Hunch's Chris Dixon wrote a great blog post over the weekend about Facebook, Google and the difference between ads that create intent and ads that harvest intent. Hunch, being a recommendation engine, is, of course, all about harvesting intent. (I failed pretty miserably at their Twitter Predictor game with about 30% correct answers. Or maybe it was the game that failed, I don't know).
His post reminded me of two things. One is the urban legend about grocery stores putting beer next to diapers to boost sales of both.
The other one is an old Internet joke that not only has a lot to do with intent and contextual advertising but also points at the ideal state of things, at least from the advertiser perspective. It goes like this.
A young guy from a village moves to a big city and goes to a huge department store looking for a job. The manager asks him whether he has any sales experience, and the guy says that yeah, he was a salesman back in his village.
The guy gets the gig, the first day passes, and the manager stops by to check in on things.
“How many customers bought something from you today?”
The guy says, “one”.
“Just one? Our sales people average 20 to 30 customers a day. How much was the sale for?”
The guy says, “$101,237.65″.
The boss says, “$101,237.65? What the heck did you sell?”
The guy says, “First, I sold him a small fish hook. Then I sold him a medium fishhook. Then I sold him a larger fishhook.
Then I sold him a new fishing rod. Then I asked him where he was going fishing and he said down the coast, so I told him he was going to need a boat, so we went down to the boat department and I sold him a twin engine Chris Craft. Then he said he didn’t think his Honda Civic would pull it, so I took him down to the automotive department and sold him that 4×4 Expedition.”
The boss said, “A guy came in here to buy a fish hook and you sold him a boat and truck?”
The guy said, “No, the dude came in here to buy tampons for his wife, and I said, dude, your weekend’s shot. You might as well go fishing.”
Showing posts with label sticky eyeball. Show all posts
Showing posts with label sticky eyeball. Show all posts
Integrating Display Ads into Content

This site found an interesting way to integrate display ads into its content. The Google/Doubleclick ad is the second one (flat belly) in the top row; it also appears on the site's sidebar in the same fashion.
Banners, Search Ads on Retailers' Sites

Interesting. So Walmart must be making more money with the display ads it serves on its home page through the last year's deal with Yahoo than it would've made by promoting its own merchandise. The entire site served over 900M pageviews in August (Compete Pro data).

I looked at the list of the top 10 retailers, and it seems Walmart's is the only site on that list to serve third-party ads on its home page. It was a pretty big deal when Home Depot was the first large retailer to announce it would start selling ads on its site three years ago. Walmart also serves text ads on its search results pages, via Google.

And so does Target:

Also,
AdSense Ads on a Brand Site
Behind The Malware Ad on NYTimes.com

The malware ad served on the pages of NYTimes.com wasn't the first such case, and not even the most prominent -- a couple of years ago, malware ads showed up on CNN and The Economist, among many other sites. Neither will it be the last, so see how it's done in this code dissection by Troy Davis. (Update on Sep 14'09: NYTimes.com has more details about the attack.)
And here's some historical perspective:
Disposable Advertising, or Don't Kill the Microsite

The Subservient Chicken site attracts 10-20K unique visits each month.
The debate about their increasing irrelevance aside, if we are going through all that trouble to build campaign microsites, why are we always in such a rush to pull the plug on them when the campaign is over?
I've heard two kinds of arguments: (ir)relevance and cost. I can rarely agree with either.
Unless the information on a microsite is so super time-sensitive that it becomes misleading as soon as the clock strikes twelve, I don't see how an old microsite -- retired and cut off from ad budgets, perhaps, but still breathing -- can hurt anyone. Coke's brand equity hardly suffers when I buy a poster on eBay that sports "A Pause That Refreshes" tagline instead of the current "Coke Side of Life". Have Burger King's Whopperetes become so off-brand that the microsite, launched for the 2006 Superbowl, is no longer live?
