|
On February 20 2012 14:10 Primadog wrote:Show nested quote +On February 20 2012 14:02 MrCash wrote:On February 20 2012 13:55 Primadog wrote:On February 20 2012 13:50 MrCash wrote:On February 20 2012 13:48 Primadog wrote:On February 20 2012 13:47 MrCash wrote:On February 20 2012 13:42 Primadog wrote:On February 20 2012 13:33 MrCash wrote:
Did this in a couple of minutes, but someone weigh in if I'm not doing this right.
2412.88/956 = $2.52 (Earned per ad)
2.52/7615 = $0.00033 (average earned per ad through the duration of the showing)
If I got my definition of CPM properly, the $0.00033 would be the CPM in this specific scenario and it's not what has been touted by any streaming companies.
Or if CPM is per 1000, then it would be $0.33, which makes a lot more sense.
Still is pretty low, as normally we hear numbers ranging from $2-$4 thrown around. This is an incorrect way to calculate CPM or eCPM. The error of this come from double-counting ads from rebroadcasts. While 956 ads may been played, most of it was played during rebroadcast time, which have much lower concurrent average than the live broadcast. Therefore you're artificially deflating eCPM. Much more accurate measure would be viewer-hour per ad dollar, which is roughly inline with the first-week analysis FXOBoss produced. How is that not eCPM? Do TV ads pay less when they run the show 3 days in a row when the same people are watching? If the ad service providers gives less revenue for running more ads to same viewers, they are either not providing enough diverse ads or are insufficiently filling the ad demand. Those numbers are based completely of total revenue and average concurrent viewers. Viewer-hour per ad dollar ad would serve no purpose and would by definition NOT be eCPM or CPM. Number of ads run per hour and number of viewers hour to hour would vary as well. The trick here is that there aren't 7615 people on average watching both the live and rebroadcasts. How do you know that? They said average concurrent viewers and total ads run. If the rebroadcast numbers don't include average viewers, than maybe the rebroadcast ads are not listed either? That seems like a strange assumption to make. I enjoy tracking data, so I have an idea of the general range of where FIS#5 runs on average. It's unlikely that 7615 number represent an average that includes rebroadcast when this is their best day: ![[image loading]](http://i.imgur.com/r33Ct.png) Note that the http://pe.nitrated.net/ tracker refreshes at 3minute intervals. Hence the instanenous concurrent peak would be higher than the tracked concurrent peak. Consider, for example, if you have 8k people watching live, 4k watching the first rebroadcast, and 2k thereafter watching the 2nd rebroadcast on the average. the 956 ads break down evenly to each of these three time slots. Therefore: 2412.88 / ( 8k * 956 / 3 + 4k * 956 /3 + 2k * 956 / 3) = 0.00054 or a eCPM of $0.5 per thousand. Not great, but at least closer to the quoted CPM and known fill rates. While a very pretty chart, it's still a lot of assumptions. Assuming that average numbers for rebroadcasts from one day of views. Assuming the total ads run is actually total for everything, while other numbers are exclusive to certain times. Either we can draw conclusions assuming the numbers given to us are accurate or we can't draw any conclusions at all. The more assumptions you ad to the equation, the less valuable the conclusion. From what you are showing there, that seems to be rather accurate for that one day, but that's as far as I can reasonably agree. ... You do realize that the the chart roughly graphs a single live broadcast, and that the 5k leveling off is exactly the first rebroadcast period right? Again, this chart comes from the best day of FIS#5. My example simply shows how to properly calculate eCPM, which demonstrates why your calculation is wrong for every case except for when live viewership = rebroadcast viewership.
Ok, so you understand that the graph does in fact not display the entire duration of the showing and any conclusion you draw outside of it are assumptions.
The numbers given to us by FXOpen are the most reliable we have available.
|
The numbers I have given are direct from twitch. My calculation was similar to that of the first person to do the cpm/ecpm rates. However there are external variables. That being said the raw figures are that of the first poster.
|
Thanks for being transparent with this information - we appreciate it.
|
One thing I would like to commend FXO on is the production value of this whole event.
Aside from a couple very minor hitch ups, it was really well done and organized, everything was very smooth and nice.
If you guys ever try anything in the future, I would make a recommendation to ad some personality to it.
Even though you guys had higher caliber participants than virtually anything other online tournament out there, the broadcasting and analysis, while clean, was rather bland.
There is a lot of appeal and attachment created from casters being on camera in between games (there were a couple videos by FXO members promoting Razer gear I believe, but that was it).
Some sort of introductory animations or player statistics would be very beneficial as well.
Of course there are many factors, all of which hard to account for, but I do hope you guys manage to find someway justify airing more of similar caliber games, regardless of format.
|
Thanks for the information.
One question I have is how will the ads on the VODs affect the total income? People are going to be watching the vods for weeks/months and should help with the revenue. But from what I've seen here it won't come close to the cost of running the event.
The current state of esports seems to be one of tourneys/events being bankrolled by venture capital or parent companies with deep pockets hoping for a viable business model sometime in the future. The real issue is that esports is dependant on the game titles - as SC2 ages the viewership will naturally drop.
