Commentary
Online Video Finally Works: Now It's Time To Make It Better
- by Adam Singolda , March 10, 2010
When the online video market began to take hold in 2005, many publishers grappled with the proper way to monetize the emerging online video-watching trend. It wasn't easy, and for the last couple of years there was a question whether it was even possible. Yet, a recent study published by eMarketer shows that while overall online ad sales declined in 2009, "video ad spending growth will far outpace any other online format, running in the 34% to 45% range from 2009 through 2014."
So it seems like publishers have finally found a way to monetize their video assets. How did they make it? And what is next in line for those publishers? The story starts with the well-known 30-second TV ad spot.
TV has popularized the 30-second ad spot. In the early days of online video, publishers and advertisers were trying to repurpose this format to online video. But what worked well in the middle of a 30-minute TV program, did not work well before a 3-minute online video clip. Users simply didn't consider the value of a 3-minute clip to be big enough to "pay for" by watching a 30 second pre-roll.
In the last two years, advertisers, beginning to see the untapped potential in online video advertising, have started producing 15-second spots for online video. The 15-second pre-roll works much better online, producing a mere 16% drop in views (versus 40%-70% with 30-second spots).
The situation for publishers of video content has now reversed. From owners of a non-monetizable video content, they became owners of inventory that is applicable to a vast amount of 15-second pre-rolls that are being produced. Yes, this is the small secret of online video: large portions of semi-professional and professional video content are fully sold.
We expect the trend to continue. Many young users are not watching TV anymore, preferring to consume their video online. And online pre-rolls have many advantages for advertisers when compared to TV spots: they are accurately measurable, the user has to watch them and cannot "change the channel" without losing the ability to watch the original video clip, and the Internet allow advertisers to better target the video ads to the right audience. These advantages, coupled with the growing availability of 15-second spots, will cause ad spend to move from TV to online.
So what should publishers do next?
The answer depends on the type of video content provided by the publisher.
UGC sites should increase the development of more professional videos by either producing it, or syndicating it. On top of it, those publishers should use the right tools to push users to those professional videos that are more monetizable. This can be done by manually promoting specific categories/videos or by automatically biasing the discovery engine engine on the site.
Professional and semi-professional sites should pursue growing their monthly streams, and monetization will then follow. There are two innovative ways to grow streams significantly:
(1) Place relevant video links on article pages to push article readers to become engaged video viewers. Today there are scientific ways to promote personalized videos on articles to best match what users want to watch next.
(2) Generate video streams from outside of your site. Traditionally, this was done by promotional, viral and SEO tools. New tools can also automatically place highly targeted links of your video content within third-party affiliate sites and blogs. In this way, users come across the most relevant videos from your inventory while visiting other sites.
So while it is becoming easier to produce/syndicate and monetize videos (meaning the monetizable inventory grows), it is also easier to distribute your videos across the Web. The bigger challenge will be to make sure users discover what they really want. Otherwise, they'll just go away.
Once the discovery code is cracked, the floodgates will open, users will stay longer, watch more, and come back. Your challenge then will be to best price your pre-roll ads, and build your brand in the advertisers' marketplace.
Shameless Plug Warning! Yup, this article is spot on - we've been working this plan for years now and only produce "television" that would be summarily rejected by broadcast or cable, but that's just fine with us. Our stuff is edgy, hyper-interactive (you can Skype right into the show) and has a VERY loyal (and GROWING) audience. We won an EMMY for www.kickstarttv.com and I don't have to tell you how popular www.medicinalmarijuana.tv is. Go to the sites and check it out. Traditional broadcast sucks.
Nice to read a straight, fact base column here for a change.
Can Adam please explain what he means by these three statements below and what tools he recommends?
"Publishers should use the right tools to push users to those professional videos that are more monetizable… by automatically biasing the discovery engine on the site.
Today there are scientific ways to promote personalized videos on articles to best match what users want to watch next.
New tools can also automatically place highly targeted links of your video content within third-party affiliate sites and blogs."
Denise, thank you for the question(s). I would be more than happy to answer each one of them, and i hope i can cover them.