And I'm not picking on Burger King; the company is better than many others in preserving its own digital ad history. Many of BK's campaign sites are no longer around -- Whopperettes, Power Sitting (blog post) that made fun of Atkins in 2005, the fictional rock band CoqRoq (blogpost, also 2005) -- but the ones still alive show that they can pull in traffic long after their best-by date. Subservient Chicken, launched in 2004, still gets a healthy 10-20K monthly uniques (the graph above) who can't be all advertising types. SimpsonizeMe, developed two years ago, attracted 30K visits last month.
Which brings us to the argument about cost. For the sites that don't need daily babysitting and are not based on a third-party technology that requires ongoing license fees, the recurring costs are domain registration, hosting and bandwidth. Here, the reasoning should be fairly straightforward: does the site generate enough traffic at a certain benchmark CPM to pay for itself in impressions? In other words, if your benchmark is $10CPM and your hosting/traffic tier costs $100 a month, the site needs 10,000 pageviews to break even. (The actual formula I use also includes time spent on site). This kind of planning for the microsite's afterlife could be done before the project even launches; it can then guide designers to include features intented specifically to generate postmortem impressions.
For decades, we've been working with media that because of their nature made advertising transient and, like paper plates, sometimes useful but ultimately disposable. With the web, we can create ads that accumulate viewership over time in a way that pre-YouTube TV spots never could, and yet we are squandering the opportunity. Thinking of microsites as an investment that pays off over time instead of an expense line in a three-months campaign budget could be the first step.
P.S.:
- GSI is the hosting service for Subservient Chicken and Simpsonize Me.
- When visiting SubservientChicken.com, enter "Crispin"
Quote of the Day
Grumpy Brit on how web start-ups misunderstand the ad biz:
"Advertising hasn’t proven the monetizing panacea it was assumed to be because the people who fund, develop and manage web-based businesses know the square root of buggery bollocks about advertising. Which is a bit like the people responsible for laying track being ignorant of anything to do with trains. Despite this modest shortcoming the geeks act like they invented it."
-- via Ad Contrarian (Related thoughts from Seth Godin and Brian Morrissey.)
And wtf is buggery bollocks?
"Advertising hasn’t proven the monetizing panacea it was assumed to be because the people who fund, develop and manage web-based businesses know the square root of buggery bollocks about advertising. Which is a bit like the people responsible for laying track being ignorant of anything to do with trains. Despite this modest shortcoming the geeks act like they invented it."
-- via Ad Contrarian (Related thoughts from Seth Godin and Brian Morrissey.)
And wtf is buggery bollocks?
Online Publishers: Sell What Others Are Buying
Yet another speculation on business model for Twitter on Techcrunch -- AdAge, too, recently joined this new parlor game -- and an insightful comment by a user known only as SAG that boils down to "sell what others are willing to buy":
ROI Math for Brands on Twitter
Tweetwasters, a new Twitter-based toy, calculates the total amount time you spent twittering by multiplying all your tweets by 30 seconds (a conservative estimate, in my opinion, since you probably spend more time reading than typing).
I thought it would be interesting to check a few of the brands I follow (the links are to the corresponding tweetwaster pages):
ComcastCares: 20,719 tweets, ~173 hours
Starbucks: 591 tweet, ~5 hours
HRBlock: 303 tweets, ~3 hours
DellOutlet: 143 tweets, 1.19hours
SouthwestAir: 1,112 tweets, 9.27 hours
Zappos: 1,226 tweets, 10.22 hours
Who is doing better, Comcast or Zappos? Let's assume that our benchmark is cost per follower and crunch some numbers.
ComcastCares is run by the company's director of digital care who makes, let's say, $50/hr and has written 20.719 tweets in ~173 hours. Total time investment into Comcast's Twitter account is $8,650. The account has 6,001 followers, which results in $1.44 per follower.
Zappos is being represented on Twitter by its CEO, who is paid, say, $250/hr and has produced 1,226 tweets in ~10 hours, which makes the total time investment about $2500 and the cost of each of Zappo's 24,049 followers about 10 cents apiece -- 14 times cheaper.
