its no surprise that those who own or sell advertising on internet video sites didnt respond favorably to my last post. Its far more profitable for them to sell 1k views an hour for 24 hours than to have to deliver 24k viewers in an hour or even worse, at the same time.
It shows the max number of users that a show can have at a point in time , which advertisers will evolve to in buying time critical ads (movie buys on Thursdays before opening on Friday). If an advertiser for Shrek 19 or Saw XX needs to reach 10mm people online before their movies open the next day, knowing the maximum and average number of simultaneous viewers for the content you are buying gives you the proper constraints required to do the math on the buy.
The number of simultaneous viewers a site can support also tells you what happens to your ad if something unique happens. Its hurricane season. Lets say there is amazing footage from a hurricane or some other event, maybe a Paris Hilton prison break and one site has the capacity to handle 1mm simultaneous viewers, the other 100k. If your ad is on one, you might be ok. If its on the other, you might be out of luck. Your ad is delivered at a snails pace if at all, with users seeing more buffering then ad.
Then there is the issue of cheating. its easy to cheat on views. Youtubers game their views all the time, as they do on all video sites. To advertisers, each site defines a view differently. is it when an ad is viewed for 1 second, 10 seconds, to its entirety, to its entirety without buffering ?
If you buy an average of 1k simultaneous viewers, thats a much more difficult number to cheat. The advertiser can get a link into the serving of their ads and monitor how many people are watching at any point in time which is something they can learn from.
The number of simultaneous viewers also tells the “intensity” or heat of an ad or content. 1mm simultaneous viewers has a completely different impact than 100k per day for 10 days which is different from 1k per day for 1k days. Tracking the number of simultaneous viewers gives you the true feel of the viral nature of the content and it smooths out much of the cheating by video uploaders
Finally, the number of simuls give you a reference point to TV . what is the maximum number of simultaneous viewers that Youtube, or AOL , Yahoo, MSN or ??/ VIdeo have supported in aggregate across all their content ? What is the maximum number of simultaneous viewers that any given piece of content or advertising has achieved ? Thats a number thats easily comparable to a TV rating. The number of views isnt.
Back in 1999, when we broadcast the Cubs during the day, a 6k audio stream could get 25k plus simultaneous listeners and often peaked past 40k. Telling an advertiser that they could reach 40k people at once was powerful. Far more powerful than the number of views or listens content had.
its time we start to use average and maximum number of simultaneous viewers as the benchmark for video .
Prekle.Com
Comment by ali gumus -
I suppose it\’s too much adv. over the internet
Comment by laminate flooring -
As always the devil is in the details. The example you gave of a live broadcast does not translate as well into the realm of on-demand content. I certainly understand where you are coming from and would like to add some additional comments.
If advertisers are asking for a run of network ad but then simultaneous viewers makes more sense. If the advertiser is looking for a specific demo or wants to buy a property, like news sports or entertainment, then total simos might not mean as much.
If I were an advertiser I would want to understand the total number of unique viewers and volume of traffic a site sees per some time increment .. say per day or week. I dont think this has to be simultaneous as in the live broadcasting world. For most purposes you could live with a day or even a week. If I am buying a subset of traffic then I want data for the subset and the total. You can throw in total capacity to round out the info.
One of the most abused aspects of internet advertising today the frequency of ad views per user. Do you remember the PSAs we had in the early days? People were so sick of seeing the ads that they called said they would never support the groups again if they had to see the ad one more time. If a site only has a few unique ad spots this can be a real problem.
Comment by Justin Madison -
mark,
i see ur point but i work in the online video biz and what you are asking will not happen anytime soon. 3 reasons:
1. $ – this is the only way to show mass dist and demand the cpm\’s at this point.
