Monthly Archives: April 2010

Nielsen fluffs up the TV ratings

The media ratings service Nielsen made two humongo moves recently that acknowledge video viewing is migrating away from live television. On Wednesday, it announced it would no longer track how many TV channels the typical U.S. consumer receives, because as MediaPost reports, “there no longer is a ‘consistent’ meaning for the term ‘channel.’ “

But the real story is this: Nielsen has also decided to change the very nature of television ratings by including “duplicate” viewing — as in, you watch a TV show tonight, and then you watch it again by playing it back on a DVR. This may sound like a nuance, but it is a huge shift in the concept of a media audience delivered. Advertising impressions have always been perceived as mutually exclusive. A newspaper with 100,000 circulation is assumed to reach 100,000 different sets of eyeballs. Broadcast is fuzzier, of course — a 100 GRP schedule could reach 100% of the viewing population once, or 1% of that population 100 times — but at the micro level of a single commercial airing, each audience has been assumed to be unique.

No more. A critic might suspect this move is Nielsen’s way of bolstering the broadcast industry, by boosting ratings numbers as audiences start to slide elsewhere. Magazine and newspaper publishers have tried similar gamesmanship with their BS “readership” malarkey; back in 2007, for instance, Essence magazine claimed its 1.07 million printed copies reached 7.8 million readers thanks to the magic of “passalong readership.” (“Look, honey, Essence magazine came in the mail — let’s hold a party and invite all our friends!”) Is Nielsen gaming the system by adding in numbers beyond the “live audience” to now include duplicates as well? Perhaps. Either way, the real challenge for advertisers is Nielsen provides no way to determine if any of that downstream DVR audience skips over commercials entirely.

Image: Tantek

Your Facebook daily newspaper

Zach Allia gives newspaper editors another reason to cry.

You see, last week Facebook expanded its “Like” button — a little icon you click inside your Facebook news feed to say you dig something, which in turn pushes that item higher in your friends’ Facebook stream — to partner sites around the web. Like wildfire, now The New York Times, CNN, Huffington Post, Vimeo, and thousands of other sites have spread the Facebook “Like” on their articles and videos so you can push their stuff back to friends.

So this week Zach cleverly launched a app to collect everything your friends are liking. You can even sort all the recommendations by categories: News, entertainment, travel, movies, reviews. If you trust the judgment of your friends, you no longer need a publisher portal. Which gives newspaper editors their new reason to cry.

EdgeRank: Why Facebook filters your stream

If you’ve used Facebook for a while, you know that sometime about a year ago the social network updated its “stream” to stop showing everything everyone you know posted, and instead magically began listing only items you might find interesting. The magic was a filter — a process Facebook calls EdgeRank. In essence, EdgeRank creates a score to judge whether something someone else creates is worth pushing to your news stream, multiplying the affinity between you and the creator by the “weight” of the type of change (e.g. a comment is worth more than a “like” click), and also by the recency of when the change was made. If your girlfriend made a long comment a minute ago, bingo, up to the top. That long-lost cousin you rarely chat with, boom, off the page.

When Facebook first started culling its news feed, technologist Robert Scoble was an immediate fan, writing (and we paraphrase) that the filter made his feed instantly warmer — tapping a more relevant, lifelike view of his contacts. But think more deeply about it, and it is startling that Facebook admits its own network was filled with so much chaff that shoddy material must be eliminated. Metcalfe was wrong when he supposed that the value of a network grows exponentially based on its number of nodes, because he failed to see that each point is not created equal, and the data flowing between them also has differing levels of quality. Marketers who hope messages can go viral inside new human networks may hit a wall if the users of such networks find filters are necessary to wall off the content they find irrelevant. Perhaps, just as the automobile amplified daily pedestrian travel, new communication technologies have reset our ability to maintain relationships at a higher level — but like your new commute to work, it’s only a set distance further, not an exponential growth curve to the horizon.

Using product ownership to slash online ad costs

Novices to online marketing are typically shocked that banner ad costs can swing in a 10-to-1 range, but that’s the modern Wild West of Internet media buying. About a year ago we were negotiating banner ad rates with The Wall Street Journal. The quotes were all over the place, north of $60 CPM in the initial round, down below $20 as we found other nooks and crannies in their pricing strata. And then, of course, with behavioral targeting, we reached identical prospects for $6 CPM or less.

The secret is simple: If you can find a way to reach the exact same audience directly, without paying the toll extracted by a major publisher, you can often free up 90% of the costs. (This issue has nuances; see Thomas Miskin’s excellent riff on why publisher context is sometimes more important than media costs.) Advertising networks are typically the best way to remove the toll; their collections of thousands of web sites can add cookie-based consumer tracking to pinpoint audiences that expensive, elite publications once owned.

OwnerIQ targets the man behind the mower

Now OwnerIQ provides a new spin on behavioral targeting. OwnerIQ is a consortium of advertising networks that uses observations of consumers who visit manufacturer pages indicating they own a product. Say, for instance, you boot up a web site with details on the owner manual for a Blu-ray device. Ping. It’s highly likely you own that gadget, so by tagging your computer, OwnerIQ can then serve future ads to you as you travel across the web — perhaps for a Bose sound system. OwnerIQ collates the data from three sources: relationships with about 30 major manufacturers; publisher partners that offer parts for specific products; and its own site, — the type of content someone only reads when they have to work on something they own.