So really the only viable business model will be to increase the amount of money from viewers (PPV style), which is most likely not going to work, or to cut the costs of production. Most likely in the future you'll see smaller prize pools to compensate for the costs. Personally I think for at least the larger tournaments they need to cut the budget for casters - I remember the djwheat reddit thread where he said the big time casters are making more than the players. That is simply not viable. IMO thats the first thing to look at.
|
I don't believe Twitch pay any VOD ad revenue. This is the reason why Day9 Daily and OneMoregame.tv hosts their archive in Blip.tv, which offers similar revenue share arrangement (50-50, same CPM but higher fill rate) as Twitch.tv has for streams.
|
On February 20 2012 14:33 MrCash wrote:
One thing I would like to commend FXO on is the production value of this whole event.
Aside from a couple very minor hitch ups, it was really well done and organized, everything was very smooth and nice.
If you guys ever try anything in the future, I would make a recommendation to ad some personality to it.
Even though you guys had higher caliber participants than virtually anything other online tournament out there, the broadcasting and analysis, while clean, was rather bland.
There is a lot of appeal and attachment created from casters being on camera in between games (there were a couple videos by FXO members promoting Razer gear I believe, but that was it).
Some sort of introductory animations or player statistics would be very beneficial as well.
Of course there are many factors, all of which hard to account for, but I do hope you guys manage to find someway justify airing more of similar caliber games, regardless of format.
It was the cards for this event, but there were technical issues with the sound that made it unworkable. It's slated for our next one. It really just comes down to manpower and cost for those things to be done.
|
I'd like to know why was the cost so high at 8k? What was exactly purchased with that money Is it possible to cut production to lower the cost to approx. 2k?
|
Staff who worked on the event:
1 tournament manager
3 casters
1 graphic designer
Thats 3k total costs for a month. Extremely cheap if you ask me.
|
On February 20 2012 15:28 PolishxThunder wrote:
I'd like to know why was the cost so high at 8k? What was exactly purchased with that money Is it possible to cut production to lower the cost to approx. 2k?
Prize pool is $5k, so that leave $3k to production. Note that FXOBoss originally stated his budget as $10k overall, or $5k in production.
|
On February 20 2012 15:30 FXOpen wrote:
Staff who worked on the event:
1 tournament manager
3 casters
1 graphic designer
Thats 3k total costs for a month. Extremely cheap if you ask me.
I would just like to commend you on the amazing graphics for the invitational. Whoever did those was amazing.
|
Then lowering the prize pool is the answer. Stream numbers are actually fairly decent imo so I don't undestand how you could increase revenue.
|
On February 20 2012 15:42 PolishxThunder wrote:
Then lowering the prize pool is the answer. Stream numbers are actually fairly decent imo so I don't undestand how you could increase revenue.
Increasing the value of advertisements is the best response. Its entirely possible to do that. Hopefully I can.
|
On February 20 2012 15:00 Primadog wrote:
I don't believe Twitch pay any VOD ad revenue. This is the reason why Day9 Daily and OneMoregame.tv hosts their archive in Blip.tv, which offers similar revenue share arrangement (50-50, same CPM but higher fill rate) as Twitch.tv has for streams.
You get ad revenue on Twitch VODs because of the preroll. The reason Day9 has VODs on Blip is because ad revenue from Blip is MUCH higher than both youtube and Twitch (ok, maybe MUCH is an exaggeration, but it is undoubtedly more). The problem with Blip is that it doesn't get the same amount of broad exposure... but for someone like Day9 it doesn't really matter, he'll get the views wherever his vods are posted, so he goes to Blip for the extra revenue
|
Ah, that explains why preroll numbers always look a bit funny.
|
i feel like its a bit late (time zone wise), youd get more viewers if it was maybe 2-3 hours earlier
|
this was a great event and im glad i subscribed for it. One thing i noticed when i bought my pass is that i needed to register a credit card with paypal, i have never seen this before i had plenty of money in my account. I'm sure this would have turned some people away.
|
I watched a good bit of the event live, and it was a great tournament.
Like others have said, I really think that they should consider a time-slot that is better for NA/EU viewers. There are tournaments with MUCH lower caliber players/matches and no better production quality that garner better viewership because of the timing. I don't think a better time slot would immediately bring the invitational to break-even, but I think it would help immensely.
A question: is there any way to know what the breakdown of viewership was by geography? And, if so, how did it compare to say, MLG?
|
Thanks Boss! Very informative and nice post!
|
Why NA/EU?
EU wouldn't have had a problem with the timeslot they could have sat down to a lovely Saturday of watch Sc2. If they wanted to watch live. or watched it at night if they where happy with a rebroadcast. Sure the Live broadcast was a bad time for NA but over here in the future I have to put up with have my MLG finals on a Monday morning.
There is never going to be a perfect time-slot, and I don't think pushing the Western Pacific side of the world to have to take a sicky or "Watch at Work" to see some awesome SC2 is a good thing to do.
|
|
|
|
|
|
EPT