2. media buyers have no idea what you just wrote. so how can they ask for something they dont understand
3. why should we hold our industry to standards that tv cant hold itself too. i have to think if the web 1.0 guys (banners) could do it all over again they would not have painted the web into this corner of stats, stats, stats and focused on the brand impression like tv.
overall we need to help marketers take baby steps bc it is fun for us (web junkies) to have this convo but i am still trying to explain to people why not to buy a myspace homepage takeover for 500k and call it innovative web marketing. and those f#$%ing brand profile pages…give me a break.
this is the world we live in. as a matter of fact if u ever want to start a real digital agency give me a call. thats where u need to start.
marx.
Comment by marx -
I know it is OT, but I just heard that Genarlow Wilson\’s sentence has been commuted. I think Mark would be happy knowing that.
Comment by Ghost of Tom Joad -
I think the numbers should be used together, but the weight should be on the number of simuls. If your site isn\’t popular, it doesn\’t matter how many simuls it could support. But even non-popular sites can get a surge of viewers via Digg, Slashdot, or BoingBoing. As bandwidth costs continue to decrease, everyone will be able to host their own videos with little to no cost to them. YouTube is like the Geocities of today- it does the hosting because it is cheap and easy to use. Soon, this won\’t matter and they will disappear into Internets History.
Comment by The Big Fez -
But the thing is that it may likes the other peoples too. Any guess why??
Comment by Christopher -
Does this site look familiar mark?
http://web.archive.org/web/19990208021642/http://www2.broadcast.com/
Comment by Jordan Young -
I do not agree with Mark regarding video streaming, I really believe that people stream videos a lot.
Comment by Ruby on Rails examples -
Mark Cuban doesn\’t worry about having everyone agree with him. That\’s one of the things I respect about him. He says what he thinks and let\’s the chips fall where they may. Keep going Mark! Let us have it … even when I don\’t like what you say.
Scott Hove
Mr. Motivational Speaker
http://www.mrmotivationalspeaker.com/
Comment by Scott Hove -
Wow, I never thought of the ad delivery speed like that. That makes sense, especially the hurricane analogy.
Comment by Sean -
Mark,
I have to disagree with you as a general point – though as some other commenters have noted, it might be more relevant in the case of event/time sensitive marketing. [Full disclosure, I am launching an ad network this summer to serve ads into applications – mobile, web & desktop, while video is not sole focus as a network capable of serving into an application, we are definitely going to look at opportunities with publisher partners that have video applications]
I think you are missing a vital and key point about online video.
Most people DO NOT STREAM online video. Yes, with the launch of apps such as Joost, vuze and others this may shift – but the vase majority of online video is either:
a. selected via link (or less often search) and watched one at a time via use choice (i.e. follow a link to YouTube to watch a video, do a search on google/youtube, watch a video which has been embedded somewhere – via youtube or 100\’s of other embedded media players
b. and very common – SUBSCRIBED to video (often via iTunes but there are other options, just yesterday Jay from Digg & Revision3 while speaking at the Future of Online Advertising conference in NYC mentioned that Diggnation gets over 250,000 DOWNLOADS of EACH episode – the vast majority of which are thru iTunes. So many of their viewers are then watching that video on a regular tv via Apple TV or an XBOX360 (or I suspect less often via another media centerlike device). Or watching \”on the go\” on a laptop or portable media player.
For these downloads the concept of \”simultaneous\” viewers is utterly meaningless (as are ad models that assume connected users – Jay was suggesting that one successful route is ads incorporated into content – sponsorship models/content creator read/performed copy etc)
I think that even with streaming apps such as Joost the concept of \”simultaneous\” viewers is also meaningless. These apps, and indeed the VAST majority of online video are NOT time sensative videos. Indeed inside of Joost you can pick and choose which bits of content to view – I choose to watch episodes of Lexx in the order of they were created for example – much as you might via a Tivo on a traditional TV setup)
There are only a handful of cases I think you are correct about:
– streaming videos of live events – from the small scale uStream, slightly larger Justin.tv, or the modern versions of what you did with Broadcast.com years ago – online presentations of live events, conferences, shows, or other performances. In those cases the capacity of for viewers at the same time is important to know
– online distribution of, for example, live sporting events.
However even in these cases I would argue that the \”best\” advertising solutions will not rely on a simultaneous user model – as the same ad would NOT be (in most cases) be being served up to all of those users.