OwnerIQ claims in its case studies that response rates can be 50% higher or more than standard web banner CTRs. The cost structure is a fraction of niche marquee sites reaching the same audience. If you believe that past product purchases are a signal for future consumer behavior, OwnerIQ may be worth checking out.

Image: Idiolector

Facebook’s 10% problem

Facebook and Nielsen have released a new study on how well consumers recalled the messages of 14 ad campaigns. Let’s ignore for a moment that the study was done via an opt-in survey instrument on Facebook, drawing an inherently biased pool of people who are willing to click on an online offer, and see what the data showed:

1. “Organic” impressions, or messaging about a product that your friends actually write to you, led to much higher rates of recall than the paid ads themselves.

2. Yet paid impressions, the actual ads, typically make up 90% or more of all impressions online — leaving only 10% the social good stuff.

Conclusion: People listen when friends talk, but it’s very, very hard to get friends to talk. In our own experience with clients, the vast majority of Facebook ads go unnoticed, much more so than other forms of online display advertising. Click-through rates on banner ads across all U.S. web sites averaged 0.08% in 2009; on Facebook, campaigns are typically in the 0.02% range. CTRs don’t track real impressions, of course, but they are a good proxy for how much of your online audience is actually digesting the offer. Part of the challenge lies in social media sites being refreshed frequently as users click to update the news streams, making “impressions” on each page relatively inflated compared to news sites where readers linger over articles. If true organic impressions are only a fraction of the actual paid ads that make it to real eyeball retinas, Facebook’s 10% problem may be more like 0.002% — a tiny, desirable dynamic in which your consumers tell others how good you are, sweetly powerful and so very hard to control.

You can download the Nielsen report here.

Facebook, the operating system

Facebook pulls you further into its loyalty schemes this month with a few clever advances. gives Facebook users a way to find, create, or share Microsoft Office documents inside the Facebook ecosystem. Facebook’s Platform Showcase touts how major sites such as ABC, CNN, ESPN, NYTimes and Yelp are integrating FB functionality (“Look, honey, we can now share the news!!”). And back at the ranch, Facebook is running ads on its own user home pages promoting the “Like” button you’ll see across such partner sites — hoping you’ll click it to push that content around the stream to make all this document creating, web site partnering, and newsmongering work.

Razorfish’s Shiv Singh suggests Facebook will use data from users clicking “Like” around the Internet to build behavioral-targeted advertising, serving you marketing messages based on the fact you dig polka-dot bikini swimsuits. (Shiv, sorry, we’re paraphrasing.) Or the data could be collated and sold to advertisers and publishers to use in their own networks, with Facebook becoming the Experian-type list company of the entire Internet. We could get really crazy and suggest direct marketers will tap such lists, combining users’ self-admitted desires and the homophilic connections between them to drop junk mail into your house because your online friends love polka-dot bikini swimsuits. Imagine the irony of technology’s hippest online network making mail work better.

If all this sounds confusing, just imagine Facebook as a software platform running inside the Internet, building a level of utility that is hard to leave, adding open architecture that encourages others to plug into it. This strategy worked once for Microsoft. Someday, somewhere, we expect a college kid to launch a new sticky platform … and that one might run inside Facebook.

Image: Bekkchen

The noosphere of margins

If you don’t work in advertising or communications you’ve likely missed the enthusiasm among middle-aged men who wear jeans and black T-shirts about crowdsourcing — the idea that masses of people can be sorted to build better solutions at lower costs. It’s a play off the “wisdom of crowds,” where groups of people, each of whom has partial information and can make an informed best guess, often when averaged together can find a near-perfect answer. The world is filled with examples, from children trying to guess the number of marbles in a jar (the aggregate answer is almost always spot on) to markets of investors trying to gauge the future value of a stock (speculative bubbles aside, the way prices of pieces of companies adjust nearly instantly to new information is almost miraculous). Crowdsourcing expands this concept by making it an economic model: You set a prize, encourage thousands to contribute an answer, and after judging the collective input, opt for the top one. That is, instead of selecting the middling average of the response curve, you cherry-pick the best outlying result on the fringe.

Want a new cell phone design? Instead of paying teams of internal engineers, throw a contest, and some brilliant young design student from the crowd will raise her hand to knock your socks off.

All of this has excited the ad industry because it provides a fundamental restructuring of human creativity (the most positive reason) while also greatly reduces costs (useful if you’re managing agency human resources). Futurists also dig it because it pulls deliverables out of the noosphere — the “global consciousness” concept by Vladimir Vernadsky that we are moving into a third phase of the Earth’s development, after physical formation and biological evolution, to a higher form of collective intelligence. This feels trippy, perhaps, until you look out the window of a plane landing in JFK and see how tiny swarms of humans are acting collectively to terraform our planet, and then, holding that perspective, realize our species may be acting like bees with a hive mind. No less an organization than Princeton has set up 65 devices around the world that randomly generate numbers — and is monitoring them to see if some form of global consciousness is making those numbers less random. Really.