(1) Tools I recommend, and what’s biased discovery engine
There are technological ways, and editorial ways to know what parts on the site are better monetizable. For instance, sometime news-publishers better sell or just get higher CPM videos on “health” or “tech” section on their site (better than they sell “broad news” section). Knowing that, some publishers actively promote those sections across the site (by mainly placing static links across the site).
However, another way to achieve this goal is using Recommendations-Engines that support “biasing” of this kind. Such recommendations solutions can be injected with the categories that the publisher “prefers”, and as long as the user experience is not hindered, those sections on the site will be promoted automatically to users that are likely to be engaged with them. At Taboola (www.taboola.com) we offer this feature to our partners, we think it is important.
(2) Promote personalized videos on article
In many cases, what happens is publishers have a lot of “flat content” page views (articles, images, etc’) and they have a strategy to become more video-centric (mostly as video is engaging, and can generate higher CPMs over time). What publishers can do in such case is to place video-links based on category, or popularity even next to the flat content to convert users to become video viewers. At Taboola, what we offer is a way for our partners to use a product we call Text2Video which essentially leverages our prediction technology to try and figure out what a user would like to view next almost agnostic to where he/she is on the site at the moment (based on many metrics such as GEO-location, day of week, video length preference, and many others). If you’d like, you can check it out here http://taboola.com/Text2Video.html
(3) Place highly targeted links on affiliate sites
For this part of the post, I am referring to a new affiliation management platform that allows publishers take a branded recommended videos widget and distribute it to affiliate sites and blogs. I would recommend you check out this post on the Brightcove blog that Aviv Sinai, our Product Manager for this product wrote. On the post he describes into more detail this new method for distributing videos across the web in a scalable way to win streams/revenue/brand: http://tinyurl.com/ygnjyym
I hope it helps!
--adam
founder & ceo Taboola
www.twitter.com/adamsingolda
Are you kidding us, Adam? The entire pre-roll model is sheer pablum, and a horrible waste of video assets.
Are you honestly trying to sell this pig in a poke with the rationale that making consumers suffer through a 15-second ad that they don't want to see in the first place is somehow an improvement over the version that tortures them for a full 30 seconds?
Why is everyone in this industry nowadays playing not to lose instead of playing to win? Chief Inspector Jacques Clouseau would be proud: "I have devised a plan guaranteed not to fail."
Gimme a break.
Jeff, thanks for the note. some thoughts below, let me know i was able to cover your thoughts on this.
As Robin Wauters (TC) said today, and I believe it makes a lot of sense from a user perspective -- “i’ve yet to come across any form of digital advertising that people truly love”
My take is always to believe in numbers, quantify/quantify/quantify things. It's also our culture at Taboola. You see -- being able to A/B test things, measure over time in scale, and compare trends & patterns help to make decisions with regards to what works, and what’s not. Other wise, it is very opinion based, and that is never a good way of making big decisions (what do one think next to another, who is right, who is smarter?). I’m sure you have similar methodologies at “Brothers Einstein” when you make brand strategy recommendations to your customers.
In my segment above, I relied at some part of it on a good job I think TubeMogule did, saying how 15sec pre-toll behave differently than 30. Not “what they think”, or what “they’ve heard”, but rather -- what (as far as they studied & published & I understand) happened. And that’s great for us (you/me/sites/companies) as we can realize this detail matters to users.
to your question if i'm convinced that this 15/30 change is the holy grail, i can't really say. i also usually don't believe in one thing that makes a whole difference, usually things are truly affected by combination of little changes that make sense. However, this is one thing that was measured & made a difference. i thought it's interesting.
About the targeting thing (“don’t want to see in the first place”): I think you touch on one of the challenges video ad networks have and that is that as opposed to search-ads, banner-ads, text-ads (=unlimited inventory)— video ads are limited in their inventory. It’s harder to create video ads. It means that if you have large Group A (users/videos) and substantially smaller Group B (video ads) and you need to match between them, you may stumble upon many cases in which you don’t have a great match. That challenge is declining as I think video ads are moving more frequently into the web, and sometime as I wrote above even being created to the web directly. Also it’s easier to make a 15sec pre-roll than 30 which also helps the inventory issue. All that combined will keep improve the targeting, and I suspect we’ll see better matching down the road of the pre-roll we face before the actual content.
I hope all of this makes sense to you Jeff. Let me know if you have other thoughts
regards,
--adam