互联网终极谜题

    广告的好处显而易见：广而告之。但没人真正了解广告的效果有多好，也没人知道如何衡量广告费用的回报率。人们以为互联网能够改变这一切，因为可以直接统计（点击量），但实情并非如此。技术在解决旧问题的同时往往带来更多的新问题。

    准确衡量在线广告面临很多困难。根据网络流量分析公司康姆斯科公司（comScore）的说法，“可见性”是其中之一。该公司上周发布的白皮书宣称，要达到广告效果，网络浏览者就必须看到展示广告一半以上的内容，而且时间要持续1秒以上。听起来不算很高的门槛，确实也不高。好歹有了个标准。但康姆斯科同时指出，即使是这样微不足道的指标，也有将近1/3的在线广告无法达到。

    比如，广告可能出现在页面的下半部，用户不见得会下拉到那里。此外，在康姆斯科采访的主要品牌中，他们的广告推广活动有将近3/4碰到过这样的尴尬情形：公司的广告被放在“不雅内容”旁边，比如色情或者仇恨言论【又或者盗版：据报道，反盗版鼓吹者猛烈抨击宝马公司（BMW），抗议它在提供盗版音乐下载的网站推出广告，宝马不得不宣布将审查公司的在线广告政策。】品牌显然不希望和这些东西扯到一起。

    根据调查，4% 到11%的广告浏览量并不是来自真人，只是自动点击程序的战果。而4% 到15%的广告在目标地区之外被浏览。

    康姆斯科认为，所有这些发现都应该在广告定价时加以考虑。虽然互联网的广告空间无穷无尽，必然打压广告价格（在酒吧里网络媒体高管都会这么跟你说，同时又要了杯昂贵的酒，眼睛都不眨一下），但是某些在线广告位就是比别的抢手得多。“确实有取之不尽的广告空间，但也不是真正的用之不竭，”康姆斯科执行副总裁琳达•亚伯拉罕告诉营销新闻网站ClickZ。“真正好的新闻网站只有那么多，一天也只有24小时。通过测量‘可见性’和广告位置，买家和卖家就能挑选出那些真正能够产生广告效果的地方。”

    康姆斯科根据对12家主要品牌的广告推广活动进行的调查撰写了这本白皮书，接受调查的品牌包括福特汽车（Ford）、食品巨头家乐氏（Kellogg）和电信运行商Sprint-Nextel。

    The conceit of advertising has always been simple: it works. Nobody has ever really known how well it works, or how to measure returns on ad spending. The Internet was supposed to change all that by enabling close measurement, but it hasn't, not really. Technology has a way of creating as many problems at it solves.

    There are many problems with accurately measuring online advertising. One of them, according to the traffic-analytics firm comScore, is "viewability." To be valid, according to a white paper published by the company last week, at least half of a display ad must be visible to an Internet user for at least one second. That doesn't sound like much of a threshold, and it isn't. But at least it's something. And even by that paltry measure, nearly a third of ad impression don't measure up, according to comScore.

    For example, an ad might appear on a page, but on the bottom half, to which a given user does not scroll. Also, nearly three-quarters of the campaigns of the major brands surveyed by comScore had at least some ads placed adjacent to "inappropriate content." That means stuff that brand advertisers don't what the be associated with, like porn or hate speech. (Or piracy: BMW reportedly has announced it will review its online-ad policies after anti-piracy advocates slammed it for running ads on sites offering unauthorized music downloads.)

    According to the survey, between 4% and 11% of ads were served to "non-human" traffic -- meaning bots. And at least 4% and up to 15% of ads were served outside their intended geographical areas.

    All of these considerations should go into ad pricing, comScore says. While the unlimited inventory offered by the Internet certainly puts downward pressure on prices (as any media executive will tell you while ordering another stiff drink), some online ad placements are much better than others. "You have the economics of infinite supply at play, but it's not really infinite," comScore Executive Vice President Linda Abraham told ClickZ, a marketing news site. "There's only so many great news sites out there and only so many hours in the day. Measuring "viewability" and placement "allows buyers and sellers to differentiate that inventory that matters."

    The survey on which the white paper was based examined the campaigns of 12 major brands, including Ford (F), Kellogg (K), and Sprint-Nextel (S).