Analysis for 'Tremor Media'
Thursday, October 29, 2009, 9:40 AM ET|
According to comScore's September Video Metrix report reflecting actual unique viewers, Tremor Media is the top-ranked video ad network, with 33.6 million unique viewers, followed by BBE with 27.4 million and BrightRoll with 25.1 million.
comScore has not yet released it monthly top 10 results, but a sneak peek shows that Tremor places #8 on the list (I believe the first time an ad network has cracked the top 10), Jambo Media, a video syndicator (and also a VideoNuze sponsor) comes in at #9 with 32.3 million viewers and Facebook shows up at #10 with 31.1 million. All 3 companies are new to comScore's top 10 and compared to comScore's August top 10 list, they replace Turner (now #12), AOL (now #15) and Disney (now #20). All 7 other top 10 sites are back, though with a little shuffling (Google/YouTube, Fox, Yahoo, CBS, Viacom, Microsoft and Hulu).
With respect to the video ad networks specifically, as I've written previously, there's an ongoing debate about which numbers are most relevant to focus on. comScore has been working to fully populate its actuals list, which requires cooperation from the video ad networks themselves. Another way of measuring video ad networks' size is by "potential reach," which considers the total number of viewers of all the sites in a network (so for an ad network that would mean all sites it has the right to place ads on). Looking at both provides a broader picture of video ad networks' size.
By the potential reach measure, among video ad networks, Tremor is the top-ranked, with 72.9 million unique viewers, YuMe is #2 with 66.2 million, Ad.com is #3 with 57 million, SpotXchange is #4 with 55.7 million, ScanScout is #5 with 54.9 million and BrightRoll is #6 with 51.4 million. Oddly missing from the potential reach list is BBE, which in August was the fourth-largest video ad network with 62.7 million unique viewers. I'm trying to get an answer to that one. Tremor also announced yesterday that 60 sites have recently joined its publisher network, including A&E, Hachette Filippachi US, Thompson Reuters and SBTV.
It's also worth mentioning that Google/YouTube continues to dominate the video landscape. In September it is up to 10.4 billion videos viewed (vs. 10 billion in August), with a 40.2% market share (vs. 39.6%) in August. As the comScore data compilation slides I offered on August 31st support, Google/YouTube's share has hovered consistently around 40% since the middle of 2008.
Data like the above is obviously extremely important for understanding the evolving online video landscape. I'm cognizant of many people's concerns that the comScore data is incomplete or does not synch with internal logs or other measurement techniques. However, comScore is the only third-party data source that consistently releases results, providing trend data to analyze. Although I wouldn't suggest "taking the data to the bank," I do believe comScore provides great directional evidence of the market's growth and the standing of individual players.
What do you think? Post a comment now.
Wednesday, August 26, 2009, 9:11 AM ET|
Below is a first look at comScore's rankings for video ad networks' "potential" reach for July '09. The rankings, which have not yet been publicly shared, reveal a relatively tight clustering of 5 video ad networks - ScanScout, Tremor Media, YuMe, Broadband Enterprises and BrightRoll - with ScanScout capturing the number 1 spot in its first month being fully measured by comScore.
The "potential reach" aspect of these rankings is important to understand. As I explained in June in "Unraveling comScore's Monthly Viewership Data for Online Video Ad Networks," the potential reach numbers account for the aggregate number of viewers of all the sites that the ad network has the right to place ads on. However, as I discussed with Tania Yuki, comScore's director of product management, it's not a perfect measure, though comScore is continually trying to improve it.
The rankings are determined through a combination of the ad networks' self-reported publisher list and comScore's own tracking. If a video network reports that any one publisher accounts for 2% or more of its viewers, comScore requires a letter proving the business relationship. There is also a self-policing mechanism as comScore provides a "dictionary" of all publishers that each ad network reports. Competitors can review the dictionary and appeal to comScore if something appears amiss. Still, there's some looseness in the methodology, and having spoken to a number of industry executives, also a fair amount of concern that it is accurately portraying the industry's true performance.
comScore recognizes the limitations of the potential reach approach and that it is just one way of understanding a video ad network's value. Actual monthly performance is equally important, and comScore has been working with ad networks to implement this reporting as well. As I wrote in June, the "hybrid" approach requires ad networks to insert a 1x1 beacon in their video players. Though this approach also has its limitations, many of the biggest video ad networks are now implementing the beacon, and soon comScore will likely begin reporting actual as well as potential reach.