Instead the ad which is shown should be adjusted for factors such as country of the viewer, language preferences, at times demographic profile, past indications of interest (via profiles, past click activity etc), probably also assumptions made based on technology (i.e. show ads for new computers to viewers running old browser versions) – no matter how it is technically implemented (pre-roll, interruption akin to traditional tv ad breaks, overlay of ad over content (which I personally find annoying along with many others), or a post-roll ad (best only when the viewer has other shows already cued up – when it then perceived as an ad break before the next show starts).
And don\’t forget that with online videos you can (except as mentioned when the video is a live stream) PAUSE the video. That is a big part of the appeal – not just of online video, but of services such as Tivo. You can also rewind, and can easly skip to another alternative entertainment source.
Caching networks and the ways large scale sites work also make a raw simultaneous user number non-trivial to calculate and think about. For example, a network might be able to handle 100k+ simultaneous users within Korea, but break down much earlier than that should 100k+ users from Vietnam tried to access the link (or from rural America). Not due entirely to a failure of the content delivery network globally – but to a capacity limitation on a portion of the caching network which might have deep capacity from high use/high bandwidth countries – and fewer and more limited capacity in lower use/lower bandwidth countries.
And a still further consideration is that online video is NOT just about the big, aggregated distribution channels (youtube, aol video, Joost etc) but also about a much wider range of channels – 100\’s of them (iTunes, Tivo, some offerings from cable networks, direct websites, rss feeds, search sites, 100\’s of smaller video hosting sites, blogs, social networks etc) In many many cases these distribution channels take a single copy of the content from the publisher – and then manage distribution of that content to their users directly. In some cases also distributing additional ads around the experience. (and Jay in that same speech talked about the business negotiations this then creates between these distribution networks and content creators)
thanks for sparking the discussion – an exciting but unsettled question – and I, for one, think that the solution will NOT be a repeat of TV models but newer models and metrics that will arise. Mixing models from old & new media.
Shannon
Comment by Shannon Clark -
\”Both abundance and lack exist simultaneously in our lives, as parallel realities. It is always our conscious choice which secret garden we will tend… when we choose not to focus on what is missing from our lives but are grateful for the abundance that\’s present — love, health, family, friends, work, the joys of nature and personal pursuits that bring us pleasure — the wasteland of illusion falls away and we experience Heaven on earth.\”
Comment by KindAndThoughtful -
Yes..this is just the way that everybody want.
Comment by Melissa -
Has anyone seen the black and white Lebron ad floating around? Cool. The music is perfect.
Comment by Brandon -
Great post and cool blog!
Comment by Adi Azar -
Interesting to note that, power outages and the rare network or local cable feed or broadcast glitch aside, TV ads are not subject to being cut off, not delivered, or slowed down to the point where the viewer gives up, as is the case with Web-delevered ads.
–Mike
http://www.michaelabanks.com
Comment by Michael A. Banks -
I think the simultaneous issue only matters for time-sensitive ads. Tracking TV ads is different because you have to watch TV at that time and are forced to view the ad, not taking DVR into account. Taking DVR into account, those ads are not viewed. You surf the web when you want and therefore get exposed to an ad at your discretion, not the hosting site\’s discretion. If not time sensitive, who really cares when one watches the ad? Tracking clciks can help an advertiser select when to run the ad or banner, but the amount of simultaneous users is not that important.
Comment by Jim -
Mark are you still into the Genarlow Wilson case?
I believe you are a man of your word and you are still trying..
Comment by MaNs -
Surely I am not the only one to read this and ponder one simply fact. Stats are always skewed. There is not an industry that does not skew the facts to their favour. For example, American Idol says they got a record number of votes for a certain show. So what? To me, their may not be more voters but more pre-teens calling in (since you can vote more than once I believe) for their favourite hunk of the year. It is similar to movies and how they quote only parts of the film critics sentence, ya know the one promoting the movie as the next best thing to sliced bread. Thank god I am not a huge fan of sliced bread. Truth is, statistics will also favour the person who is providing the statistics and the truth may never appear. This industry is following a long precedence set within our society, and as a consumer, or ad buyer, or whatever, we are getting screwed (not in the good way either). If the truth was always there for us to see, people would not pay accountants very good money to make things seem better than they actually seem. Our ignorance to not ask questions before committing to a sale is our error, not the company providing the statistics. You don\’t buy a car based on everything the salesman tells you so why be so stupid to apply this concept to paying for advertising?