Crowdsourcing employee firings: A thought experiment
The trouble with group consciousness is sometimes it moves in ways that hurt individual participants. Here’s a thought experiment we’ve posed with Edward Boches, the Mullen agency creative chief who is a strong proponent of crowdsourcing: Imagine you run a business with 50 employees, each of whom you pay $52,000 a year or $1,000 per week. One day you decide to fire all your employees and instead hold a weekly global competition for each of their jobs, with a prize for each winner of $100. Thanks to nearly perfect information systems, thousands of people apply for each job post, and you have no trouble filling each slot. (There are lots of smart people starving in the world; now you find them and they find you.) Training would be required of course, but as part of your competition each winner must agree to conduct three weeks of home study, so all the fresh employees hit the ground running. Then, every future week, you fire everyone again, hold another competition, and hire a new crew of winners — ensuring a constant flow of improved ideas and talent, at a 90% payroll savings!

If information systems were perfect enough, this model could become real. But is it fair?

The economic challenge that few raise about crowdsourcing, or Chris Anderson’s broader theory that all services want to be free, is that there is some value in inefficiency — a value extracted from imperfect exchanges that is passed onward to support society. Your business exists because it charges a profit margin, and that profit exists only because you build and defend fiercely some form of inefficiency between supply and demand. If customers could get food teleported to their kitchen table from the farm fields in California, all the inefficiencies of transportation and packaging and storage would be gone — and all the margins of those businesses in the middle also. If your own customers could get what you produce without touching your operations, your source of income disappears. Do we want to live in a world where there is perfectly seamless transfer of goods, services and information? Or the real question is, can we?

Image: Pensiero

Dongle prisons, or how iPods pull society apart

On an American Airlines flight inbound to Chicago, light fading as wings shredded fog outside, we noticed about 80 passengers in various stages of reading, typing on laptops, punching numbers into smart phones (should be off? oh never mind), listening to Bose QuietComfort 15 Acoustic Noise Cancelling headphones, little screens glowing everywhere … and it occurred to us: We have entered a world of dongle prisons.

A dongle, as futurist Jaron Lanier explains on page 109 of “You Are Not a Gadget,” is a proprietary mechanical device that acts as a key to make software or content work. Think of the CueCat in the 1990s (a pen-like scanner that allowed early web users to tap magazine ads for more online information), or more recently Kindles or Google Droid phones (complete with hard key for booting Google Search) — all are gadgets meant to lure you into content prisons. “Prison” may be a harsh word, but the strategy is obvious: The aggregators who profit off the sale of information (Apple with iTunes, Google with search sponsored links, Amazon with print or electronic books) can only thrive if they convince you to become chained to their systems and avoid competitors. To lock you in, they convince you to buy a shiny piece of glass that unlocks their content. Lanier notes that all such material — music, ebooks, search results — really is just bits that could conceivably be accessed from any device with a headphone jack and screen, but “dongles” create the illusion of artificial scarcity. Laugh at CDs and vinyl records all you want, technologists: The iPod in your pocket is just as antiquated a delivery device, cleverly designed to limit access.

Societal costs

Attempts to build “sticky portals” are nothing new, but the proliferation of new dongles tied to content ecosystems is starting to cause rifts in how the Internet, and society, function. Trouble is, by limiting access to content, such devices create incentives for consumers to create self-centered feedback systems that in turn polarize society. Do you read the entire daily newspaper anymore, or just RSS feeds and blogs from people who think like you? Do you watch straight-up news, or a cable channel that leans in your political direction? Tech gadgets are bifurcating Western culture. We are building self-service walls, of political opinions or content tailored only to your specific needs that draw you into extremes of mental behavior at odds with your neighbor next door. No one wants the middle anymore; CNN is plummeting in the ratings as liberal MSNBC and conservative Fox News lap viewers up. The desensitization of consumers by free, limitless content means they require ever stronger stimulants; the fluid access of online tools makes it simple to find the content that bolsters preconceived prejudices; this hunger and access in turn drives polarization of content, as the producers of video, news and writing find that extremism builds audiences that might, with the proper dongle device, be tied down to pay.

More is splintering here than the Internet, as Josh Bernoff suggested: content portals are fragmenting society into isolated islands filled with groupthink opinions. The great irony of our age is that in the rush to build mass audiences, the economic winners are instead creating clans who want only to believe what they want, and ignore other opinions. Perhaps this is the next evolution in humans, a division of the gene pool guided not by speciation to survive the elements but by minds trying to weather the storms of too much content. We see macro shifts with the Taliban fighting Western culture; we see national shifts as the Tea Party and Democrats accuse each other of radicalism; we see relational shifts as moms, dads, and children get pulled into different gadget-fed virtual worlds personalized to their demos, insulating family members from each other.

Until we figure it out, plug in your earphones, and enjoy that dongle in your hand.

Image: Swami Stream