Here's a pretty inclusive list of brands on Twitter. Maybe someone will write a script to run them through Tweetwasters to find the most sociable and efficient brand account?
Earlier:
From Hamnet to Mad Men: Fan Fiction in Real Time
Facebook, Twitter Buzz Visualized
Block Twitter Madness Out of Your Life
I thought it would be interesting to check a few of the brands I follow (the links are to the corresponding tweetwaster pages):
ComcastCares: 20,719 tweets, ~173 hours
Starbucks: 591 tweet, ~5 hours
HRBlock: 303 tweets, ~3 hours
DellOutlet: 143 tweets, 1.19hours
SouthwestAir: 1,112 tweets, 9.27 hours
Zappos: 1,226 tweets, 10.22 hours
Who is doing better, Comcast or Zappos? Let's assume that our benchmark is cost per follower and crunch some numbers.
ComcastCares is run by the company's director of digital care who makes, let's say, $50/hr and has written 20.719 tweets in ~173 hours. Total time investment into Comcast's Twitter account is $8,650. The account has 6,001 followers, which results in $1.44 per follower.
Zappos is being represented on Twitter by its CEO, who is paid, say, $250/hr and has produced 1,226 tweets in ~10 hours, which makes the total time investment about $2500 and the cost of each of Zappo's 24,049 followers about 10 cents apiece -- 14 times cheaper.
Here's a pretty inclusive list of brands on Twitter. Maybe someone will write a script to run them through Tweetwasters to find the most sociable and efficient brand account?
Earlier:
From Hamnet to Mad Men: Fan Fiction in Real Time
Facebook, Twitter Buzz Visualized
Block Twitter Madness Out of Your Life
Pay-per-View as Advertising Model for YouTube
Mark Cuban: "From all appearances, Youtube is trying to squeeze every last nickel they can out of Youtube. They are doing everything they can think of to create advertising inventory. Pre rolls, overlays, display ads, you name it."
That's what many online publishers think advertisers want to buy: pre-rolls, overlays, display ads. But advertisers rarely think like this: "If we buy $100K worth of overlays, we'll hit our revenue targets for the quarter." Advertisers pay for the traffic of certain quality they get sent their way, and, ideally, how this traffic is generated shouldn't be of their concern. This is the principle behind lead-generation sites, and Google's search ads are similar, too.
There's one thing that advertisers who look at YouTube really want to buy. Google, despite all the overlays and pre-rolls, hasn't really been selling it to them, at least not until the recent launch of the sponsored videos program.
I'm talking about video views. I don't think there's an agency out there that hasn't uploaded a creative to YouTube. It's safe to assume that advertisers hope to have as many people view these videos as possible. Why isn't there a straightforward way for advertisers to pay for the views?
Yes, you can buy display ads and featured video units (a minimum order is in, what, six figures?) and now you can also put your videos in video search results, AdWords-style. But shouldn't the burden of driving traffic to your video be shifted from you to Google, who has deep insights into site users' behavior and more computing power than an ad agency ever will?
When a company is being paid for ad impressions, it naturally looks for ways to increase the number of available impressions -- the inventory -- and hence all those crazy ad formats Mark Cuban laments.
But when it is paid for driving views to specific videos instead of ramping up ad impressions, its focus and priorities shift and suddenly an entire universe of possibilities appears. Google knows what tags work best for what videos -- then how about promoting videos by automatically choosing the most optimal tags? Or using the "social graph" of users to identify people with similar tastes and propagate the videos the graph's branches? Or adding a sponsored thumbnail to the "related videos" end frame?
So yes, maybe Google could experiment with selling what advertisers are willing to buy instead of ad formats.
That's what many online publishers think advertisers want to buy: pre-rolls, overlays, display ads. But advertisers rarely think like this: "If we buy $100K worth of overlays, we'll hit our revenue targets for the quarter." Advertisers pay for the traffic of certain quality they get sent their way, and, ideally, how this traffic is generated shouldn't be of their concern. This is the principle behind lead-generation sites, and Google's search ads are similar, too.
