Online Advertising Conference Recap

By Eric Goldman

Last Friday, the High Tech Law Institute and the Berkeley Center for Law & Technology co-sponsored the Law & Business of Online Advertising conference. We had first-rate panelists and an enthusiastic audience of over 100 attendees. Rebecca blogged the event (tutorials, consumer issues, publisher issues, advertiser issues) and Sarah Bird posted her own recap. My understanding is that BCLT will post the audio from the conference to the conference website in the near future.

I’m going to focus my recap on just three of the talks.


Joel Winston from the FTC (speaking for himself, not on behalf of the commission) spoke on the consumer issues panel. He said that consumers have a feeling of lack of control on the Internet. He thinks that consumers are generally aware of online tracking, but the tracking process is opaque, and consumers don’t understand the future implications for use and disclosure of the tracked data. Surveys say that most consumers think tracking shouldn’t be done at all or should be governed by an opt-in or opt-out process. Many people like targeted ads but they are worried about other uses of the data, such as security breaches, government misuse and secondary uses.

So consumers want transparency and control, and trust is the key. Adults are concerned about posting data online but kids will post very intimate details online. People don’t understand the privacy tradeoffs, such as the connection between targeted ads and free content. And transparency isn’t working when consumers don’t read privacy policies. Self-regulation is the right approach, but the FTC will step in to protect consumers.

The FTC’s behavioral principles:

* transparency and control

* reasonable security and limited data retention

* express consent for material changes to a privacy policy

* express consent to use sensitive data.


Mark Cooper spoke about an interesting new paper he’s working on. He starts with the premise that everyone hates “interruption marketing” such as TV ads: consumers hate TV ads (interruptive) and advertisers hate TV ads (can’t measure efficacy). In contrast, he thinks newspaper ads are clearly better because they are easy to skip, easy to store and contain more useful information. [Eric’s note: Mark is making a highly stylized argument. I explored the relative merits of different ad media in this paper.]

He thinks online advertising can improve on interruption marketing because the Internet is a two way conversation, not push marketing. He outlined 4 dimensions to measure the acceptability of advertising:

* influence. Online advertisers don’t create audiences, they chase audiences created for them.

* intrusiveness. Online advertising isn’t in the “middle of content.” [Sounds like Mark has never experienced an unconsented adware install…and I’m not sure how he’d explain spam in one’s in-box…] But he worries that data collection may be more intrusive than other media.

* ubiquity. I think he argued that online advertisers don’t devote as much on-the-page real estate to ads as newspapers do. [Sounds like he’s never been to a domainer’s website…]

* efficiency (delivery of useful information). The cost of online advertising is less than TV, which expands the market for advertisers. This also facilitates the creation of hyper-niche content sites.

Despite some of the benefits of online advertising, he worries about how much information he needs to give up to get these improvements. He thinks behavioral targeting and tracking is inherently deceptive but in-session contextual advertising is OK, and maybe informed consent may be OK.

[Eric’s comments:

* I’m not sure the four dimensions he uses are the right dimensions to measure advertising

* His arguments relied on a number of assumptions that aren’t very robust, which limits the extensibility of his analysis.

* I think the statement that behavioral tracking is inherently deceptive must be overstated for rhetorical emphasis. Otherwise, I don’t think that statement stands up to critical scrutiny.

* I argue (in great/excessive detail) that some types of behavioral targeting are both good and inevitable in this paper.]


Rebecca Tushnet spoke on intermediary liability. She made two main points:

1) Intermediaries aren’t good representatives of the speakers who they facilitate because the intermediaries are adverse to their users. Ex: the 512 notice-and-takedown provisions, Google’s policy on TMs in ad copy.

2) There is pressure to move away from a robust interpretation of 47 USC 230. Ex:, the recent Adult Friendfinder case (which I hope to blog on soon) and the Quiznos case.

She thinks there may be merit to looking at the New York Times v. Sullivan case, which people sometimes forget is an advertising case (i.e., the plaintiff was trying to hold the newspaper liable for ad copy supplied by an advertiser). The newspaper wasn’t liable in that case unless it had actual malice about the definition–a very high scienter bar. Perhaps the actual malice standard could be more widely used in the online context; among other benefits, notice alone wouldn’t create liability.