Friday, January 25, 2008

Flash Penetration and Bullshit Statistics

I started this blog today because I don't have any forum for my musings on technology and other stuff. I'll probably keep this blog eclectic, limited only to my non-Tango rantings.

I'm trying to figure out what platform to use for website designing. I've done a bunch of Flex programming recently (see www.kbaseweb.com) so the natural choice is Flash and Flex (or Flex and Flash, or some combination).

But it's time somebody looked at the issue of Flash "penetration" seriously. I'm sure somebody has, but the only thing I find on the Web is the much-repeated idea that "Flash is installed on 98% of browsers." The Wikipedia entry says that there's no other data than the official Flash stats, so we are left with the official survey.

Is this survey valid? Let's start with the idea that it is probably not and then figure out why (scientific, no?). Because if we could ever get at the Google numbers for Flash usage, then we'd know the truth. But instead...

First off, how is the survey done? Well, users get an email asking them to participate in the survey. Wait, you mean they use junk mail to get people to participate? No. They send email to their users. Wait, users of what? Users of Lightspeed are the possible Universe of respondents. Lightspeed is one of those sights that gives out small cash prizes for participating in surveys.

First off (for the second time!), how does that universe of respondents relate to the total population of web users? Well, we can take some guesses. Perhaps the users are younger or older than average users (probably both: would you fill out a survey for $10 if you're working a 100K a year job? A 50K a year job? A part-time summer job which pays you $7 per hour?)...

How do people find out about Lightspeed? At places like Cashsurfers.com, Volition.com and other coupon sites, which route you to LightspeedPanel to get cash for your time. Again, only a small proportion of Web users see the sites that refer one to become a Lightspeed memeber.

Here's the official blurb:

Panelists are recruited from multiple sources such as RDD, in-person interviews, Web partners, as well as banner ads. 100% of Lightspeed Research’s panel members have opted to participate in the panel.


But let's say that this is all crazy and that EVERYBODY is checking out Lightspeed.com. In that case the page would be very popular and would have a lot of links to it. What is the current Google Pagerank of the join page found here? 3 of 10 (same as the infinitely famous TheKBase Web). And what about the LightSpeedPanel home page? Same: 3 of 10 (dismal, if the Web is a popularity contest).

Also, although one can join LightspeedPanel without using Flash, the LightSpeedPanel home page contains a Flash movie! The means that the users who have logged in at least once have already downloaded Flash. So when they get the email to take the Flash Penetration survey, they have already had one encounter with the Flash download thanks to Lightspeed itself!

Anyway, do not believe the hype regarding Flash penetration. The only people who know the real answers are Google, Yahoo, or basically any big site (any Webmasters want to share?). But 98% if a fantasy statistic if there ever was one. Check out these numbers, which might be more realistic.

Personally, I think the proof is in the pudding, as it were. Flash apps are OBVIOUSLY way better than Ajax apps (anybody who is not seriously drugged can see this instantly). So why doesn't Google use Flash for Gmail or for the Google search page (although check out SearchMash in Flash from Google to see that they are playing with it)? Because not everybody has Flash.

Some people don't have Flash because they didn't get it. Other people don't have Flash because they are behind corporate firewalls that disallow it. Other people don't have Flash because their Flash install got screwed up somewhere along the way (usually upgrading to version 9) and they will not have Flash again until they reinstall the OS or switch to another browser.

What do the real statistics look like? Well, for Flash penetration I'm not sure. But if you look at some Browser-version sites like this one, you'll see that almost 50% of users are still using IE6. Now I don't know how reliable those statistics are nor what that means, but I'll bet that you'd be mad to be programming in Flex 2 or Flex 3, both of which require Flash 9, unless you have some control of the audience ("you wanna use my site? then you'll install Flash 9!").

Don't believe the hype. Flash is big, but it's not Google. And Microsoft knows that and is trying to get into that niche with Silverlight and their always famous hegemonic distribution techniques.

I can't wait for 2050, when all of this will be history and your browser will be your broom or maybe your fingernail.

No comments: