The Savior Of Web Advertising?
John Battelle -- my friend, former Wired co-worker and founder of the late, great Industry Standard -- writes in this month's Business 2.0 about the putative savior of web advertising: contextual advertising. John's as smart as they come in understanding media and the Internet and I hope his prediction is correct. But my own experience and understanding of the web consumer leaves me doubtful.
Consumers have one of two modes when they visit websites: browse or search. In browse mode, they're surfing from one site to another, maybe reading, maybe being entertained but generally not stopping in any one place too long. In search mode, they're looking for something specific and have an end goal in mind. Most (but not all) browse mode traffic occurs off search engines and much (but not all) search mode traffic occurs on search engines. Search type traffic also occurs on product review sites and comparative pricing engines.
I would argue that contextual advertising of the type John discusses in his article is extremely effective when a user is in search mode but a needless distraction in browse mode. So in the example of a user reading a product review, contextual advertising for additional information about the product makes sense. But that type of targeted advertising already exists: in five clicks through car review site Edmunds.com I viewed ads for Toyota, Chrysler, Mitsubishi, Dodge and nothing else. Contextual ads on a Slate piece, where the user is browse mode, will not result in anything close to the targeting effectiveness (or desired click through rates) that an Edmunds would see with a car advertiser.
In other words, websites with a product review focus in interesting verticals (such as cars, technology or travel) with large amounts of endemic advertising are already successfully mining those prospects and were among the earliest to hit profitability with the ad model. And search engines have successfully been using contextual advertising (originally known as a keyword banner buy) to redirect searchers away from search results and to advertisers' websites. But I don't believe contextual advertising will be the savior of everyone else. I hope I'm wrong.
Categories
Consumer Internet & Media3 TrackBacks
Listed below are links to blogs that reference this entry: The Savior Of Web Advertising?.
TrackBack URL for this entry: http://ventureblog.com/cgi-bin/mt/mt-tb.cgi/79
Lubasin, et kirjutan tänaseks otsimootorireklaamindusest. Peale praeporgandit ja -paprikat mehhiko kastmega on küll uni peal, aga mis seal ikka :) Nagunii Helen juba kommenteeris, et "vahel sa kirjutad nii pikka ja tarka juttu"... ma juba hoo... Read More
Andrew Anker considers contextual advertising: "Consumers have one of two modes when they visit websites: browse or search... contextual advertising is extremely effective when a user is in search mode but a needless distraction in browse mode." I... Read More
VentureBlog: The Savior Of Web Advertising? Consumers have one of two modes when they visit websites: browse or search. In browse mode, they're surfing from one site to another, maybe reading, maybe being entertained but generally not stopping in any... Read More
Actually there are three modes. The first two ones, as you correctly indicated, are surf and search, but there is a third one, where contextual advertising actually can make sense: Transaction. You will find this transaction mode in different variants, one of the more common ones would be e-banking. The difference between the surf mode and the transaction mode is that you know what you want to get done before heading to the web site. It is not an open-ended surfing experience or search mission.
Christoph... I wonder if contextual advertising would be any more effective in that instance. If you are directed towards a certain goal (for instance, buying a stock via your online stock account) you have a mission (as you say, "you know what you want to get done"). In that instance, clicking on an advertisement would only take you away from that mission. And more likely than not, the company providing the service makes more money on the transaction you're intending to do than they ever could on the advertising placement.
From my own experience, I would have to agree. I regularly follow adword links from google's search engine, but when I have seen them on sites like Slashdot or SFGate.com my personal CTR has been much lower.
Also, I would have to ask if Google/Overture/etc. are going to let people bid different prices for contextually-placed ads versus keywords. Personally as a buyer I'd pay more for search keywords since I know the user has typed in that phrase and is definitely looking for it, whereas if they are just looking at a page with the keyword in an article, the user may be somewhat less interested and therefore less likely to buy my widget.
J.Nielsen mentions context advertising on search engines. It actually works.
When I search something on Google I usually look at the right side of the results page. Since those small context advertisings are sometimes accurately targeted, they often contain exact info that I need.
Andrew -
My take is that it all depends on what the context of the information is. Today, the only “contexts” that have been defined are:
a. A link to an advertisement, which is disproportionately noise over signal from a contextual richness standpoint.
b. Links to specific web SITES or deep links to specific web PAGES.
If you add additional contexts to the model – business listings, product listings, user profiles – suddenly a great article on a trip to New York has very specific contexts that can be called upon (“gee, that restaurant sounds great, I’d like to bring up the listing in a single click and if it looks good, add it to my rolodex,” or “that briefcase the writer was talking about sounds very cool, I’d like to bring up the product listing in a single click, and if I like, add it to a wish list or buy it.” And on the pay for performance model, each of these actions is monetize-able.
This gets to a second point, available actions. As much of a quantum leap as the Googles and Overtures are over prior context-based advertising models, the fact remains that at the end of the search for “red nike shoes” is an algorithmic search return of very high quality and an advertisement of (often) lesser quality. That such a value proposition is going to generate $2B+ in industry-wide revenues, and 50% of Yahoo’s cash flow in 2004 tells you how low the bar is. Think about it: the purchaser of the key word may be a fantastic operator or a scumbag and the link might take you to a web site, a product page or something entirely different. The quality and specific context is un-defined and you have to visit the web site or specific page to find out. As such, it may serve the advertiser’s purpose without fully satisfying the consumer’s. Providing a deeper context in the form of a very structured product, business or user profile is part of the solution, but so too is the ability to act on the contextual ad "out of band," as a separate channel from the action of searching and browsing. Amazon gets the importance of this, and that’s why they provide you a wish list, but you must be on Amazon to take advantage of it, and you have no way of associating the articles, reviews, price comparisons, etc. that lead you to add the product to the wish list in the first place. Similarly, Expedia will provide you a map of hotels in an area you are thinking of visiting, but again, you have to be on Expedia, and their view is limited to hotels. In the course of my web travels, I may come across lots of interesting places that I would like to visit, and “drop” on to such a map, but today, I can’t. These are just two examples. Needless to say, there is also no way to organize these nuggets of relevant information into specific activities of interest, as anyone who struggles with the decision of whether to bookmark, email or save to a folder can attest.
Disruptiveness to the user’s workflow is the third element in extending contextual advertising’s reach. One of the biggest deterrents to clicking on an advertisement besides (pick one: pop ups, spam fears, irrelevance) is disruption to a consumer’s workflow. How many times does something look potentially interesting but clicking on it either takes you off to another destination (and away from the view you were browsing) or launches another browser window, which muddies the desktop while starting the countdown to the next time IE ties up your system? If these activities were non disruptive to workflows or at the very least, meaningfully less so, then I would be much more inclined to populate my wish list, rolodex and maps with the people, products and services I am most interested. And if the mechanisms were in place, I would also share them with people in my sphere of influence/relevance.
Taken together, that's a context-based advertising model that "clicks" with consumers, advertisers and web site operators alike.
Andrew - I haven't been specifically thinking about online stock trading. Living in Europe I am more used to e-banking in the sense of having access to all core banking functions, which includes bill payments and mortgages. Once I'm through with my transactions my attention can be grabbed by something else and an ad (maybe in the form of a deep link) can be quite effective. You can find an analogy in grocery stores at the check out counter. Goods that are located close to the exit tend to sell well.