Comment by Jeremy Rutledge -
The general market is what everyone is concerned about – the growth patterns / show that the Ethnic Market is the new area
or common sense area – why does no one ever mentioned the Ethnic area in any of these conversation – I own an Ethnic Ad Agecy –
there is a huge pie !!! think outside the box / the Ethnic Market grabs new trends like everyone else. I just want you to think
about that. I Have have the Ethnic Numbers – people lie numbers dont. Come up with benchmarkets in the ( Ethnic market ) that is what advertisers are not being shown
Comment by andy cyphers -
with regards to the example you gave, ads for a time sensative product, I completely agree with the \’simultaneous view\’ concept.
(ironically, youtube is the best place for time sensative ads, however i still think \’simultaneous view\’ is not needed. more so… x number of view before a certain date.)
And I was about to point this out to you, but you did it to yourself… \”Finally, the number of simuls give you a reference point to TV \”
here in lies the problem… the VOD concenpt is completely distructive to the \’tv\’ model. so why should the same metric be applied to the new model where viewing something is all about YOU (the individual) not the masses.
vod is for the individual (or group of friends), while TV is for the relatively untargeted massess. you can not use the same metric. The only caveat to this is during a live broadcast over the internet… then \’simultaneous view\’ would be very important.
Comment by echotoall -
Mark, are you talking user generated content or long form video? If you\’re talking UGC, then I agree with you. But, if you\’re talking about long form video, such as that on ABC.com and other major networks, then I completely disagree. The number of simultaneous viewers does not matter in long form video unless the campaign is targeted for a high number of impressions/views in a short amount of time, such as your movie ads example.
Though, with on demand, long form video, you can now spread that advertising out a bit. Right now the studios have to advertise on Thursday night because not many people watch TV on Friday and Saturday because there aren\’t any good programs on during those times; besides sports. Now, with on demand video, you can run that advertisement all the way through Sunday morning/afternoon and still help drive movie goers to the box office on opening weekend. Sure, you\’re going to hope that you get the bulk of your viewers on Thursday night, but you don\’t have to.
I\’m sure you\’ll argue that long form doesn\’t command enough simultaneous viewers to do that type of advertising. You\’d be right, at least for now. But, in the very near future (1 year or less) you will see TV quality and possibly even HD video being delivered over the Internet at costs that will allow the distributor to make significant profits with the new type of ad types that are making their way to the video screen soon.
Just look at the pause ad on ABC.com. Imagine the advertising possibilities with just that one new ad type. You now have a viewer that has disengaged from the program, thus more prime to pay attention to an ad. You can also direct that user away from the program, give them new info or have them sign up to receive more info or a discount (such as a $1 off a movie ticket), and then direct them back to the program that is waiting for them. Considering the average person, according to Nielsen, pauses 3 times during a 30 minute program, there\’s plenty of opportunity to engage the user into take action on an ad as long as the ad is targeted correctly.
Mark, I think you\’ll soon sing a different tune when it comes to advertising on On Demand, long form video. Especially when you figure out you can target your viewers better and earn more advertising revenue by doing it.
Comment by Tyler -
I am confused why it seems like aside from advertisers people just don\’t \”get\” the value of being able to broadcast your message to the highest concentration of people at the same time.
I\’m sure somebody in the advertising industry can quantify the value (to the advertiser) of reaching the maximum # of people at the same time with the same message.
This sense of \”shared experience\” seems highly valuable to the advertising community in terms of how products are received into the world.
