Tuesday, August 26, 2008

Facebook Ads...not so great.

Some of you have emailed me about the CoNotes Facebook Ads that you have seen. I'm impressed that you have even seen them given the atrocious click-through rate that I have been getting. I am going to assume the only reason my friends have actually seen the ads is because I have emailed them about 50 times about CoNotes. So they are pretty aware of my logo and can recognize it at a quick glance.

I have been running ads for a few weeks now, and below charts the CTR data for the past 2 weeks. I ran the exact same ad with different university targets to see if I could detect any differences in market segments according to which university a person went to. The results were pretty inconclusive as you can see.

Read this document on Scribd: Facebook Ad Report

I had a theory that there was an underserved recruiting market at schools not typically targeted by startups. Therefore, I was expecting to see higher click-through rates at the liberal arts and top non-Ivy schools. While the average CTR of liberal arts schools was the highest, I can't really make any statements as of yet. I think I will really have to wait for spring to come around to really make any conclusions -- when college students tend to make their employment decisions. Running the ad in the summer time probably hit the alumni at those schools, which is also another interesting group to study.

Nevertheless, despite any trends that are there or aren't there, the CTRs are horrible. The highest CTR I received from ANY segment was 0.2%. Social networks are known to average about 0.1%, so my results confirmed those results. It's well documented that the users of social networks are not in the right mentality when looking at ads. They are there to find out what is going on within their "social graph" and to find some entertainment through social apps. This is much different from the transactional mentality a user has when he/she searches for "digital camera" on Google.

On the positive side, the leads that Facebook does send me (note: this is in the single digits per day) are quality leads. So the targeting functionality is pretty effective. And I'm happy that I can target ads to very specific self-defined interests of users and their school affiliation.

But here is my advice to Facebook. If you are going to build out an advertising platform, you really need to target it to the usage pattern of your users. So what do I mean by that? Well let's think about the main reasons people use Facebook:
  1. To find out about people in their network (in their network = friends or just people "in their network")
  2. To share their experiences with people in their network
  3. To spend time on social activities (not necesarily in their network)
Given these 3 main usage patterns, Facebook's advertising platform should enable different kinds of ads.

For usage pattern 1, self-expression advertising is the most logical fit. People self-affiliate themselves with brands and products. You have Apple fan-boys, Audi tuners, and Jimmy Choo shoe lovers (I'm not totally sure why I know about Jimmy Choo). Why not let companies advertise through these people? It could be a mutually beneficial relationship where brands pick specific people they want to advertise their brand through, and those people are compensated because of the special role they play in their social network (e.g. "mavens" or "salespeople" as Malcolm Gladwell would describe them). Compensation could be money, but even better is probably special privileges to products from the company. Brand companies are very concerned about diluting their brand value, so they want to make sure the right people are promoting their brand (e.g. Kobe lost a lot of endorsement deals after his rape allegations). So to enable such a advertising platform, Facebook would need to build a mechanism for advertisers to find which users to advertise through.

Usage pattern 2 lends itself to more of a transactional advertising model. People often post pictures of food from a restaurant they ate at or write a post about the hotel they stayed at. These are all opportunities to enable transactional commerce. WOM advertising is strong, so why not enable advertisers to take advantage of the best WOM advertising -- through your friends. Imagine that your friend took a picture of an awesome slice of prime rib from Lawry's. If I hadn't eaten at Lawry's, I would either want to save it to a list of "places I want to try" or maybe even make a reservation.

Usage pattern 3 is pretty hard to create a specific advertising platform for. But it doesn't really matter, as this type of usage is actually enabled by 3rd party Facebook apps. Facebook app developers currently can put their own ads on their apps, but it would be helpful as a developer if Facebook enabled ads through their own API. Basically Facebook would create an inventory of ads that developers could tap into and place on their own apps. Depending on the nature of the app, developers could select the best types of advertisements.

The difficult thing about some of my suggestions are that they would be changes in behavior for advertisers. Advertisers like buying advertising products with defined standards and specs. However Facebook has an opportunity to really shape social network advertising if they build out the right advertising platform.

So that's it. Do I really have any evidence to say this would be much better than the current Facebook Ads or their even worse Beacon product? Not really. But I'm pretty sure it is a step forward.

No comments: