Summary: In response to the FTC Staff Report (on which we blogged earlier), online advertising industry associations joined together and came out with their own principles to improve the data collection and use principles of the online experience. In essence, it is a last-ditch effort by the industry to keep its role of self-regulation. You can read the entire report at www.iab.net/behavioral-advertisingprinciples.
The online advertising industry has responded to the February 2009 FTC Staff Report on the topic (which is called “behavioral advertising”). Here is a summary of the main principles in our words (not theirs):
- The Education Principle—an 18-month campaign to educate consumers.
- The Transparency Principle—which mandates clear and easily accessible data collection and use practices and changes to websites.
- The Consumer Control Principle—enhanced options for determining which and how data are collected. It is expected that a “data consent” toolbar will be created for a broad range of online providers (ISPs, browsers, publishers) that will enable consumers to consent to data collection. There will also be steps to “de-identify” the data.
- The Data Security Principle—essentially a data security and retention policy to improve security of the data and limit the period such data are retained.
- The Material Changes Principle—requiring consumer consent to any material changes to privacy and use policies. If successful, this will address one of the principal concerns of the FTC, which is the now-current practice of empowering online providers (e.g., publishers) to change their TOUs, etc., with retroactive application. This is something courts have now found invalid in such online agreements, too.
- The Sensitive Data Principle—addresses concerns for data from sensitive groups—e.g., children—and sensitive data—e.g., health and financial records.
- The Accountability Principle—should be the implementing programs for these principles and disciplinary procedures.
So What?
One thing is clear: the online agreements that govern the use of websites are now in the crosshairs—meaning that those agreements should be revised now. The FTC Staff Report provides pretty good guidance on what to include and what to exclude.
The industry report is something more abstract but it also has some gems that can provide some great innovations for new practices. Among them, the “data consent” toolbar is a good idea.
In addition to that idea, we really like the idea of steps to “de-identify” the data. In fact, we think that this is probably the most important step forward. Those data can be tremendously powerful in that form. That is the fulcrum point—and where fortunes will be made. And some lost, too.
(See some TOUs, etc., we have drafted: www.npbn.com and www.photospin.com for some examples.)
Summary: It is true that we will continue to pound on one of our more popular themes–the ever-increasing value of the ever-increasing piles of data from Internet usage. Now comes research on the value of Google Trends, in this case to track changes in economic activity or consumer preferences as they happen.
File this not under “We told you so” but “We will continue to tell you so.” This is about data–how Internet usage data can be mined, sliced, diced and mashed up for valuable insights. Someday, of course, someone will address the (trickier) question of making money from it. (Oh, details, details, details.)
First, a goldmine of a resource. Go to Google Trends (http://www.google.com/trends) to see what search terms people are using (your own name will probably not register, unless you are, oh, Steve Jobs). You can also use more advanced analytic techniques at Google Insights for Research (http://www.google.com/insights/search/#). This is real-time stuff. And, get ready for this: You can download it as a CSV file.
Berkeley professor Hal Varian (also wearing a second hat of Chief Economist at Google) and Google colleague Hyunyoung Choi wrote a paper on predicting the present (clever twist, that) and on their blog, happily quoting Yoga Berra as to predicting the future. (The article can be found at the Google Research blog at http://googleresearch.blogspot.com/2009/04/predicting-present-with-google-trends.html.)
They correlated search term volume with statistics gathered on economic activity by other sources unrelated to search engines. For example, they matched the search for certain real estate terms against home sale volumes (reported from another source). Another example is the search for travel destinations against reported arrivals. True, their model requires a certain amount of “re-jiggering” but not so much as to make the results suspect. Their “modifications” look to be within the norm of statistical analysis. Nonetheless, the results are stunning.
The point is not so much getting data that are exceptionally precise but rather to give relatively accurate directional results and to improve the reliability and precision of subsequent models. For the corporate user, then, the directional results can be compelling because they are certainly accurate enough for strategic purposes. In other words, the predictive power is sufficient to understand probable upturns or downturns within a few percentage points but not so great that you can get an accuracy of, say, three per cent. Hence the use of the word “strategic” above. In addition, it should be emphasized that it looks like the predictive power is short term.
On May 4th I will be participating on the panel “Valuing Media Assets in a Down Market” as part of the Media, Money & Technology Symposium of the DigitalHollywood 09 conference, May 4th through May 7th in Santa Monica. The conference will be held at the Loews Santa Monica Beach Hotel. http://www.digitalhollywood.com.
I will be discussing some of the due diligence minefields, especially in licensing agreements that serve as the basis for ownership claims on (and therefore value of) media assets.