There's one thing that advertisers who look at YouTube really want to buy. Google, despite all the overlays and pre-rolls, hasn't really been selling it to them, at least not until the recent launch of the sponsored videos program.
I'm talking about video views. I don't think there's an agency out there that hasn't uploaded a creative to YouTube. It's safe to assume that advertisers hope to have as many people view these videos as possible. Why isn't there a straightforward way for advertisers to pay for the views?
Yes, you can buy display ads and featured video units (a minimum order is in, what, six figures?) and now you can also put your videos in video search results, AdWords-style. But shouldn't the burden of driving traffic to your video be shifted from you to Google, who has deep insights into site users' behavior and more computing power than an ad agency ever will?
When a company is being paid for ad impressions, it naturally looks for ways to increase the number of available impressions -- the inventory -- and hence all those crazy ad formats Mark Cuban laments.
But when it is paid for driving views to specific videos instead of ramping up ad impressions, its focus and priorities shift and suddenly an entire universe of possibilities appears. Google knows what tags work best for what videos -- then how about promoting videos by automatically choosing the most optimal tags? Or using the "social graph" of users to identify people with similar tastes and propagate the videos the graph's branches? Or adding a sponsored thumbnail to the "related videos" end frame?
So yes, maybe Google could experiment with selling what advertisers are willing to buy instead of ad formats.
Targeting Messages by Referring Site
MySpace to Hit $1B in Ad Revenue
So, on one hand, MySpace is expected to hit $1B in ad revenue for the year. On the other hand, the click-through rates for the site are 3-5 clicks for 10,000 impressions (0.03% CTR; from this source and my own experience). At $3 per 1000 impressions, that's $30 for 5 clicks. Not bad.
Crisis in Advertising: Privacy vs. Relevancy
This is the first part of the guest post by David Rostan, the co-founder and president of ListensToYou. With ListensToYou, David hopes to improve the online advertising market by giving users control over the ads they want to see. Prior to founding his company, David worked in marketing and strategy for technology and ecommerce companies. He has earned his MBA from the Kellogg School of Management.
If you want to take a closer look at ListensToYou to understand David's approach, you can use the invitation code "adverlab" (limited to the first 100 sign-ups).
Most consumers realize that there is a tradeoff they make between getting things they want online (anything from news to product offers) and the protection of private information. However, it is currently only framed as a tradeoff between letting companies know things about you (in a pejorative way) and getting things you want; we’re missing the opportunity to maximize everyone’s benefit through a fair, consumer-controlled transaction.
To stop the flow of personal information, users employ anonymizers, delete cookies and participate in the outcry over online privacy. This really begs the question “what are companies doing that makes users want to be invisible to them?” The answer is that companies are gathering more data and creating better predictive algorithms to reach (or trick) customers instead of listening to (or even asking) what they want.
Internet users enjoy control over content in a way not previously possible in other media. What they read and consume, the time of day or night they consume it and the format in which they view the information are highly customizable.
Yet, in the most consumer-driven medium ever, internet users are still completely left out of the conversation in one main area of internet content – advertising. While I may be able to choose a website, an author or an article, video or post, I cannot exert any control over the ads I will see surrounding my chosen content.
Publishers need to monetize their sites, but advertisers, brand marketers and publishers are all cast in a bad light when customers feel mistargeted or stereotyped or feel that their private information has been abused. Exploiting user information is not the way to build good relationships with customers: advertisers lose because people won’t click ads; brand marketers lose because users won’t form positive associations with the advertising companies; and publishers lose because mistrusted ads are a constant drag on the website’s brand and revenue source.
This “everybody loses” situation is an over-arching problem with online advertising: web users cannot control what personal information (their habits, interests, behaviors, etc.) they trade for the online activities they value (product and service offers, socializing, local event notification, etc.) and, therefore, do not get things they value or trust the ads they do get.
