Today in Social

Facebook recently started offering Like topic-based targeting in its self-service ad platform and via its advertising API. While its implementation seems simple, Facebook appears to be doing some data analysis to suggest related topics in addition to those an advertiser specifies, somewhat akin to paid search broad-match techniques. Such targeting and analysis could be applied to a Facebook Connect-based ad network that serves up ads outside of the Facebook site, should Facebook or anyone ever build such a thing. (Are you interested, Microsoft?) At the same time, comScore shows Facebook becoming an increasingly important site for video consumption. High-quality online video advertising inventory is valuable, though Facebook isn’t exploiting that opportunity yet, and its quality content is often fee-based.

Today in Social

The Wall Street Journal hears that Yahoo is preparing a hybrid content syndication/ad network product for third-party sites. AdExchanger thinks it sounds like a combination of Facebook Likes plus an ad network, but that doesn’t capture it, nor would such a thing exploit what Yahoo is good at. The Journal says sites would get a widget with Yahoo content personalized for the reader on the third-party site, based on contextual targeting of other content on the site and/or on behavioral targeting drawn from Yahoo’s data.  Accompanying the content on the widget would be display ads whose revenue Yahoo would share with the other site. There’s no info on how advertisers would purchase these ads. In theory, this sounds like a good offering, particularly for second-tier sites. Yahoo has plenty of quality content, is pretty good at lightly personalizing its own sites, and could offer a brand-friendly environment with the scale of a network – if enough sites sign on.

Today in Social

I’m a little late to this – tip o’ the hat to Business Insider for the link – but it’s a great conversation. I’m a fan of ad targeting, using social media Big Data for that purpose, and applying online advertising and direct marketing techniques to traditional media. To a point. This post by online ad expert Doug Weaver and this interview with Simulmedia’s Dave Morgan point out that you can get carried away. Too much targeting delivers unscaleable audiences and unnecessary complexity. One to one marketing isn’t really cost-effective for most products, not to mention its privacy baggage. And building something that media buyers can’t understand, or that demands they piece it together from ten different suppliers, is self-defeating.

Privacy Legislation’s Potential Impact on Online Media

Because the Kerry-McCain online privacy bill is watered down relative to prior proposals, it will face less industry resistance, and is more likely to be passed this year. That has far-reaching implications for online advertising and targeting.

Today in Social

It’s online advertising forecast season. EMarketer projects nearly 11% growth in US spending next year, to $28.5 billion, and double-digit growth out to 2014. Agency Magna Global is slightly more bullish, and it and big media buyer ZenithOptimedia compare online to other media growth very favorably. NewNet technologies are a new catalyst for building off of proven Internet media models, so we’ve got our own Modern Media Manifesto on how to capture this spending. Meanwhile, could do-not-track or anti-cookie legislation derail this potential gravy train? Or would some publishers be happy with contextual, rather than behavioral, ad targeting?

Today in Social

This Guardian interview with the head of Facebook engineering is interesting, and a little scary. If this is an accurate portrayal, it’s definitely a hacker culture at Facebook. Very small teams work on “several dozen to 100” mostly short-duration projects, with little in the way of formal oversight processes. Their project mix is 80/20 focused on features that will increase usage or grow the number of users, versus generate revenue or cut costs. Top priorities include mobile, infrastructure (server, database, development tools), and machine learning. Machine learning is critical to feed the algorithm that powers users’ news feeds, as well as for ad targeting and security – spotting spammers, etc. And Facebook is mostly a Mac shop.