JD on EP

H.264 + Flash = TheFutureOfVideo
Take a look at this sample, TV serial in HD at only 500Kbit/s!!!
http://flashvideo.progettosinergia.com

Wow, 4500 can be extrapolated to represent all computers in the world. Amazing.

John,

I have to object... as Jason said there's no way you can extrapolate like this .. I just checked the google analytics data for flexbox (http://flexbox.mrinalwadhwa.com) .. it has had close to 40,000 visitors since the release of Flash Player 9 Update 3 on December 3rd 2007 ... now the audience of flexbox are professional flex developers who can be presumed to be early adopters of this update especially because of the flex related features ... even with this early adopter demographic I see only 35% people have Flash Player 9.0.115.

Please don't make such claims they reduce the credibility of even the genuine claims that people at adobe make about flash player penetration.

_
Mrinal

Ours is 72% still on version 9.0
9.5% on version 9.0 r115 (ranked 2nd).

As other commentors have mentioned, "real world" adoption on web sites is far lower than 62% of your small sampling. I see adoption rates in the 15% to 20% range.

"That's 62% of today's computers" No, that's 62% of the 4500 people you had surveyed.

You talk trends, but you did not mention trends you spoke in absolutes (62% of the worlds computers support 9.0.115) based on a small sample size of a user that is registered as a lightspeed panelist. Sorry I don't equate this with absolutes.

I’m not convinced of your pitch here. I understand your motivation and I get what the Flash Survey’s execute upon, but you’re fumbling the ball here a bit.

For example - Could you elaborate on the part where you mentioned cross checked against server traffic? Who owns the traffic and how are they checked and what are they checking? (the link to that methodology link doesn’t provide this break down).

[jd sez: "Cross checked to server traffic" means checking against the daily successful installation rate, tracked by an anonymous ping to the Adobe servers when a new Player installation first starts up. It had held steady at about five million successfully completed installations per day for a few years, then increased to 8M/day early last year, and at the end of last year jumped to 12M/day (I've got links in archives). Over the past quarter we saw a very significant jump to a new plateau level. Server activity matches consumer audit results.]

What I do understand is that 4500 various persons are surveyed at certain periods of the year, and these persons have to elect to opt-in (which I believe will skew the statistics in terms of candidate selection). They must also be 18+ and only reside in certain regions of the world and some of these regions are assumed to be one in the same. An example comes to mind where 400 are selected from China? Only 400! Given the population density inside China this is really limited amount being sampled.

Not to mention that Oceania is poorly represented as to assume Japan/China speak for Australia and New Zealand is completely flawed in theory and practice.

[jd sez: Yes, check the methodology for more details on how consumer survey panels are formed. The smaller sample within a region may not suffice to say "67% of people in China have the latest Player", but it does help make the entire group of 4500 more representative. As I've insisted in previous quarterly audit discussions, regional samplings would necessarily have a smaller confidence level than those regional contributions to global samplings.]

Vietnam is currently one of the fastest growing technology regions in APAC right now. They aren’t being accounted for.

Last but most important of all, these do not account for corporate or below 18+ consumers. MySpace isn’t run by 23+ consumers and I think this data is not being factored into the equation accurately enough. Corporate firewalls aren’t even being measured in this regard, so in all this sampling of consumers is purely a best guess backed by a methodology that really no person on record has openly challenged.

[jd sez: Yes, as discussed in previous consumer audit conversations, general consumer norms would not predict *your* particular audience. These overall consumer trends are used to buttress and confirm your own individual server logs.]

Ok, as Jason stated, this is trends not absolutes. Admit it, you got caught doing what all Adobe staff have been doing for many years, you made us think that 62% of todays computers have Flash 9.0.110 installed.

You didn't however state that all trends based off a weak methodology that has more holes in than a piece of swiss cheese - indicate that 62% of the worlds computers today are likely to have Flash installed.

Keyword being likely.

I'll admit, I use the 95% of the worlds computers have Flash spin as much as the next person. I've also known about the holes in the Flash survey for sometime, but now that you have put the 4500 out there in the public eye, you've just hurt Flash.

As now just as easily as everyone saying 95% of the worlds computers have flash, the competition is now going to say - no 4500 of people sampled around the world had Flash.

There is no comeback for that, and sure we could argue of methodology but that argument is lost before it starts.

I agree with Mrinal, shut the hell up. [jd sez: Nope. ;-) ] We aren't stupid so stop being arrogant and making out in your Update (talk about bad blog ethics, stick it in the comments like us all and less hijacking/controlling the conversation) we are all clueless because you feel we didn't get the way the Survey works.

We do, we just don't buy it and there is a difference.

I develop flash for a site that gets about 100,000 visits per day and it has flash on it. We've discovered that the 115 rate of adoption is about 40-50%. So the figures are not that far off. I think the point is that specific publics adopt it more. People who watch video probably are all updated, etc. Youtubers, etc.

Im actually surprised at how fast it was adopted, It's absolutely awesome.