Here's a rant: I hate the cute stupid videos on startups' homepages. Look at this brilliant article by DJ -- he basically says that your product can't have any "mental overhead" -- meaning, if your UI and copy isn't clear enough, no one is going to watch your silly video. Love this. This is my interpretation, of course, but engineers and designers often overlook this reality.

Much as I want to agree with you, the folks at Dropbox swear by their homepage video.

It's a recommended Growth Hack: http://pandawhale.com/convo/3396/growth-hack-make-a-video-not-a-tutorial-tour-or-tooltips

Another example: http://blog.jasonshah.org/post/18921010795/lock-down-a-paid-subscriber-on-visit-1

Well, Dropbox has reached a certain scale where the intended user (at this point) has probably already heard of them, right? Maybe I'm just reacting to seeing so many of these for startups with very small user bases. Good point.

Conventional wisdom (READ: Paul Graham) has told these startups that it's the right thing to do, because it worked so well for Dropbox and Airbnb.

As with most things startup, it's bad to copy what others have done. Only the originals make it big.

No one is impressed by the checklist item: "a video". It has to be good.

I tend to prefer screenshots and great blurby copy, but if a video is really good, I appreciate it too. And there is certainly a large portion of the mass consumer audience that *prefers* a video.

Even if they prefer videos I'm not sure they actually watch the videos.

Videos spawned a new acronym TL;DW.

I think a well written comic strip would be better.

Comics are great too! But also remember many people are slower readers or even more auditory than visual.

They'll find friendly, well-paced voices, with measured and appropriate emotional intonations, much easier to comprehend than walls of web text and images.

(With the right UI for offering and controlling web audio narrations -- *not* strictly screen-reading, but optimized for the sighted but speech-preferring -- it might be possible to have a parallel web that's better for listeners.)

A web for listeners is a fascinating idea. Like books-on-tape ... Er, podcasts. :)

Yes, though in smaller interactive (and almost conversational) clips than audiobooks/podcasts.

Perhaps Qwiki's audiovisual reference presentations, as they'd originally launched, were one hint of the possibilities. (Qwiki appears to have discarded that original offering to pivot to a mobile-presentation-content-creation offering.)

There's a company called VoiceBunny (beta) that translates text to spoken word with very high quality. It's a plug-in for blogs and content sites. Fred Wilson uses it.

There's also SpeakerText: http://speakertext.com/

No, wait, that's reverse. Voice to text. Darn.