One to one marketing is great — it is. But so is Coca Cola hitting everyone with \”I\’d like to teach the world to sing, in perfect harmony\” at the SAME time, in the SAME way. We know for a fact that Coca Cola highly values the ability to do this and will pay extra for the privilege.
Comment by Robert Seidman -
freegeorgemartorano
George Martorano is serving one of the longest federal sentences for a nonviolent offense in America. Is it because he wouldn\’t drop the dime on his late dad? Books. Screenplays. Poems. Set in places like South Philly, South Jersey and South Florida. Tender teenage love stories, tough Mafia crime tales, screenplays about prison life and essays about God and the afterlife
Comment by mark anthony given -
Couple of issues here: There\’s no reason for the advertiser to buy \”around\” certain kinds of content. The advertiser should buy the audience they want.
The shallow demographic metrics of \”what kind of viewers are these?\” around a specific web site – has already given way online with the growth of \”leads sales.\” Taken to final step – the attention span of the exact individual the advertiser seeks is worth an exact knowable amount. In the entire free market of attention span economics – everything is knowable and has a incentive / tendancy to expose itself and become \”frictionless.\” *shudders at the word*
Let\’s be clear… TV has been the most powerful for brand advertising and it does so at an exact knowable cost: about 3 cents per viewer. Now for the first time online, advertiser web sites (not banners or pre-roll) are better than tv commercials for branding.
Put it all together and I think you have obvious play even for companies like Goowy to simply just compete with TV.
Comment by Morgan Warstler -
I do agree with you to a certain extent, for anything time sensitive measuring the selling power of a website by the number of viewers at once makes sense, but for any other ads it doesn\’t really matter. On TV, we see a combination of both as well.
Comment by Ash Haque -
In the hierarchically organized multiprocessor system EGPA, which has the structure of a pyramid, the performance of concurrent programs is studied. These studies are assisted by a hardware monitor (ZHLMONITOR III), which measures not only the activity and idle states of CPU and channels, but records the complete history of processes in the CPU and interleaved I/O activities. The applied method is distinguished from usual hardware measurements for two reasons: it puts together the a priori independent event-streams coming from the different processors to a well ordered single event stream and it records not only hardware but also software events. Most useful have been traces of software events, which give the programmer insight into the dynamic cooperation of distributed subtasks of his program. This paper describes the measurement method and its application to the analysis of the behaviour of a highly asynchronous parallel algorithm: the projection of contour lines from a given point of view and the elimination of hidden lines.
Comment by wow power leveling -
Sorry Mark, I have to disagree with you.
Let\’s take Dallas Mavericks ticket marketing for example.
1. Advertising a home game for the coming weekend.
2. Advertising season tickets during the off-season.
Your model is relevant to one of these but not both.
For event marketing i.e. perishable inventory, I think your model makes a great deal of sense and why TV still works for such things as upcoming movie releases and Dallas Mavericks weekend basketball. You have perishable inventory with a limited window of opportunity so you need to reach as many people as possible for your mass-market product. As an advertiser, you should be willing to pay more for this kind of advertising because of the time-sensitivity and my ability as a video content site to aggregate the eyeballs you need.
In the 2nd example, you have 4 – 5 months to fill-out those season ticket seats so while inventory is perishable its not critical. Also, the consumer of a high-end luxury item such as Mavs season tickets may not respond as well to a mass-market media message with a long lead time. In this case, the sort of advertising necessary is more skewed to finding the quality of consumer. In this case, I am willing to accept 1,000 every day for a 1,000 days to meet my goal if I know I am getting quality consumers. Its the reason the Harvard Business Review can command premium pricing even though it has a smaller circulation then Forbes.
Which leads us to the ultimate question…Are video content sites geared more towards this event or more personalized marketing. I think we can safely assume a more personalized marketing approach. The Internet and content sites are catering to a specific defined niche audience (the long-tail if you will) and should be priced as such. Why would you pay for the spill if I could deliver an ad to 100 qualified, prospective season ticket holders without sending it to another 99,900?