Video ad networks are a very important part of the online video ecosystem, responsible for placing millions of dollars of ads each month. Importantly they allow a level of targeting and reach that brands seek, but are often unable to attain on their own with a handful of direct site relationships. With the online video medium still relatively new, buyers require data helping them understand their options. However, the comScore data is just a first filter, diligent buyers still must dig in to understand how each network, or individual site meets their needs.
What do you think? Post a comment now.
Monday, June 15, 2009, 9:55 AM ET|
A monthly reminder that online video remains a work in progress is comScore's viewership data for online video ad networks. Even as someone who follows the industry closely, I find these reports confusing. The press releases often distributed by various online video ad networks touting their progress only adds to the confusion. I touched on this last month, and to clear away some of the fog, last week I spoke to Tania Yuki, comScore's product manager for its Video Metrix measurement service.
comScore's traffic reports are extremely important for the online video industry's growth because they are a key source of data for advertisers, media buyers, agencies and others looking to tap into this new medium. Ad networks in particular are an important part of the online video ecosystem because they provide significant reach, targeting and delivery technology, all of which are required by prospective advertisers.
A key part of the current confusion is that each month comScore's Video Metrix Ad Focus report - which details the total audience of unique viewers for online properties and ad networks - combines both the actual audience of destination properties with the potential reach of video ad networks. For example, here's the top 10 for April:
As you can see, 5 of the top 10 listed are ad networks, whose measurement is potential, while the other 5 are actuals. "Potential" is supposed to represent the aggregate number of viewers of all the sites that the ad network has the right to place ads on. However even the validity of this number is amorphous, because networks are only required to provide comScore proof of their relationships if the site accounts for more than 2% of all streaming or web activity.
Recognizing the need to provide more clarity, comScore has recently made available the option for networks to participate in a "hybrid" measurement approach, meant to track networks' actual viewership. To participate, networks need to place a 1x1 pixel, or "beacon" inside any video player where their ads appear. comScore takes the data reported by the beacons and combines it with its 2 million member panel of users whose behavior it tracks. It reconciles differences between the two through a "scaling" process that looks at the intensity of users' non-video behaviors.
To give a sense of the difference between potential and actual, comScore reports BrightRoll - which along with Nabbr are the only video ad networks to have implemented the beacons by April - as having 26M actual viewers vs. the 62M potential reported.
comScore's hybrid approach, which fits with its recently-announced "Media Metrix 360" service, is an important step forward in providing more clarity on how video ad networks are actually performing. Still, as Tania explained, even the hybrid approach has its own idiosyncrasies. For example, some publishers resist having a network's beacon incorporated into their video player, because they want to receive traffic credit themselves. Further, it is a voluntary program. Tania said that in addition to BrightRoll and Nabbr, other networks like BBE, YuMe and Tremor are all working through the implementation currently.
The actual numbers are important for buyers, so that ad networks' viewership can be assessed on an "apple to apples" basis with online properties, as well as non-video options. Tania said that media buyers tell comScore they value both potential and actual numbers. Though that sounds right to me, I think that for the online video medium to mature, buyers are going to put increasing emphasis on actual performance, particularly as it relates to existing media. That's why recent efforts from YuMe and Tremor to translate online video's impact into TV's gross rating points (GRP) paradigm are also important.
In short, comScore seems to be doing its part to improve reporting clarity. However, this isn't going to resolve itself overnight; the market will continue to experience reporting confusion for some time to come.
What do you think? Post a comment now.
