Where Online Advertising Fails: The Future of Web Advertising
by
on February 25, 2008,
A huge proportion of tech news has to do with online advertising: studies on its efficacy, technologies for serving ads, ad revenue ups and downs, and acquisitions having to do with any of those. The reality, however, is that we don't know much about how to really leverage online advertising, and we have a lot to learn.
What set me off to begin with was an opinion piece on WSJ.com by Esther Dyson, which hovered near getting it, but ultimately missed by a mile. Ms. Dyson made the point that the “traditional” online advertising is fading, but asserts that advertisers' audience members are best accessed in social communities, and that data will ultimately be compiled and disseminated to advertisers by ISPs. As much as I think any ISP would love to get their hands on that sort of ad revenue (as noted by the newish hijacking of my URL typos by Time Warner Roadrunner), it's impractical. How can you assume that the same person is always using the same IP? Are all ISPs going to move to a static IP addressing scheme? What about VPN clients and WiFi access points and mobile browsing? There are too many holes in this concept for it to work until they can track me by serial number on the plug in my head. Three UK ISPs are willing to give it a try, however.
The formation of quadrantONE completely misses this boat, and somehow thinks that trying to compete with Google in the blanket-your-ad model still makes sense. The main problem with newspapers as they exist now is that they are still in the business of trying to have something for everyone, and aren't able to sort content for readers. I'd be far more willing to subscribe to an online newspaper that customized content specifically according to my needs (and feel free to subsidize part of the cost with targeted ads) than I am for a paper copy that arrives in my box every morning with last night's news and ads I never even see. Of course, quadrantONE took some of their advice from Nielsen, who can't even manage to figure out if advertising actually accomplishes anything, so that may have been their first problem.
Ms. Dyson was right, however, when she said that the current model is losing ground. Google themselves saw a dip in their share of the U.S. ad market last quarter, and while they are still performing well, there is already a trend to head more toward niche marketing. Vertical ad/content networks like BlogHer and Glam.com are already eating into Google's model in small bites. What makes more sense as an advertiser? Buying your AdWords and splashing them over how many thousands of sites where people are already immune to seeing the same ads (yelling smilies, anyone?), or focusing your efforts on a specific area of content that appeals to your target demographic?
Here's what we do know: the people who are clicking through on ads aren't the people advertisers really want to go after. A joint study by comScore, Starcom, and TACODA recently discovered that over 50% of click-thrus are from 6% of the online population, and that 6% skews toward a household income under $40,000. In other words, not the people who are going to be doing a lot of online shopping.
The traditional ad folks like Google and Yahoo are trying something new in going to video, with Google testing video ads and Yahoo snapping up Maven. Other ad companies like Gorilla Nation also think that video ads are the next wave in advertising. And of course, we have the social networking idea, which is being flayed virtually everywhere. People don't go to social networks to shop; they go to socialize.
The reality, however, is that if users aren't clicking through ads as they are now, why on earth would they click to sit through a video ad? This is old-school advertising, designed for a time when television also served all things to all people and you sat through commericial breaks instead of fast-forwarding with your DVR. Niche cable networks have found success providing a smaller, more concentrated market for advertisers, and online companies are following suit. GoFish has launched an ad/content network geared at children and teens, while MTV Networks (which also owns Nickelodeon) acquired a whole parenting content network in Babunga.
But even in a combination ad/content network model, an assumption is made about the audience. Fashion and gossip web sites must be read by women. Tech sites must be read by a predominantly male audience. The ads are still going to be fairly predictable, and ignore those outside the “typical” demographic. So who really has the right idea?
Amazon has the right idea, that's who. With their latest move, testing ads, I'm convinced that Jeff Bezos is a genius. Amazon has all the information at its fingertips that Beacon went so very wrong with, and they have been very upfront about collecting it. They have your purchase history dating back to your very first purchase on the site, and display it for you right in your account. They have a wide variety of products from nearly every possible retail market, from groceries to media to electronics, and with the addition of merchant storefronts through the Amazon system, they increase that data set. Who knows you better than Amazon, and who could target you better? They already show you what other people who've purchased what you purchase. The make product suggestions based on your purchase history. Serving the ads will only take it a step further, providing advertisers with as targeted an audience as they could get, and reducing the amount of ad noise you are subjected to. They may or may not succeed, but at least they are on the right track. Imagine Amazon tied into your content network. It may be a little Big Brother-esque, but it's an advertiser's dream.
If you enjoyed this post, make sure you subscribe to profy RSS feed!
Maryann Jensen
Oliver
Hank
Opensource 
Michael Fidler 
1) “The main problem with newspapers as they exist now is that they are still in the business of trying to have something for everyone.” No, that’s their main advantage, and it’s why they won out over the “vertical slicing” of earlier publications: pamphlets, gazettes and registers. Read enough newspapers and you’ll learn it’s called “serendipity.”
2) Amazon don’t know a thing about me any more. I stopped using it after I it took me over a month to report someone using my email address with a stolen credit card, during which time they allowed the crook to try at least three different stolen cards. I don’t expect I shall ever make another purchase there.