I cannot foresee with our current broadband deployment in the US video content sites being anything more then just supplements to event marketing campaigns. However, video content sites are the best place to develop on-going, involved, and intensive relationships with consumers – something TV fails at on both a costs and execution basis.
One final note – While video content looks like TV, its not TV and should not be treated like TV unless its trying to command those premium dollars then I say break out those max and average numbers and compare. However, be ready to do the inverse when a TV program is trying to cultivate a long-term relationship and gauge actual conversion from those TV spots.
Comment by Grant Case -
you\’re right, but i think something infect the development of internet business like trust between sellers and buyers.As for online ads, I don\’t pay any attention to them at all. I do make note of the really obnoxious ones that manage to get past my blockers. The note I make is to never buy that product or service.
Comment by wow power leveling -
hasn\’t the problem as far as advertising is concerned been that the networks and the statisticians have long maintained an obfuscatory method of taking/serving numbers in order to preserve the integrity of the current advertising rate model?
it just seems to make sense that better numbers across the board are in the interest of the advertisers and less so the providers/hosters of content. That would then lead one to believe that the providers who have mindshare and detailed unbiased and perusable metrics would have a competitive advantage in the marketplace at least as far as attracting advertising dollars.
that said, are there numbers that indicate the effectiveness of online adverts? working in film and music I\’ve become aware of the power of product placement – especially in music videos. For many artists – the advertisers are really getting over – as they are getting prime placement for pennies on the dollar for multiplatinum selling artists – usually just enough for the artist to offset some of his/her video costs. those would be killer metrics: maybe studies need to be done – first using controls – what kind of adverts work best online? Then advertisers can spend smart dollars – and the system becomes a bit more efficient and transparent.
Comment by blyx -
Maximum simuls won\’t be the most important to advertisers for long. I would bet this is just important right now because they aren\’t used to counting views or cpm advertising. If the ad is time critical obvious the situation certainly changes and the simultaneous viewers is the only reasonable way to measure their exposure.
Comment by loaders -
How can I get email or phone number of Mark Cuban?
Comment by Michel Drouin -
Is there a good reference point to begin with estimating and ultimately tracking simultaneous viewers? Thank you
Comment by Mark Chew -
Of course, the problem with TV now is Tivo and DVRs, but you make a great point about advertising online. If companies can track the number of users it would be more reliable than TV (and Nielsen Ratings) and advertisers could probably even charge more for this accuracy.
Comment by ac sports -
Interesting idea, but it avoids much of the possibility of truly targeted video ads, and Tivo-like timeshifting altogether. Would it be more valuable to have this \”real time\” advertising ability, or for a Comic Convention to deliver an ad to 10k single men, 18-40, who live in the area and enjoy SciFi programming? Or tell 100k people who enjoy murder mysteries that a new book is coming out? I\’d much rather pony up $4/1k unique viewers when I KNOW they might to be interested, rather than a 30 second 1am spot when I don\’t know who will see it – and you still might only reach 17 viewers then.
While the \”something dramatic\” ad is interesting, that does not reflect the reality of what we see on TV now. Ad campaigns shifted from being planned out months, or years in advance. Chevy advertises on Comcast\’s on-Demand network, and I don\’t believe they would care about simultaneous views.
Comment by dave -
Definitely some good thinking, and there\’s a lot to be said for the maximum-simultaneous metric for advertisers in the mass market, like the movie example you cite.
However, there are plenty of smaller market advertisers who are more interested in long term exposure than huge swaths of simultaneous viewers. I guess I\’d argue that the more niche a product is, the less relevant simultaneous views are, and conversely that the more mainstream/mass market a product is, the more simultaneous views matter.
And, to date, almost all internet advertising has been niche. Well, niche plus mainstream companies just throwing money at the \’net lest they be left behind, whatever that means. But you\’re right that, as internet advertising goes more mainstream and wants to attract more mainstream advertisers who actually want to see ROI, simultaneous views will become a more important metric.
Comment by Brooks -
Very insightful look at how users can look at different architectures in advertising and how to approach media in numerous ways.
Comment by AMERICO -