Ping me for more information. You can also try me on Twitter at globalcapjcr.
Summary: If you provide content to, or receive content from, newspapers or you advise newspapers, new strategies (deployed by European newspapers) raise some questions about controlling the use of the content and your revenue. Re-read the agreements.
We just posted some (good) news about what look like effective strategies for newspapers in Europe to survive (see digitaldumonde.wordpress.com, for example). In a nutshell, they have revived two old concepts–repurposing content and offering premium services. The premium services are not necessarily news or other content but, for example, a weight-loss club.
That got us to thinking about content licenses. Too often, we see that the grant of license is either too narrow for the newspaper to use the content on other platforms or it is too broad to provide adequate protection to the content provider. And, if you are providing the content, it is possible that derivative works are not effectively prohibited.
Narrowness occurs because the agreement was often drafted only with a website in mind. The problem of breadth occurs because the website is not sufficiently defined (narrowed) to mean only one website. Another problem may be that there is no limit on the number of places or times content can be posted.
The problem arises in the revenue section, too. Often it is driven by the definition of revenue–which is limited to revenue generated from the display on a website. Thus, it might be silent as to other platforms.
And The Point?
It is a simple fix by specifying the platforms, the number of places and times content can be used and a better-drafted definition of revenue.
As for the premium services, things are a little more difficult. A content provider could argue for a bit of that revenue if the user leaves from the content providers content (say, a story) and goes to the premium se3rvice (e.g., the weight-loss part of the site on another page). Bit of a stretch but worth discussing.
And, of course, one can–and should–always press for access to the user data, even where they go after the content. No, not the amount of weight lost, but simple use.
Perhaps you can get a discount on the weight-loss club membership.
Quick Summary: Whatever the agreement with a vendor (e.g., content, etc.) for your digital initiatives, own the data. You will figure out later how to use them.
What do data, The New York Times, basketball and your agreements have in common? Well, I am sure there is a joke in here somewhere, but actually it is a lesson about using data. But first you have to own them.
And Therein Lies the Rub.
We have often seen parties not press to own data collected from digital platform deals (say, a license for content). The party’s management would say something like “We have no idea how to use it.” Well, lots of things start off with no one having a clue how to use it (which is itself a topic for a post), but at least get access to those data. Because they will be very valuable.
And that’s where The New York Times comes in.
Well, more precisely, an article in the Sunday, March 14th edition of The New York Times Magazine, entitled The No Stats All-Star, by Michael Lewis, author of Moneyball, a book in which he develops the theme he carries into this article. The article can be found at http://www.nytimes.com/2009/02/15/magazine/15Battier-t.html?scp=1&sq=Shane Battier&st=cse. Lest the New York Times gets angry at me for a “deep link,” you can go to http://www.nytimes.com and search for “Shane Battier” (with the quotes).
The article is about the Houston Rockets guard, Shane Battier, who is an NBA guy with unimpressive stats, IF you look only at the usual stats (points, rebounds, blocked shots). But IF you look at the right statistics, then you realize that he is actually quite formidable.
The first lesson is not so much about basketball as it is about the the volume and types of statistics they collect. And they collect massive amounts.
They then slice and dice these statistics. Then (and this is important), they communicate these data to Shane in a way that he can act upon them. Then, and this, too, is important, they create an on-gong feedback loop.
What they collect can be translated into all kinds of businesses, but that is not the point of this post (though it would be a good exercise for any manager). The point is:
What Does This Have to Do with My Agreements?
Websites, mobile phones and, very soon, mobile device (in some countries they are called MVDs, others MIDs and in the US the “netbook” is emerging) collect (or at least can collect) mountains of user data. And no, you do not have to collect personally identificable information (“PII”) because the other data can be very powerful. In fact, these platforms have not yet begun to figure out just what kind of data they can collect.
It is true that no one knows what to do with a lot of the data. So, we have a situation where platforms do not yet know what data to collect and people do not know what to do with the data, which leads to people negotiating agreements where data ownership is either not addressed or too narrowly tailored to what is known now.
And this is where the article provides an object lesson. The Houston Rockets did not know which data to collect or even what to look for–but they found out. And their game improved.
So What?
So make sure that:
- Own or License the Data. You own the data or, at a minimum that you are granted a license to those data.
- Collect Other Data. You have the right to collect other types of data (or have access to other data that the other party collects)
- Direct Future Data Collection. Specify that other data will be collected and provided to you.
For those of you familiar with entertainment legal agreements, there is a “future technologies” clause. Perhaps this will be the “future data” clause.
And don’t leave rights to the data on the table merely because you do not know what to do with the data. You’ll figure it out once you have them.
Global General Counsel 