Thursday, May 14, 2009, 9:21 AM ET|
Amid the steady stream of sneak peek press releases I'm sent each day, one I received late Tuesday from Tremor Media, the video ad network and monetization platform, caught my eye.
The release cited March data from comScore indicating that Tremor's network now had potential reach of 137M unique users and 57M unique video viewers (both unduplicated). The former number is from comScore's Media Metrix Ad Focus report and the latter from its Video Metrix Ad Focus report.
In particular, the latter number stuck out because I recalled comScore numbers from just 2 weeks ago that revealed the viewership for the top 10 video sites. Google (YouTube) was #1 with about 100M viewers, and Fox Interactive (mainly MySpace) was #2 with about half the amount, 55M.
comScore's new data meant that Tremor's potential reach was second only to YouTube's actual reach. And if you make the argument that much of YouTube's viewership is still UGC, while Tremor's network focuses solely on premium publishers, Tremor would be #1 in potential reach against premium video, a key point of the release. It's also worth noting that 2 other video ad networks focused on premium publishers also show up in comScore's top 10 for potential unique viewers- BrightRoll with 56M and YuMe with 41M.
Tremor's VP of Marketing Shane Steele and market research manager Ryan Van Fleet walked me through the data further yesterday.
First, it's important to read these numbers carefully, as there's a little bit of apples vs. oranges going on. The Video Metrix Ad Focus report combines actual viewership by the destination sites (e.g. YouTube, MySpace, Yahoo, Hulu, etc.) with potential viewership by the ad networks. The report clearly denotes what's considered "potential." If I understand it correctly then, the comScore numbers for ad networks should be read as "here's the total potential audience of viewers you have access to." However, what percentage of this accessible audience actually gets an ad served by the ad network is only known by the ad network itself.
VideoNuze readers will recall there's been a lot of sensitivity around these comScore numbers, since last summer a minor kerfuffle broke out over comScore's ranking of YuMe's traffic. Initially it attributed MSN's full audience to YuMe, but later revised YuMe's ranking down by only included pages against which YuMe ads could be served. comScore also stated that on an ongoing basis it would report "potential" reach for ad networks based on documented agreements and "actual" reach for those networks that included certain tags. The new Tremor numbers reflect this potential reach measurement.
It's also important to remember that comScore filters its data to arrive at unduplicated reach. As I understand it that means that if for example Tremor had USAToday.com and Fox.com in its network (note Tremor doesn't disclose its publishers except to its advertisers) and a single user watched video at both sites, the user would only be counted once in Tremor's potential reach. I don't know how exactly comScore de-duplicates viewership, but let's assume it's accurate.
The extent of Tremor's reach (along with BrightRoll's and YuMe's), particularly against premium video is an encouraging sign. I've written in the past that key inhibitors of TV ad dollars moving over to online video are both scale and various friction points in the ad buying process. The comScore data demonstrates that a cluster of ad networks is emerging that can deliver against TV ad buyer's reach expectations, while adding new targeting and reporting capabilities unavailable in TV. There have also been recent enhancements to these companies' reporting/analytics (particularly around GRPs) to synch up with TV ad buyers' expectations.
The online video ad model continues to grow and evolve in spite of the current recession. This is particularly important for expensively-produced premium video where effective online monetization is crucial.
Chime in here with a comment if you think the comScore data or its implications needs further clarification.
Video Research Around the Web
- Cord-Cutters Show Interest in Discovery Channel: Survey Multichannel News
- Pew: Mobile Broadband Users Double Since 2013 Multichannel News
- D2C Advertisers See Sharply Higher TV Ad Spending Mediapost
- Forecast: OTT Revenues To Hit $23.7 Billion By 2023 Mediapost
- Data-Enabled TV Buying Rises, Despite Concerns Mediapost
- Netflix Again Blows Past HBO, Others as No. 1 Pick for ‘Best Original Programming,’ Survey Finds Variety
- Cord-Cutting Got 75% Worse in Q1, Most Terrible Quarter Ever for Pay TV Multichannel News
- Many 4KTV Owners Don't Use Advanced Content Features Mediapost