Skip to content

In search of the feedback loop: Google Audio

The Wall Street Journal offered extensive coverage today of Google’s decision back in February to exit the off-line radio advertising business. This isn’t the first time one of Google’s many growth initiatives has failed to achieve the hoped-for results, nor does this particular failure expose a fatal flaw in Google’s extraordinarily scalable business model. It does, however, offer some important lessons that have meaning for start-ups.


  1. Online media is about measurable interactivity. At the end of the day, this is what sets the Internet apart from other types of media. The Internet can be truly interactive, which means that advertisers and agencies can measure engagement, attention, and even effectiveness. If you want to wow people with a gorgeous large-format photograph of the new Porsche Boxster, the Internet may not be the ideal place. But for advertisers who know how to value, measure, and price engagement, the Internet is unparalleled. It’s a vast over-simplification to say that Google’s secret sauce is its ability to leverage data on user responsiveness to optimize its pricing model -  but it’s also not too far from the truth. On the radio, there is almost no ability to measure responsiveness – and sophisticated algorithms such as Google’s have a lot less value.
  2. The human side of advertising can not be ignored. As the article points out, the radio advertising market was dominated by close interpersonal relationships. Try as it might, Google was unable to displace these bonds of trust with an algorithm. This is true offline – but its also true online.
  3. Auctions are only worthwhile when they create real value for someone. Lots of start-ups build “bidding platforms” or “marketplaces” into their business plans, but this doesn’t always make sense. Bidding platforms and/or marketplaces require an investment of time and energy for customers to learn how to use, and they require a non-trivial leap of faith that the price generated by the system will be one the customer can live with. If lots of different players value a good or a service (or an advertising impression) differently from one another, a marketplace can make sense because it provides a meeting place where a “fair” price can be determined. Similarly, if each player has a really good sense of what a particular thing (or advertising impression) is worth to him, he might be willing to engage in an auction/bidding process. I think its fair to say that neither of these conditions are true for off-line radio. Because there is no feedback loop – the real value of an impression is (a) unknowable on a per-customer basis and (b) effectively a commodity. Pricing mechanisms that work brilliantly online didn’t add enough value offline.
  4. Ad agencies are here to stay. Google may have tried to displace ad agencies in the radio business and is probably still hoping to displace ad agencies in the long run in other businesses as well. This is unlikely to happen. Advertisers will always require sophisticated agents to help them navigate the complex world of advertising – both online and offline.
  5. Pricing and measurement are inherently connected. How well you can measure something is tightly related to how much you can charge for it. Because online advertising is inherently measurable, publishers have little choice but to submit to the brutal rationality of the online market places. This is part of the reason that ad exchanges are growing to rapidly and ad networks are struggling. Offline, where measurement is harder, publishers (or in this case, radio stations) still have reason to cling to their ability to sell inventory at whatever price their salesforce can achieve and have little incentive to submit to a pricing mechanism which may dramatically lower the price of their real estate.

So how do the online and offline worlds converge in a way that makes sense for advertisers? One example that has been quietly doing this with significant success has been Yahoo’s partnership with AC Nielsen, which was launched back in 2003. You can view the official Nielsen Homescan Online PDF here. This platform combines data from Yahoo’s online properties with offline consumer panel data with AC Nielsen’s famous Homescan panel. The combination lets Yahoo measure the impact of online campaigns on offline purchases in a statistically significant way.  Below is an AC Nielsen presentation from March 2009 that illustrates how the company is working to link offline and online data to drive real-world strategy for retailers.  Another example is recent research prepared by MySpace, Comscore, and Dunnhumby that found a 28% ROI in offline sales on a $1M online campaign on the social network. What’s important here is that the online results (traffic & impressions) were not impressive, but the resulting offline sales were. This experiment was covered here and here, but the research itself isn’t available online yet. Privacy concerns aside, there is no doubt that offline data and online data will increasing merge to generate actionable intelligence for advertisers – and that, in the long-run, is good news for online publishers.

Nielsen Homescan Online