数字广告业缘何恋上欺诈

    询问任何一位非广告从业者“留下印象”是何意，你很可能会得到诸如“做某件让人记住的事情”这类合理的答复。然而，在广告业内，“印象”的定义几乎没有这么简单。为什么呢？原因在于，如果你在电视、电台或印刷媒体上打广告，根本没办法确切掌握这些广告究竟给多少人留下了“印象”。多少人抬头看过某个特定的广告牌？多少人在商业广告播出期间关注电视？所以，媒体公司历来测量、销售的其实是留下印象的潜力。这套体系产生了预期效果，因为营销商可以估算多少人或许真正看到了他们的广告，几十年下来也计算出了这些“潜在印象”的真正价值。广告圈所称的CPM，即获得每千个（潜在）印象的成本，就源自这段历史。CPM市场不仅效率惊人，而且非常庞大，市值达数万亿美元之巨。它推动塑造了无数大品牌，并为人们喜闻乐见的新闻和娱乐内容的创造提供了资金。

    接着，数字时代来了。

    借助于数字媒体，媒体公司和广告客户终于可以知道到底有多少人看到了它们的广告。科技让市场来衡量实际的印象，而不是潜在的印象。但有一个问题：尽管比起传统的“潜在印象”，“实际印象”要更加罕见，更有价值，但数量更庞大的“潜在印象”业已成了数十亿美元的市场形成的基础。

    为了满足需求，同时让这项定义与电视和其他不易衡量的版式保持一致，广告客户和媒体出版商都同意有意维持“印象”的松散定义。这个故意留下的漏洞旨在维系“印象”市场的流动性。问题是，在数字时代，媒体公司可以简单地创造印象。起初，它们具备低价值，但就如同任何一件拥有无限供给的事物一样，这个行业最终创造了零价值“印象”。为了方便理解这两个例子，设想你拥有一个网页。今天你在这个网页上发布了两则广告，然后你重新设计，到了明天，你就发布了10则广告。你刚刚“创造了”更多的印象，但并没有赢得更多的关注（而这恰恰是广告客户寻求的东西），自然也就没有带来更多的消费支出（这是广告客户的终极目的）。这只不过是价值更低的印象。更大的问题是，鉴于印象的宽泛定义，那些真正精通互联网的人士完全可以创造没有人能看到的“印象”，比如在网站背景上播放视频或者制造虚假流量【在《广告周刊》(AdWeek)的一篇文章中，迈克•希尔兹列举了一个采用这种手法的绝佳例证】。

    所有这些低质量和欺骗性的印象持续压低高质量内容生产商因“真正的”印象而获得的价格。好吧，你或许会说，那又怎样？如果存在这么多欺诈行为，价值就应该下降。说的没错，但随着印象价格的下降，高品质内容公司面临三种选择：

    1. 把越来越多的广告加载到内容之中（这种做法对消费者不利，他们反过来会寻找一种避免观看这些广告的方式，从而持续降低印象的价值）。

    2. 找到一种让消费者为无广告内容付费的方式【家庭影院频道(HBO)，PandoraOne广播频道、奈飞公司(Netflix)和资讯网(The Information)就是这样做的】。

    3. 关门大吉。

    Ask someone who doesn't work in advertising what it means to "make an impression," you will likely get a reasonable response like "to do something that someone remembers." However, inside the advertising industry, the definition of "impression" is not nearly as simple. Why? Because if you run an advertisement on TV, radio, or print, you have never been able to tell exactly how many people those advertisements really made an "impression" on. How many people looked up at a particular billboard? How many people paid attention to the TV during the commercial? So what media companies have historically measured -- and sold -- is really the potential to make an impression. This system worked because marketers could make estimations on how many people might actually see their advertisements and over decades figured out the real value of these "potential impressions."From this history came the cost per thousand (potential) impressions, better known as CPM in media industry. The CPM market was astonishingly effective, and large, to the tune of trillions of dollars. It drove the creation of major brands and subsidized the creation of the news and entertainment content people love.

    And then digital came along.

    With digital media, media companies and advertisers could finally know exactly how many people saw their advertisements. Technology allowed the market to measure actual impressions, rather than potential impressions. But there was a problem: While "actual impressions" were much more rare, and much more valuable than the traditional "potential impressions," the multibillion dollar market had already been created on the more numerous "potential impressions."

    In order to meet the demand and to keep the definition in line with TV and other less measurable formats, advertisers and media publishers alike agreed to keep the definition of an impression intentionally loose. This was a loophole created to help keep the market for "impressions" liquid. The problem was that in digital, media companies could simply create impressions. At first they were low value, but as with anything that has an infinite supply, the industry was eventually creating "impressions" that had zero value. For examples of both, imagine that you have a web page, today you have two ads on that web page, and then you redesign so that tomorrow you have 10. You just "created" more impressions. But there isn't any more human attention (what advertisers are seeking) and there is no more human spending (what advertisers are really seeking). And that is just lower value impressions. The even bigger problem is that given the loose definition of impression, really Internet savvy people could create "impressions" that no human could ever see, like videos playing in the background or fake traffic to sites. (Mike Shields over at AdWeek highlights a perfect example of this practice.)

    All of this low quality and fraudulent impression continues to depress the price quality content producers can get for "real" impressions. Okay, you might say. So what? If there is that much fraud, the value should drop. And that is true, but as the price of the impression drops, quality content companies have three choices:

    1. Load up more and more advertisement impressions into their content (which is bad for consumers, who in turn find a way to avoid, and continues to lower the value of impressions)

    2. Find a way to have consumers pay for content without advertising (HBO, PandoraOne, Netflix, The Information).

    3. Go out of business.