3) Come to think of it, Google don’t know a thing about me either. Like everyone I know, I use ad-blocking software, not just for convenience but for sped and security too.
The gaggle of latter-day pamphleteers who expect to have their lives subsidised by advertising forget that placing ads before prospects is a privilege. Abuse it, and they’ll realise that not only are you irritating them, but you’re also pushing up the prices of the advertised goods in the shops. There is a socially acceptable level of advertising, reached over generations of unconscious negotiation. “New media” advertisers are in danger of forgetting this.
Newspapers won out against the pamphlets how long ago? Their reign is no more, and it’s simply because there is too much information out there and far quicker ways to get it. You don’t have to look at the newspaper layoff numbers to see it; any street in America will demonstrate exactly how few people get newspapers anymore. If the grocery stores in my area would cease their moratorium on using online coupons, the papers would vanish even faster than they are. The reason that newspapers originally took over was because they were a faster way of disseminating information than individual pieces of print. Now the opposite is true; there is a glut of information and people don’t have or want to take the time to sort through that open flood gate to find what they wanted. Major news stories will always hit a vertical content network, but picking and choosing from online content providers is the same as grabbing the world news section and the sports section from the paper and tossing the other sections.
I can’t speak to your specific Amazon situation (although I’m not understanding how someone else could hijack your email for purchases without any sort of confirmation process), but I find it hard to believe that Google doesn’t have data captured for every single person who has been online at some point. Even if you aren’t using their services, any search you create, any mail you’ve sent to someone who IS using Gmail, any web site you post on goes through their gigantic reach. Blocking ads doesn’t do a thing when it comes to data collection, and unless you are using IP masking, a pseudonym, and cash payment for absolutely everything, Google knows at least something about you.
If you read my article, which I’m not entirely sure of based on the last paragraph in your comment, you would have realized I was talking about exactly how the web is going to move away from the constant glut of advertising and move it to a far more focused model that does away with the blinking ads for nothing you’d ever be interested in. Amazon will be able to do that for their customers. Others will be able to do the same, minimizing the barrage that is currently underway as companies try to get their ads to everyone hoping that the audience they really wanted to target will see it as well.
I think you’re spot on about the direction of advertising. The flood of information on the internet can be used to target products more exactly, increasing the usefulness of the ads I see, increasing the likelihood I’ll buy the product. This helps advertisers as they can pay for a smaller ad campaign, and helps me as the noise to signal of ads decreases.
I disagree with some points, though. I think ISPs will play a big role in the collection of this information. Even with dynamic IP addresses, the ISP is assigning me that number and can identify me even as it changes. Yes, the family might all use the internet, but 5 is still a far more granular measure than say, the 10,000 visitors on a website. Plus clustering techniques could help identify the different users of a connection. Wireless ports that serve large quantities of users would be useless, but then you just ignore data from connections with too much throughput to be one user.
The Amazon model is a start. What I buy says a lot about what I’ll buy in the future. It’s not perfect, as lots of purchases are gifts and skew my profile. As this form of advertising evolves and more information gets collected, I think that the targetting will get even more precise. Do I tend to read environmental articles on news sites? Then target me with ads that promote how eco-conscious your brand is. Do I seem to be swayed into buying discounted items a lot? Then include a special offer in the ad. Figuring out what I want is just the first step. The next step is figuring out how to make the sale.
Paul, thanks for your comments. I still think that customer purchase history is more valuable than the IP address. Beyond even dynamic IP addressing and multiple family members, you will have always have open WiFi networks, mobile use, hot spots, libraries, work places, etc. that will go unused when it comes to collecting data. And there is an ingrained distrust of anyone collecting data that you didn’t give permission to do so, which is why I think that the underlying premise of Beacon was correct. We feel that it’s spying when a 3rd party tracks what we do, but we know that our purchase history is saved. Have you ever heard a giant outcry about Amazon’s tracking other than the specific purchase circles? We’ve been seeing “Customers who bought this also bought this” and “Of those who viewed this item, x% bought these other products” and “Here are your recommendations” for ages without balking.
ISP’s can identify you more accurately than your IP address will. Consider when we used to pay for the internet by the hour (do people still do that?) Your IP address might change, but clearly the ISPs were still keeping track of your account as a single entity. And there will be plenty of connections that don’t lend themselves to this data aggregation, but that’s always an issue. The percentage of people connecting to the internet via a public location is much smaller than the percentage of people without an Amazon account.
I agree that the spying aspect is a bigger issue though. In regards to ISPs, the net neutrality furor is a good example of how people react to them stepping out of their role of simply providing a connection to the internet. I really can’t predict whether privacy advocates will grow stronger or if as data portability and use grows people will start accepting more intrusion into what they present of themselves to the internet. I recall in the pre-Broadband days some ISPs would give you a free connection if you let them show you ads. I wonder if a similar strategy would make ISP data collection more palatable.
Anyways, good article. Its provoked some interesting thoughts about where the trend of using customer data to target ads will take us.