SXSW 2005 Panel: Making Money with Online Ads

Friday, March 18, 2005

[This is a (mostly complete) transcript of Philip Kaplan, Jason Calacanis, Gokul Rajaram, Henry Copeland, and Bill Flitter talking about making money with online advertising. Official details about the panel here.]

Calacanis: I now run Weblogs Inc. that runs about 75-or-so boutique blogs, kind of like Gawker. So this is a pretty good panel of people who make tools that can help people like you make money.

[Introductions.]

Calacanis: Let me start with Google. How many advertisers in the network? And why don’t you tell us the percentage of the revenue we get?

Rajaram: There are thousands of advertisers. We don’t reveal because we want publishers to focus on the actual number, the actual money they’re making. And we want to reserve the flexbility to reward or publish of publishers who do well or poorly.

Calacanis: Isn’t the real reason because people are constantly trying to game the system?

Rajaram: That’s the third reason. But really, we’d prefer the users focus on the actual money.

Calacanis: [Pulls up a screenshot of a page with Yahoo! Marketplace ads.] If Yahoo! displays percentage, will Google?

Rajaram: No, because it detracts from the ECPM [effective CPM].

Calacanis: Let me ask you, Phil. You learned a lot with FuckedCompany. Now you’re persuing Google’s business. Why?

Kaplan: Running FuckedCompany was and still is a big publication. We got a lot of requests to link to peoples sites, but we don’t do this. We tried advertising in some ways. So we built a system to automate the system. Not trying to compete with Google, we just filled a need. We give publishers 75%. And we tell people what the average cost per click is. Total transparency. We even show page views. We also show repurchase rate. [Looking at AdBrite site on the screen.] And we have no exclusives.

Calacanis: Google, also, no exclusive. Does this make Google nervous?

Rajaram: It’s very complimentary. There’s not just one revenue option. There are other options that attract other sets of advertisers.

Calacanis: If you have ads that look similar …

Rajaram: You can’t run contextually-targetted ads on the same page with GoogleAds.

Calacanis: I’ve heard the Overture wants an exclusive. Google doesn’t. Now, to Henry. BlogAds. Very focused. You chose to do a different style of ads.

Copeland: Our theory was that blogs and blog audiences were very different. So we have different kinds of units. Kind of a bloggish unit. It helps get advertisers into the blog mindset. We’re very friendly — publisher does what they want to do.

Calacanis: Google forces pubs to put “Published by Google” in their ads.

Rajaram: We want to make sure the users know that it’s really an ad and not a part of the content of the site. The success of any program is dependent upon the advertisers being satisfied.

Copeland: [Showing ads on Wonkette and talking about what’s a success and what’s a failure.]

Calacanis: I have a little problem with it as a publisher. It’s hard to tell what’s an ad sometimes. Why not label?

Copeland: We’re here to serve bloggers. We’re not going to tell them what to do. Except don’t serve porn.

Calacanis: We started selling RSS ads back in July or August. Not a lot of demand for them. Tell us about Pheedo. And your issues with RSS advertising.

Flitter: Basically what Pheedo does. It’s an ad server plus analytics engine. We to want to ad value to that process. Ad content that’s relative to that feed. But you’re right — RSS does have some unique problems to it that we’ve solved. Like understanding your stats. Like where expired ad links go to. And what standards to use. There are no standards. So we’ve been working together to make these standards. Top publishers: Can’t talk about some right now. Topix.net. Lockergnome. Paidcontent. And they’re rebuying. Advertisers pay by click. Measuring impressions is part of our secret sauce.

Calacanis: One way to tell if an ad has been seen is to add a 1x1 pixel image that no one can see. We get a lot of traffic for spiders from Technorati and all these other places. And you can’t block them or you’re not getting indexed on Technorati. DoubleClick came up with this a long time ago.

Kaplan: I looked at the P&L statement when they went public and it came out to between 25% and 32%, what they take for the ads (publishers get 75% or 68%).

Calacanis: We pay our writers based on traffic. And our big blogs (engadget) pay for the smaller blogs down the line, like a cancer weblog. The goal of the company is to have 700 topics, so that’s the goal.

Question: What is your position on full-text versus partial text in feeds. Does it impact readership?

Calacanis: If it starts canibalizing on our viewership, we’ll change it.

Question: Two schools of thought. One says that RSS should just be a notification system. The other is that you should post full text in the RSS and do ads.

Calacanis: We don’t do that.

Question: Talk to me in a couple years. You will.

Calacanis: I think there will be a trend towards giving more to the user and less “selling out.”

Question: Can I hear some success stories?

Copeland: A handful of bloggers are making more than I am right now.

Calacanis: Mobiletracker makes $5-$10k per month from his site. paidContent.org also makes about $70,000/year.

Flitter: One thing he does do. paidContent is very particular about what ads he’ll accept. It has to be exactly related.

Calacanis: He’s doing a digital music blog with Billboard. They’re paying him. And then he’s building his own brand in parallel.

Question: [About money.]

Calacanis: A Gawker writer makes $30,000/year.

Question: What sort of rates would a blogger get. How do you pay a part timer?

Calacanis: Anyone getting paid to blog is paid because they have a passion and because they have other revenue streams. We like the idea of having four or five blogger working on a site and then doing something else.

Question: Some ad networks run out of ads and you end up getting ugly, flickering ads for mortgage rates and such.

Copeland: We let publishers decide. The number of ads shown floats depending on how many fit.

Kaplan: We give you the option as to whether you want those filler ads.

Question: Google, are you punishing or rewarding me?

Rajaram: We believe that if advertisers are not happy, they will go elsewhere. There are a ton of different factors and we don’t publicly disclose them. If your clicks don’t result in sales, then the amount of money you get goes down. Protects us from click-through fraud a bit.

Question: Follow up. Can publishers, then, be punished when the ad doesn’t represent the actual product or service.

[…]

Calacanis: Google created a link that’ll go through your page and redraws the page with their own links on keywords. [I don’t fully understand.]

Rajaram: We should have an offfline discussion. It’s a sensitive topic.

Copeland: “Do no evil” is an interesting proposition. We try to do good, in the positive. I admire their engineering. Philsophically it’s odd.

Calacanis: It’s one of those things that’s technically possible and sounds cool. Then you make it, and it’s not that good of an idea.

Question: How do you deal with click-fraud?

Rajaram: We think we’re very good at finding false-positives.

Kaplan: We do some per-click and some flat-rate stuff. For a click to count, it has to jump through a bunch of hoops.

Question: Lots of low-quality sites in AdSense. How do you deal with this?

Rajaram: Combination of technology and humans terminate problem sites.

Calacanis: It’s a total non-issue. A technology starts doing well, and the press latches onto the downside. It doesn’t happen that often. It’s a manufactured controversy.

Question: How do you get new advertisers?

Copeland: I think Google has about five-hundred people. We have two right now.

Kaplan: We have about five or six people right now. We got $4mil in venture capital..

Flitter: We have three.

Question: I heard that 15% of Google clicks are fraudulent.

Calacanis: That’s outlandish.

Rajaram: I don’t know. Not that high. We have a large fraud-detection department.

Question: […]

Rajaram: Content has lower click-through rate than search. [Etc. More about this…]

Copeland: All clicks and conversion are not created equally.

[Note: I might alter this transcript in the coming days to remove typos and clarify the content. I post it now just to make it available as soon as possible.]