DISQUS

JX is exactly right. These "evaluations" of Facebook advertising are usually sloppy, do not make accurate comparisons, are based on a single creative, and are run by people with little or no online advertising experience. They are generally conducted with extremely low amounts of money and over short time periods, which doesn't help either.

Click-through rate hardly matters with CPC advertising on a site where users spend an average of 30 minutes a day and 50% login in daily. That is a ridiculous number of impressions per user and understandably will generate low click-through rates when there is no user intent driving the ad placement, but when you can get a million uber targeted impressions in a day, 800 clicks is pretty solid. Add in the new social referral system available through beacons and CTRs will no doubt improve.

Facebook is a social web application, not a search engine, so you need to accept low CTRs in exchange for insane targeting to users who have expressed no intent, and a correspondingly huge number of impressions. If you're so worried about CTR just do search advertising and get exactly the users who have self-selected an interest.

It's simple: you have to bid higher for ads that have lower CTRs, otherwise Facebook will stop showing them.

See my comments on how you are not advertising effectively.

http://www.allfacebook.com/2007/11/is-facebook-...
http://www.allfacebook.com/2007/11/sex-drives-f...

The reason why your USV.com ad received more impressions is because it has a much higher CTR - Facebook estimates a minimum bid price based on the CTR of the ad. If your maximum CPC is below this minimum bid, Facebook won't show your ad.

Yes, I have had fine advertising success with Facebook, as I have already stated. If your max CPC is $0.10 because you are an affiliate market (i.e. web spammer), you will not generate any impressions (or clicks) because your bid is ridiculously low. It's not rocket science people: set a realistic bid and you will get your clicks.

Your "mediocre ROI" link is 1) a terribly poor example and 2) a misuse of the term ROI.

1. The linked test is bad for several obvious reasons: they have a max CPC of $0.10, they have very narrow targeting settings, and they generated less than a thousand impressions. I checked the targeting and it yields a whopping 2,460 users.

2. ROI is a simple acronym: return on investment. You can calculate it as: profit/cost. Let's plug in the values for the linked test. Profit: 439 targeted impressions. Cost: $0 (outside of the minimal labor cost). How does that translate into "mediocre ROI"?

If you do a legitimate test of Facebook advertisements, you will find it is in fact competitive to other services.

In terms of pure sampling error, the margin of error for that poll at 250 people is +/- 2.7%, and +/- 1.4% if 1,000 people were sampled. What would be more interesting is to see the demographic breakdowns so that they could understand which users were more likely to pay.

A few points:

1. I think you mean AdSense, not AdWords.

2. Users can opt-out of tracking through a combination of behavior (don't use Google, et al.) and technology (browser settings, plugins, anonymizers, etc.).

3. To the extent that consumers want to avoid being tracked, there is a business opportunity for supplying such a service. But for it to become pervasive it will have to provide benefits that outweigh the personalization and targeting opportunities supplied by tracking.

4. As I said in my last comment, users can easily opt-out of Beacons.

What would a privacy bill of rights entail? Do you really think a political solution is a better tactic than a technical or behavioral solution?

I don't think the legal issue will hold, because users must consent to every beacon that is posted. If you don't want it posted just decline the action - no problem.

Btw, the group has registered over 4,000 Facebook users in a span of six days (actually, up to 5,100 now): http://www.mydd.com/story/2007/10/30/172215/51

Jeremy: do you mean 0.02% or 2%? Big difference there. Do you have a write-up somewhere? I'm doing some Flyer experiments as well.

When you're analyzing flyer data, you need to be careful as Facebook's algorithm is nuanced and self-adapting. In this case the affiliate marketer is strategically purchasing their flyers to minimize the CPC they have to pay (thereby maximizing profit). This can be done by creating large numbers of flyers and setting a low maximum CPC that is quickly reached. You're missing the point if you focus on CTR.

This is a rumor and has not been substantiated - don't you think your title is a little too certain? The story reported in the linked article isn't even believable (imo).

Interesting, this also explains why they recently changed their login interface to use permanent cookies, allowing multiple browser sessions to remain logged in.

This is amazing news - thanks for putting the word out! The targeting capabilities warm my heart.

Regarding the magic spells application, I will just report that there was no "recruit five friends" requirement when I added it a week or so ago. Maybe it was added more recently though.