An Example: A user may be required to give her age (34) to register to use a website. She agrees because the website claims that this is to insure that children do not use websites without supervision – i.e. as a safety precaution – but then she sees the information used to target her with ads for children’s clothing (she’s happily single with no kids). So, the ad is not only irrelevant, but it is also a surprise to her that her age is being used in a way she did not intend. She stops looking at ads, has a mistrust of the website brand, has a mistrust of the advertiser and, even when served appropriate ads, still carries a perception that the ad is not trustworthy or not for her.
The flaws in online advertising can be described as problems of relevance and trust. Relevancy has been greatly improved by behavioral targeting (even if privacy has not), but behavioral targeting is backward-looking. If you used the internet for a recent house purchase, for example, you really know what I mean. How long will you be receiving and ignoring those mortgage ads?
Contextual ads and demographic data are also used to produce relevancy in advertising, but site context is only a stereotype and demographic information is a best-guess. Just because a 54-year old reads blogs about classical music does not mean that he wants to buy retirement property more than he wants to buy the newest Nike sneakers.
To a consumer, relevancy is not only about getting what he wants, it is also about seeing what he expects. Relevancy is about expectation and emotion, but the content and subject matter of advertisements cannot always be consistent with user’s interpretation of the content and subject matter of the website, making the user feel that the website publisher does not understand why she visits the site.
Trust is even more emotional. If the user does not feel that he has consented to the ad or the information used to serve it, he consciously or subconsciously mistrusts the brands that brought it to him.
The hot topic in trust and advertising is behavioral tracking, of course. Tracking scares consumers and may invade their privacy. A relevant ad often signals to us that THIS is the brand that is tracking me the most, which is bad for any brand. Further, consumers have no control over or protection from unwanted or inappropriate advertising content: they have been removed from the discussion and forbidden from making those decisions.
This lack of empowerment, itself, destroys trust whether the online ads are “appropriate” or not. Even widely respected online publishers suffer the consequences because they have no way to lend their own brand credibility to the ads served on their sites. Publishers are often unable or unwilling to vouch for all ads that are served and, even if that were not the case, there is no understood signal to do so.
If you want to take a closer look at ListensToYou to understand David's approach, you can use the invitation code "adverlab" (limited to the first 100 sign-ups).
Most consumers realize that there is a tradeoff they make between getting things they want online (anything from news to product offers) and the protection of private information. However, it is currently only framed as a tradeoff between letting companies know things about you (in a pejorative way) and getting things you want; we’re missing the opportunity to maximize everyone’s benefit through a fair, consumer-controlled transaction.
To stop the flow of personal information, users employ anonymizers, delete cookies and participate in the outcry over online privacy. This really begs the question “what are companies doing that makes users want to be invisible to them?” The answer is that companies are gathering more data and creating better predictive algorithms to reach (or trick) customers instead of listening to (or even asking) what they want.
Internet users enjoy control over content in a way not previously possible in other media. What they read and consume, the time of day or night they consume it and the format in which they view the information are highly customizable.
Yet, in the most consumer-driven medium ever, internet users are still completely left out of the conversation in one main area of internet content – advertising. While I may be able to choose a website, an author or an article, video or post, I cannot exert any control over the ads I will see surrounding my chosen content.
Publishers need to monetize their sites, but advertisers, brand marketers and publishers are all cast in a bad light when customers feel mistargeted or stereotyped or feel that their private information has been abused. Exploiting user information is not the way to build good relationships with customers: advertisers lose because people won’t click ads; brand marketers lose because users won’t form positive associations with the advertising companies; and publishers lose because mistrusted ads are a constant drag on the website’s brand and revenue source.
This “everybody loses” situation is an over-arching problem with online advertising: web users cannot control what personal information (their habits, interests, behaviors, etc.) they trade for the online activities they value (product and service offers, socializing, local event notification, etc.) and, therefore, do not get things they value or trust the ads they do get.
