Semantic advertising. [...] Since conversations on [Google Wave] waves have to go through the server each time, a semantic engine could parse them on the fly and serve up relevant text ads. With enough data and training, a semantic engine could decipher intent, i.e. whether you’re talking about your trip [...] last summer, in which case ads would be useless, or whether you’re setting up a wave to plan a trip [...] with your friends, in which case ads for cheap flights and hotels are relevant.

Intent is the reason why nobody clicks on ads in social networks but they do in search engines. A semantic engine would know that 99% of the times you’re waving an ad would be irrelevant at best. So 99% of the time people wouldn’t see ads at all. Wave, Inc. might set this up on their own servers and allow others to set it up on theirs under a rev[enue]-share agreement.

Over the long term, Wave, Inc might also open an ad network [...]. This would be good for consumers since they would get few ads, and only relevant ones, good for advertisers since they’d get high clickthrough, and good for Wave, Inc, since they’d have a high quality, expensive inventory. This might be the thing that makes online ads something other than punishment for using stuff for free, but actually something useful and exciting.

Problem: semantic technology is still inchoate and execution would have to be flawless for people not to find it annoying and/or creepy.

Read that 2nd to last sentence in bold again. Therein lies the crux of the failure of online advertisement in most areas other than search ads.

For more proof of how badly current ad models are failing, witness the stats that MediaPost Publications just posted in ComScore: Most Clicks Come From ‘Natural Born Clickers’ 10/02/2009:

Indeed, the number of people who click on display ads in a month has fallen, from 32% of Web users in July 2007 to only 16% in March 2009. Worse still, an even smaller core of consumers — representing just 8% of the Internet user base — accounts for the vast majority, or 85%, of all clicks.

Now while the strong "few users generate most clicks" imbalance is predicted by the 80/20 Principle, the reduction BY HALF in less than two years is indeed stunning, and cannot likely  be explained away with the recessionary economic background alone.

It is far more likely that users have systematically trained themselves to mistrust and hence to simply ignore online ads altogether (other than MAYBE search ads).

The situation of internet users ignoring internet ads is apparently becoming so dire, that comScore et al. are beginning to sound like TV advertising execs in their pitch to ignore the bad news (from the same post):

"Marketers who attempt to optimize their advertising campaigns solely around the click are assigning no value to the 84 percent of Internet users who don’t click on an ad," said Linda Anderson, comScore VP of marketing solutions and author of the "Natural Born Clickers" study. "That’s precisely the wrong thing to do."

Rather, as comScore research has shown, marketers need to embrace the fact that non-clicked ads can also have a significant impact on consumers.

Basically, the argument goes: Ignore the fact that no one is clicking your ads anymore, those ads are still somehow reaching consumers on a subconscious level, and will (magically) continue to build your brand (and ultimately sell your stuff) that way.

When the Internet guys are telling you to forget "measurable" and instead extol the virtues of Madison Avenue-style image advertising, you know something is afoot.

Now it is fair to point out that people do in fact retain contents that they saw even if the "seeing" never reached the level of top level consciousness. As such, in a recession, one might argue that there could be value in a company advertising simply to "ping" the consumer in a "we’re still here" sense? (Especially if the collapsing ad rates make it cheaper than ever to do so…)

But now remember the phrase from the first quote: "…the thing that makes online ads something other than punishment for using stuff for free". Then think about how sure anyone could be that their mere ad presence actually engenders much love from anyone at this point.

Rather than ask themselves the hard question on how to finally make online ads work, online ad stake-holders give you evasive platitudes. That’s simply not enough.

But what makes Google’s new "sponsored videos" feature on YouTube even more relevant is today’s news that YouTube searches now represent the second largest search engine in the world according to ComScore, ahead of both Yahoo and Microsoft’s MSN/Live! So there should be ample room for YouTube to generate profits for advertisers and in turn for itself (Silicon Alley Insider estimates that it could add $1B to Google’s bottom line).

However, as I began to lay out yesterday, there are a number of caveats that need to be kept in mind by the internet marketer looking to take advantage of this opportunity:

1) Marketing within Social Media (vs. search ads PPC) is generally tricky due to a deeply rooted differentiation by most people between social and business contexts: People don’t like them mixed, and can react very negatively if they are (read Dan Ariely’s excellent "Predictably Irrational", chapter 4 "The Cost of Social Norms").

2) So if you are going to market in any social context, you need to get the tone and the context just right, else you are not only wasting your ads, you are likely hurting your brand. The backlash may also be much stronger than in other situations, because you will be dealing with a perceived violation of social trust.

Whatever initial offer you make needs to still fit into the "friends" context somehow, or else be so targeted that the prospect truly sees your offer as a form of "friendly service", e.g. if you are offering something that would help with a social task they are about to undertake, like offering flowers at a special price if someone is surmised to be going on a date, etc. (judging from e.g. a Facebook "action" of theirs).

3) While YouTube is overtly the least directly social (compared to say Facebook, etc.) and instead more entertainment oriented, the social aspect of sending/receiving video clip links to/from your friends is still clearly there. So to stay in tune with the viewer/prospect, you still need to get the CONTEXT just right:

If the search keyword (or individual video for that matter) is an entertainment vehicle first-and-foremost, then offer them more (hopefully related) ENTERTAINMENT products, NOT shoes or cars or deodorant. This goes for pre- or post-roll ads as well by the way, which prospects tend to gladly view IF they have something to do with the actual video content requested.

With more educational keywords/videos, there may be more latitude to offer things, though they still need to be related and represent a LOGICAL follow-up, else your sponsored video will get largely ignored/filtered out by the prospect just like most other ads (even though, as I said yesterday, Google appears to be embedding the ads very discretely, so that they don’t scream "ad" vis-a-vis the other video content).

So the formula would be, create videos that are highly relevant to your keywords, while also being disruptive enough to get attention.