Monthly Archives: February 2013

Virool Viral Video Advertising Network

Virool: The Viral Video Black Box

Self-serve YouTube video ad network Virool seems to be on a tear, raising a $6.6M seed round in February from investors that include Thomvest Ventures, Menlo Ventures, Draper Fisher Jurveston, 500 Startups, Phenomen Ventures, TMT Investments, DominateFund, FundersClub and a cadre of well-known individual investors. Virool also boasts some solid traction, including 30,000 registered advertisers and a network said to be capable of reaching 22 million viewers.

I first experimented with Virool in August 2012, and, even back then can say simply: Alex and Vlad built an intuitive product that works. Enter the URL of your YouTube video, pay money and it gets views through their embed network. The team has also clearly done a good job scaling their affiliate network, making the product an interesting alternative to AdWords for Video. As a result, since mid-last year, I’ve run several different Virool campaign split and A/B tests, with frequently interesting results.

However, once you get past its basic product efficacy, Virool is, at its core, a bit of a black box that leaves some important questions unanswered about it’s ad network. Specifically:

1. Who’s watching?Virool currently allows you to “target” your demographic (demo) by age, gender, placement, geography and keyword. However, the results have yet to inspire full confidence. Geographic demo-targeting at a basic level seems to work: Virool maps your views in the dashboard and assuming the views aren’t being run through proxies if you select United States only you’ll get United States-based view views. However, after that, the system offers very little transparency. If I select “Blogs/Sites” only for my view placements, what blogs and sites are these? And how is a blog or website segmenting its traffic by keyword or visitor age/gender? I can say that video campaigns targeted at a very specific demo (say, 18-24 year old male baed on a small set of keywords) doesn’t produce results consistent with YouTube’s demo analytics on the back end.

2. Where are they watching? I’ll give Virool some credit here because in a lot of cases Google/YouTube hasn’t been able to do any better, but going to the digital team at a sophisticated consumer brand or agency and presenting campaign results as “5,000 people from Brooklyn watched your video on some blog or website” is going to invite a lot of hard follow-up questions. What website(s)? News sites? Tech blogs? Are videos ever run through gates or video view exchanges? Currently with Virool it’s impossible to tell.

3. Do Virool viewers care or pay attention? Beyond the question of who’s watching where, an additional open question is how engaged are they? Specifically, when split-testing the same video across different distribution “channels,” I’ve found in particular that Games and Virtual Currency API’s can rapidly chew up ad spend and generate views that seem to have near-zero content engagement, even it it’s something like a Gamer Walkthrough or Live-Action Shooter clip. Why is that? My hypothesis is simply that audience-members want to play their game or build their currency base, not necessarily actually watch your content (some of these services you can even bot and run through video after video on mute in the background while you do something on a different browser). Trading virtual currency for YouTube views is hardly a new model, and if Virool is plugging into virtual currency view exchanges or gated ad networks to run through inventory it’s generating non-engaged shadow views at the expense of the advertiser.

4. So engagement is really a total dice roll (Do Virool viewers care or pay attention – Part 2)? For those who haven’t noticed, YouTube’s focus is on subscriptions and minutes watched, not views. Yes, there’s still a huge psychological reinforcement mechanism when you see a video with 4 million views, and at scale video virality can in some cases become a self-fulfilling prophecy (see: the Harlem Shake), but YouTube’s algorithm heavily weights audience retention, minutes watched and click-throughs at the expense of views. Here again, Virool’s view service is awfully sexy (particularly because it can be achieved quickly at lower CPV levels than Google’s PPC networks), but it doesn’t build audiences and it doesn’t give video marketers actionable data and insights about their viewership.

In order for me to really see Virool as the next generation of video ad exchange, Alex, Vlad and the rest of the team really need to develop the product to make it more transparent and address these unanswered questions. I’ve followed Alex on Quora for a while, even pre-Virool, and it’s clear to me he has significant expertise in the video space, the arbitrage he’s put together is really clever, and it would be awesome if Virool can apply its big raise to bring a lot more transparency to its viral video black box. Without it, it’s hard for a knowledgeable video marketer to really trust the product, because I know the shortcuts that can be taken and need conviction to know that Virool is taking the high ground.