An Example: A user may be required to give her age (34) to register to use a website. She agrees because the website claims that this is to insure that children do not use websites without supervision – i.e. as a safety precaution – but then she sees the information used to target her with ads for children’s clothing (she’s happily single with no kids). So, the ad is not only irrelevant, but it is also a surprise to her that her age is being used in a way she did not intend. She stops looking at ads, has a mistrust of the website brand, has a mistrust of the advertiser and, even when served appropriate ads, still carries a perception that the ad is not trustworthy or not for her.
The flaws in online advertising can be described as problems of relevance and trust. Relevancy has been greatly improved by behavioral targeting (even if privacy has not), but behavioral targeting is backward-looking. If you used the internet for a recent house purchase, for example, you really know what I mean. How long will you be receiving and ignoring those mortgage ads?
Contextual ads and demographic data are also used to produce relevancy in advertising, but site context is only a stereotype and demographic information is a best-guess. Just because a 54-year old reads blogs about classical music does not mean that he wants to buy retirement property more than he wants to buy the newest Nike sneakers.
To a consumer, relevancy is not only about getting what he wants, it is also about seeing what he expects. Relevancy is about expectation and emotion, but the content and subject matter of advertisements cannot always be consistent with user’s interpretation of the content and subject matter of the website, making the user feel that the website publisher does not understand why she visits the site.
Trust is even more emotional. If the user does not feel that he has consented to the ad or the information used to serve it, he consciously or subconsciously mistrusts the brands that brought it to him.
The hot topic in trust and advertising is behavioral tracking, of course. Tracking scares consumers and may invade their privacy. A relevant ad often signals to us that THIS is the brand that is tracking me the most, which is bad for any brand. Further, consumers have no control over or protection from unwanted or inappropriate advertising content: they have been removed from the discussion and forbidden from making those decisions.
This lack of empowerment, itself, destroys trust whether the online ads are “appropriate” or not. Even widely respected online publishers suffer the consequences because they have no way to lend their own brand credibility to the ads served on their sites. Publishers are often unable or unwilling to vouch for all ads that are served and, even if that were not the case, there is no understood signal to do so.
YouTube Enters Affiliate Business

YouTube blog: "Today [Oct.7], we're taking our first steps to providing YouTube users with this kind of instant gratification, by adding "click-to-buy" links to the watch pages of thousands of YouTube partner videos. Click-to-buy links are non-obtrusive retail links, placed on the watch page beneath the video with the other community features."
Apparently, not all of the wrinkles have been ironed out yet. The Sims 2 trailer above has an Amazon link to Spore (both games are by EA that is among YouTube's new partners). YouTube's aff id with Amazon is you09f-20 (probably one of many).
Incidentally, a similar affiliate model could work for Facebook, if the virtual gifts stats are any indication. See how Birthday Calendar app inserts a Send Flowers affiliate link to 1-800-Flowers.
Is Google More Expensive Than TV?
Denuo's Rishad Tobbacowala compares TV buys with Google search ads: "It has actually become far more expensive to buy advertising on Google than on network television. Google has a product called AdWords, which marketers use to bid on a particular key word [that consumers might type in during an Internet search]. On average, across all categories, it tends to be about 50 cents. Let's say on television you get a $20 cost-per-thousand rate. Fifty cents a click is equal to $500 cost per thousand. You can see how much more expensive it is, but the difference is there's some sort of action."
Media equivalency is a tough problem.If we compare CPM rates, yes, TV's CPMs are cheaper than search CPMs {umm, that was stupid), but Rishad's comparison is not entirely fair or accurate when viewed from at least two angles. First, you cannot call up a TV rep and say, "Hey, I've got 20 bucks here. Can I please have 1000 impressions today?" Second, Google impressions are free, at least to a certain point. It's the action (click) that advertisers are being charged for. A more accurate comparison would be the cost-per-click for an AdWord vs cost-per-call for a TV ad with a 1-800 number.
Media equivalency is a tough problem.
Subscribe to:
Posts (Atom)

