’s not all on YouTube)
2. It’s difficult on YouTube to find and manage the content you want to watch

These are both absolutely true yet it doesn’t matter – video discovery startups are flawed products and even worse businesses. Why? Because they don’t fit into a consumer’s mental model. They’re fine products – several are very well-designed – and I’m sure many user studies provide answers like “oh yes, I totally need this,” but in reality there’s no habit being formed. Here’s why:

Verticalized content needs context not just collections. Many of these startup founders would provide the example of a particular interest and describe how an enthusiast would want to come to their site to see [fishing, cooking, fitness, travel, etc] videos.  But just a collection of embedded thematic videos isn’t enough – the real fan wants content, community, editorial. They already get their videos as part of the sites they visit today – along with text and images. A bunch of three minute videos with varying quality, metadata and sources isn’t enough value, not when some blogger is already picking the best of these and adding content around them.

Horizontal content needs to solve search not just browse or curation. At YouTube we thought of a user’s video discovery needs as occurring across a spectrum of intent. On one side you have “high intent” which means someone searching for a very particular individual video. On the other side you had low intent – this is why millions of searches a day start with something very generic such as “funny videos.” In between the two would be more thematic queries such as “Lady Gaga” or “Louis CK standup.” Today even though YouTube isn’t an A+ in fulfilling needs across this entire spectrum it’s largely good enough to be the only dedicated video site consumers turn to for short form content. By and large they don’t want to sort their video destinations by intent (I go here to search for news video, here to browse for recipe videos, etc) because that’s unecessary mental load. Furthermore, and anyone who has ever gone down the YouTube rabbit hole can attest, a single user session is likely to include multiple points on the discovery spectrum. You start perhaps by searching for a video on how to fix a flat tire but before you know it, watching clips from High School Musical.

Social video is not a standalone product, just a signal. For a while many companies were simply pulling the video shared by your Facebook friends and people you follow on Twitter to create social collections of content. These failed as standalone products because (a) social graph != interest graph and (b) removing the context and conversation took social videos and made them non-social. Fail.

It’s hard to make money on other people’s video when you can’t monetize it directly. Finally, even those people who’ve been able to get a bit of traffic find it hard to create a business. Video sites have a great business model – the ad follows the content so when you watch a YouTube video offsite, you’re still allowing the content owner to get paid. When you’re simply embedding content your monetization is really limited. You can’t put another ad in the player. From a YouTube terms of service standpoint they’re perfectly happy to let you ad support your site so long as the ads don’t target the video content AND if you removed the video, the page would still have enough content to be interesting to the consumer (that’s the reasonable man test).

Some believe that native ad models are the solution – that content owners and advertisers will pay to promote their content. YouTube does this via sponsored videos and indeed it does work but requires a scale of both users and operations that these startups can likely not reach.

So what would I do? Probably create a great mobile app and a .99 per month subscription service to deliver a handful of curated YouTube videos to you each day. Highly thematic and where context is less important – hockey fights, vegan recipes, kid-friendly videos. It would be a niche business but one I think you could scale by getting folks to subscribe to multiple channels. But don’t take venture capital. Just build it.

[For a smart response to this post, please read this post from ShowYou CEO Mark Hall]
[Update: Some additional thoughts from the CEO of Telly, a social video site]